[BOOTLIB]: When failing to find a block device, keep going searching for more, instead of giving up (critical, because the CDROM FAT12 image is now device path #1, not #0).
[BOOTMGR]: Correctly use the right logical operator in EfiInitpGetDeviceNode to get the deepest-level media device node. We now get the CDROM node, not the raw node.
[CDMAKE]: Don't actually create an EFI/BOOT directory on the CDROM itself, but rather in the FAT12 image. Otherwise, this can confuse UEFI implementations to boot the boot manager off the raw CDROM, instead of the FAT12 image on the CDROM.
svn path=/trunk/; revision=70542
[BOOTLIB]: Additional text and graphics console support. Among other things, we have now gained the ability to fill a buffer with a particular color.
[BOOTLIB]: OEM Bitmap (ACPI 5!) support.
[BOOTMGR]: Continue making a few more initialization calls now that bootlib supports more functionality.
svn path=/trunk/; revision=70539
Update the registry configuration to reflect the recent changes to rosvboxmgmt.
Just uncomment the line matching the behavior you prefer.
CORE-10032
ROSAPPS-303
svn path=/trunk/; revision=70537
Introduce another automagic thing. Instead of creating shortlinks once shared folders have been enumerated, just assign them a drive letter!
See it in action: https://twitter.com/HeisSpiter/status/685229912748371968
Choose the one you prefer for 0.4.0 :-)
CORE-10032
ROSAPPS-303
svn path=/trunk/; revision=70536
- Reenable RTL_QUERY_REGISTRY_DELETE: the "PendingFileRenameOperations(2)" and "AllowProtectedRenames" registry values are deleted after being parsed: we don't retry at every reboot to remove/rename files that were already deleted/renamed.
- Remove superfluous parentheses.
- Warn a FIXME if SFC-protected file renames are allowed and we are processing file renames (the "protected file renames" stuff is unimplemented atm.)
svn path=/trunk/; revision=70530
- Use CRT string functions to loop through multi-strings.
- Do not fail the full KnownDll setup if one dll fails to load. Instead, just properly skip the file (and related to that, fix a bug when skipping excluded known dll files).
- Page files should be created only when we are not in WinPE mode (aka. MiniNTBoot = FALSE).
- Formatting fix in smloop.c, no code change.
svn path=/trunk/; revision=70529
- Add support for opening ICO, EMF, TIF and WMF files with the shell image viewer since it already implements what is needed (via gdiplus).
- Add some missing "Content Type" (MIME type) and "PerceivedType" values for the image extensions.
[SHIMGVW]
- Enable the resource strings for EMF, TIF and WMF files since we actually support them.
Should be merged for 0.4 branch.
svn path=/trunk/; revision=70528
- Implement support for manifest prober routine in LdrpWalkImportDescriptor, in its simplest way
- Implement such routine in kernel32 (that's mostly a copy of Wine's create_module_activation_context which seems to nicely correspond to what Windows does).
- As create_module_activation_context is moved into kernel32, delete it from ntdll
- Improve debug output in RTL's Activation Context implementation to ease debugging
- Now, ReactOS indeed looks for and actually loads manifests when loading modules. Thanks to http://blog.tombowles.me.uk/2009/10/05/winsxs/ for a really good description of how activation contexts work
CORE-7313
svn path=/trunk/; revision=70522
[BOOTLIB]: Fix incorrect use of flag introduced a few revisions ago. This fixes <1MB RAM descriptors.
[BOOTLIB]: Enable BmFwMemoryInitialize, since it works now.
All (known) bugs are fixed, bootmgr is back on track.
svn path=/trunk/; revision=70516
[BOOTLIB]: Fix two big device I/O bugs (it's a miracle it worked)
[BOOTLIB]: Use defined constants and comment some more device I/O code.
svn path=/trunk/; revision=70515
[BOOTLIB]: Add a bunch more graphical console support functions.
[BOOTLIB]: Cleanup some older graphics-related code.
[BOOTLIB]: Support graphics re-initialization. We now correctly fallback to text mode since font loading is not implemented.
svn path=/trunk/; revision=70514
Improve the bootvid font generator of Colin (see r70507) in many ways:
- Allow the user to specify the font name (optional), and optionally a font size and x- and y-offset so that (s)he doesn't have to recompile the generator each time (s)he wants to test a new font.
- The default settings (in case the user doesn't specify all or part of the font definition) are those of Windows' bootvid font (i.e. "Lucida Console", font size 10, x_offset = y_offset = 0).
- Create only once the DC and the font handle (to be used for all the generated characters) instead of each time a character is being generated. I still keep the recreation of the HBITMAP for each character (I don't know if we can also improve there by creating it once and zeroing it out before generating each character).
svn path=/trunk/; revision=70512
Add a commented out way to automatically start `rosvboxmgmt auto' on start so that shortcuts for VBox shared folders are properly created.
This is to be enabled on 0.4.0.
CORE-10032
ROSAPPS-303
svn path=/trunk/; revision=70510
Change the Blue Screen Font hardcoded into bootvid.dll to the Open Source "Anonymous Pro" font, which makes a nice 8x13 font.
Blue Screens now look like this: http://fs5.directupload.net/images/160106/ehv6245t.png
[BOOTVID_FONT_GENERATOR]
In case you find an even better font, you now have a nice little tool to test that font and generate the corresponding FontData array for bootvid.
svn path=/trunk/; revision=70507
[BOOTLIB]: Add mostly full support for resource and locale management.
[BOOTLIB]: Stub out bootstat.dat (Boot Status Data Log) support.
[BOOTLIB]: Some refactoring.
[BOOTMGFW]: Add a message table with a few initial strings. Also add an empty bootmgr.xsl stylesheet. We'll have to see what we want/need to do here.
[BOOTMGFW]: Correctly loading HTML and message table resource data, and add a few other additional initialization calls.
Next steps are to re-visit graphics code to support re-initialization.
svn path=/trunk/; revision=70504
Add a new feature to the tool:
Let it auto start VBox Shared folders and to browse all the available shares in order to create shell links on desktop.
The purpose is to workaround the missing network shares discovery feature in ReactOS while keeping the VBox shared folders usage 'user-friendly'.
It has been designed specifically for the coming 0.4.0 release: just put it on autostart for the default user. If there are shares (even new ones), it will create links, if there are no shares, no guest additions, no VBox, it will just exit.
See it in action: https://twitter.com/HeisSpiter/status/684506579555741696
CORE-10032
ROSAPPS-303
svn path=/trunk/; revision=70500
- Use the LockQueueIoDatabaseLock spinlock to guard the device list instead of blindly raising the IRQL
- Add missing lock to IopEditDeviceList
svn path=/trunk/; revision=70496
- Fix a bug in FltpDetachFromFileSystemDevice so it correctly bails when we've walked the attached device list.
- FltpDispatch can come in at high IRQL. Thanks to Thomas for noticing that err.
- Add newlines to the end of DPRINTS (it's been a while...)
- The filter now loads and runs in the reactos FS stack.
svn path=/trunk/; revision=70495
[BOOTLIB]: Implement routines to enumerate subkeys and to read a key value.
[BOOTLIB]: Implement routines to parse, enumerate, and convert registry data into BCD elements, and to convert BCD elements into boot library BCD options.
[BOOTLIB/BOOTMGFW]: Lots of cleanups, mainly around the fact we now use ntintsafe.h instead of the manually self-inlined code from before, and from documenting additional flag values, and from using our newfound BCD powers.
[BOOTMGFW]: Implement BmGetOptionList, BmpUpdateApplicationOptions.
[BOOTMGFW]: Prepare for post-BCD library reinitialization. We correctly read the 3 BCD options so far in the hive.
svn path=/trunk/; revision=70492
- Plug in the dispatch routines. These are just pass through methods for now to get the filter up and running.
- Implement the FastIo handlers. The majority of these call the FastIo routines of the attached device object.
- Make sure we detach from devices that are being deleted in FastIoDetachDevice.
- Move the FastIoDetachDevice routine to a deferred call as it's too expensive to tie up a FastIo request.
svn path=/trunk/; revision=70488
Fix the fix from r70482: Our DIBs are actually 32 bit aligned, so pass even uncompressed bits to GreCreateBitmapEx instead of copying them later with wrong alignment.
svn path=/trunk/; revision=70486
NtGdiCreateDIBitmapInternal: clear CBM_INIT flag, if no bits are given.
IntSetDIBits: Always calculate the bitmap size for uncompressed bitmaps, do not pass uncompressed bits to GreCreateBitmapEx, since they don't have the required alignment (the width in bytes for DIB is 16 bit aligned, bitmaps 32).
Remove a broken ASSERT.
svn path=/trunk/; revision=70482
Implement LsarStorePrivateData and LsarRetrievePrivateData.
[ADVAPI32]
Implement LsaRetrievePrivateData and LsaStorePrivateData.
Encryption of secrets (aka private data) is NOT implemented yet!
svn path=/trunk/; revision=70479
Factorize the code that creates ReactOS freeldr.ini entries, and use it to create a freeldr.ini at a clean ReactOS installation, and when it is installed in the same partition as Win9x/DOS.
CORE-4870 #comment Should be fixed with revision 70467, please retest.
svn path=/trunk/; revision=70467
Properly check the bitmap size in IntSetDIBits, taking into account that it might be set to 0, then we have to calculate it ourselves. Fixes a crash when invalid data is passed to NtGdiCreateDIBitmapInternal.
[GDI32]
Convert COREINFOHEADER to BITMAPINFOHEADER before passing it to NtGdiCreateDIBitmapInternal, which doesn't support it.
CORE-10583 #resolve
svn path=/trunk/; revision=70464
Export NtUserGetClassInfoEx (as it's called in 2k3) both as NtUserGetClassInfo (as it's called on XP) and NtUserGetClassInfo. They have the same prototype.
svn path=/trunk/; revision=70463
- "Fix" an export name: NtUserGetClassInfoEx on 2k3 is the same as NtUserGetClassInfo on XP (note that we currently use the default, aka ros specific, win32kdll on the WHS testbot, but it's mostly identical.
[WIN32KNT_APITEST]
- Fix a typo
- add missing tests to testlist.c
svn path=/trunk/; revision=70462
- Convert win32k native api test to actual wine style api-tests
- Hack around a bit with the win32k dlls, some renaming, etc.
- Delete old apitest stuff
svn path=/trunk/; revision=70458
Don't use DATA and PRIVATE together. DATA means no stub function is generated in the import lib, PRIVATE means nothing is generated in the export lib.
svn path=/trunk/; revision=70454
Now, the fix: the FAT disk checker expected that the free cluster count member of the FAT32 FSInfo sector be really initialized (not to 0xFFFFFFFF), and it expected the presence of the backup of the FSInfo sector as well as the free sector following the FSInfo sector.
svn path=/trunk/; revision=70449
- DPRINT messages printed via VfatPrint by the check-disk procedure.
- Addendum to r70434 and r70435 : The return value of fs_close is equal to the number of corrections made on the disk (0 corrections --> success; 1+ --> something was corrected).
That way, the old code "return fs_close(FixErrors) ? STATUS_SUCCESS : STATUS_UNSUCCESSFUL;" that we had since eons (which always happened to return STATUS_SUCCESS) was actually
indicating to us that our FAT32-formatted volumes were corrupted.
If you test this revision, you should see something happening if you try to install ReactOS now (something that should have happened since ages actually).
svn path=/trunk/; revision=70447
- Formatting + RtlZero/FillMemory modifications as addendum to r70445.
- Do not hardcode FSInfo signature values, use instead the same set of macros fastfat_new introduces.
svn path=/trunk/; revision=70446
Fix handling of NULL buffer in streamout. Fixes a few msvcrt_crt_apitests, while it breaks a few for crtdll. crtdll is broken anyway and obsolete, nothing uses it, so it doesn't really matter.
CORE-9105 #resolve
svn path=/trunk/; revision=70443
It seems suspicious to directly return the value of fs_close as the NTSTATUS code of the check-disk operation (for FAT32 volumes it happens to return 1 whereas for FAT16 volumes it returns 0).
The documentation of this function says that it "returns a non-zero integer if the file system has been changed since the last fs_open, zero otherwise."
Maybe somebody has a more precise idea on that subject? In the meantime I also add some DPRINTs to attempt to diagnose the conditions where this problem occurs.
svn path=/trunk/; revision=70434
- Gather common code in one file instead of duplicating it everywhere.
- Implement full format for FAT12 and FAT16 (using the same function Fat1216WipeSectors), based on similar code for FAT32 implemented by Eric Kohl in revision 63693.
svn path=/trunk/; revision=70433
- fat.asm: remove trailing whitespace.
- fat32.asm: backport the fixes from fat32.S introduced a long time ago.
- fat32.S/faty.S: use .space 12, 0 instead of the long .byte array (generates the same code, but is better readable).
- isoboot/btrt: whitespace fix only.
svn path=/trunk/; revision=70432
Add another test for strlen, that tests, whether the direction flag in ELFAGS has been modified. While clearing it is legitimate to do according to the ABI, the native implementation does not change it, so we don't want to do it either.
svn path=/trunk/; revision=70429
- disable GCC builtins
- Add simple tests for strlen, showing that a NULL pointer will cause an access violation, which is broken in our asm implementation
svn path=/trunk/; revision=70427
[MSVCRT] Explicitly pull in _vsnprintf, bsearch and strcspn from CRT, to prevent the MS linker from exporting stubs from libntdll instead.
See CORE-10753
svn path=/trunk/; revision=70425
- Use the old fat.S source for compiling the FAT12 boot sector, until FATX.S is completely implemented.
- Fix an assembly error in FATX.S
- In comments I added experimental FAT12/16/32 targets based on Timo's FATX.S.
svn path=/trunk/; revision=70423
- Implement AbortSystemShutdownW by a call to BaseAbortSystemShutdown.
- Implement InitiateSystemShutdownW by a call to BaseInitiateSystemShutdown.
- Implement InitiateSystemShutdownExW by a call to BaseInitiateSystemShutdownEx.
- Add RPC binding code for the winreg server (in winlogon.exe).
svn path=/trunk/; revision=70420
SHUTDOWN: Fix the hack be calling InitiateSystemShutdownExW instead of ExitWindowsEx.
WINLOGON: Add a hack by calling ExitWindowsEx in BaseInitiateSystemShutdownEx.
svn path=/trunk/; revision=70419
Disable again the DR7 check until we correctly understand why bit 10 of DR7 (reserved) is almost always set (this lead to MSVC-built ReactOS crashing when being starting in non-debugging mode). See also r69244.
CORE-10165 #comment DR7 check again disabled in r70418.
svn path=/trunk/; revision=70418
For MSVC builds (using KDCOM kd64 WinDbg protocol): Bail out for droppable packets after a given number of retries, if no debugger is attached (list of droppable packets can be found in http://articles.sysprogs.org/kdvmware/kdcom.shtml section "KDCOM protocol", subsection "Droppable packets").
CORE-10541 #resolve #comment Finally fixed in revision r70417!
CORE-7106
svn path=/trunk/; revision=70417
Fix the value for EIP used by KDBG after an INT3 set by KDBG itself. The address is already fixed by KiDispatchException, but only in the context frame, not in the trap frame and KDBG insists to use the trap frame for a lot of things. Also, after a cont from such an int3, KDBG uses a single step to re-enable the breakpoint (it needs to disable it after it was hit to be able to execute the actual instruction), but it used to dismiss *any* single steps after that. So make sure, that an actual single step, as created by the debugger is not being dismissed, but the break point is still restored after the next single step entry. You might expect that a kernel debugger would at least support setting breakpoints, but this is KDBG.
svn path=/trunk/; revision=70416
- Change the start type to demand.
- Should stop free loader from throwing a tantrum because it can't load a non-existant boot driver
svn path=/trunk/; revision=70414
Correctly fix the definition of DRIVER_FS_NOTIFICATION (done the same way as the other DRIVER_xxx "callback" functions; by the way you'll notice they are all NTAPI aka. __stdcall. This is not explicitely mentioned in the W(D)DK, because MS supposes you compile all your kernel-mode code in stdcall convention, the WDK environment coming with preset default compiler switches enabling that. But if you try changing them, you'll run into big troubles. In our headers on the contrary we explicitely mention the calling conventions).
[FLTMGR]
Fix FltpFsNotification to adhere to the correct DRIVER_FS_NOTIFICATION definition.
Addendum to r70410.
svn path=/trunk/; revision=70413
Pass cjMaxBits parameter from GreCreateDIBitmapInternal to IntCreateDIBitmap and check the image size before creating the bitmap.
Should fix crash when hitting about tab in Office2000 WinWord.
CORE-10583 #comment please retest
svn path=/trunk/; revision=70412
- Add the beginning of the fltmgr. It's called rosfltmgr for now so I can test in Windows.
- It's currently just base code which registers for file systems appearing (or disappearing), and attaches itself to them and all other mounted devices in their chain.
- Although it builds (touch wood), don't add it to a running system. The IRPs aren't plugged in yet and it'll just bugcheck lower down the stack.
svn path=/trunk/; revision=70409
- Raise the IRQL when enumerating device lists so it doesn't get edited mid-listing
- Don't hardcode the pointer size when checking the buffer size
svn path=/trunk/; revision=70408
Save parameter registers (rcx, rdx) before calling CallRealMode. Patch by andy-123, applied to a secondary location by me.
CORE-10511 #resolve #comment Thanks. I added the same fix to PnpBiosGet below.
svn path=/trunk/; revision=70407
Hack, that fixes the "Volatile Environment" bug by deleting the "Volatile Environment" key before the user hive is unloaded.
CORE-10705
svn path=/trunk/; revision=70406
Do not use the "LoaderBlock->u.I386.CommonDataArea;" hack to use FreeLdr's DbgPrint function inside KDCOM, for the simple reason that while it works when KDCOM just initializes, the memory area where FreeLdr's DbgPrint function is, gets erased when ReactOS is running. As a result, all attempts to call the DbgPrint function when tracing something in KDCOM at the end of life of ReactOS, just fails lamentably (it hangs).
We therefore cannot rely on external code to provide us with DbgPrint. Instead, implement a very simple DbgPrint function inside KDCOM, as done by KDGDB.
The KD serial port serving for debugging KDCOM is chosen dynamically amongst the other free COM ports.
This needs to be also fixed in KDVM.
[KDGDB]
Instead of hardcoding the KD serial port number serving for debugging KDGDB, determine it dynamically amongst the other free COM ports.
svn path=/trunk/; revision=70405
- Allow to specify in freeldr.ini the debugging port parameters, similarly to what NTLDR allows to do: http://naesten.blogspot.fr/2010/07/undocumented-debug-section-in-bootini.html
- Update the FREELDR.INI example file to reflect these changes.
- Because some FreeLdr components, that need to be initialized before the debugging port parameters can be read from freeldr.ini, use debugging features, we still need to choose some default parameters. For serial ports, we select the possible port starting from COM4 down to COM1.
CORE-9023 #comment Revision 70403: Allow to specify a debugging port in freeldr.ini .
svn path=/trunk/; revision=70403
When the root item in the treeview is selected, remove all items from the listview and wipe the address bar and status bar.
svn path=/trunk/; revision=70402
- Improve implementation of RtlActivateActivationContextUnsafeFast / RtlDeactivateActivationContextUnsafeFast by replace magic numbers by flag values, which are already defined in rtltypes.h, and adding various debugging checks. Two of them are triggered for yet unknown reason:
* Assert in RtlActivateActivationContextUnsafeFast
* "Trying to activate already activated activation context"
They are commented out in trunk not to annoy everyone.
svn path=/trunk/; revision=70396
Initialize the defauls users locale settings based on the selected locale Id from first stage setup.
#CORE-5989 #resolve
svn path=/trunk/; revision=70389
Fix bugs in pen implementation:
- Do not use the x coordinate to adjust styles, instead they start where the lines start
- Don't leak allocated styles
- Make sure the PEN fields are initialized correctly, even for BRUSHES, so that the destructor can do it's cleanup work
- Fix numerous parameter checks
gdi32_apitest:pen now shows 0 failures!
svn path=/trunk/; revision=70388
- Add french translation (not complete, but enough for now: only what can be currently displayed is translated).
- Don't hardcode the "History" item in the travel-log, but use the localized string. Use the hardcoded string *only* if loading the translation failed.
svn path=/trunk/; revision=70376
Fix hebrew "registry editor" translation.
[SHORTCUTS.INF]
Fix regedit shortcut name (localize it instead of using "regedit")
svn path=/trunk/; revision=70373
Rewrite the two existing tests for ICMP and add a third one (for IPv6).
Patch by Tim Crawford.
ROSTESTS-212 #resolve #comment Commited in r70367. Thanks!
svn path=/trunk/; revision=70367
Surround welcome & debugging messages in #ifndef NDEBUG blocks, that should be compiled in trunk but will be disabled in the release.
This is to "cure" the bad-looking problem described in http://www.reactos.org/forum/viewtopic.php?f=25&t=14620 .
svn path=/trunk/; revision=70365
Apply patch by Mark Jansen (learn_more) with a few tweaks. Fixes calls to Shell_NotifyIconA such as those from VBoxTray.
CORE-10497 #resolve
svn path=/trunk/; revision=70360
- When cleaning the clipboard, cancel any pending delayed rendering and increment all the sequence&serial numbers.
- Fix the clipboard access check when performing a delay rendering. This fixes VBox host <-> guest copy/paste amongst others...
CORE-5512 CORE-9620
CORE-8367 CORE-8627 #comment Please retest with revision >= 70355.
svn path=/trunk/; revision=70355
- Reduce indentation level of some code.
- No need to initialize pWinStaObj when it is reinitialized just after.
- Fix some comments.
svn path=/trunk/; revision=70354
Fix the computation of where bitmap pixels start, depending on whether we are dealing with DIBs or DIBs version 5+ (depending on the header BITMAPINFOHEADER or BITMAPV5HEADER etc...).
I also noticed a bug in Windows that few other people already discovered, related to the automatic conversion of CF_DIBV5/CF_BITMAP to CF_DIB clipboard format. See the comment in the code.
CORE-10679
svn path=/trunk/; revision=70349
- Add support for scrolling with the mouse wheel, by Ricardo Hanke.
- No need to check for CloseClipboard failure & display an error message if the clipboard was successfully opened just before.
- Use ARRAYSIZE macro instead of hardcoding the length of available clipboard format list.
CORE-10679
svn path=/trunk/; revision=70347
* Implement AppBar docking (defaults to floating because we don't properly support AppBars yet -- works in Windows).
* Implement loading and saving the AppBar settings from the registry.
* Remove the minimize button from the options dialog.
CORE-10691
svn path=/trunk/; revision=70344
* GetGUIThreadInfo: When they said "foreground thread", they meant the one with the keyboard focus, not the currently running one.
NOTE: This needs verification, as I have no idea how to properly handle IntGetFocusMessageQueue() returning NULL.
[MAGNIFY]
* Make use of the new-found ability to call GetGUIThreadInfo with 0 as the threadId.
CORE-10691
svn path=/trunk/; revision=70338
* Use GetGUIThreadInfo instead of multiple separate "hacks" to obtain all the necessary data in one single call.
* Keep the off-screen DC for later use, and only recreate when needed.
* Restore caret follow back to default-on.
More changes toward CORE-10691
svn path=/trunk/; revision=70337
* Replace the menubar with a notification icon and context menu.
* Change the dialog button to "Ok" instead of "Exit" (applied to en-US and es-ES only, the translators will have to do the rest :P).
* Due to a rather bad problem presumably caused by AttachThreadInput, set the "Follow Caret" option to disabled by default. This can be re-enabled once the problem is fixed.
More changes toward CORE-10691
svn path=/trunk/; revision=70336
Some improvements to the drawing code:
* Draw as soon as the mouse/caret/focus changes, or after 100ms if nothing changed.
* Copy only the needed area of the screen into the internal buffer.
* Blank out the area of the buffer where the client region of the magnifier overlaps, to avoid re-zooming the pixels.
* Keep the source region contained within the closest monitor, instead of only the primary one.
* Invert colors when copying from the screen, instead of when stretching, since there's less pixels to draw.
First step toward the improvements described in CORE-10691
svn path=/trunk/; revision=70335
- Adjust the type of some variables.
- Correctly set the fClipboardChanged flag to TRUE when emptying the clipboard: now this correctly triggers sending the WM_DRAWCLIPBOARD message to windows. Adapted from a patch by Ricardo Hanke.
CORE-10351
svn path=/trunk/; revision=70334
Handle WM_PALETTECHANGED and WM_QUERYNEWPALETTE: realizes any palette found on the clipboard, if there is any, if it receives one of those messages.
svn path=/trunk/; revision=70329
- Display the application icon in program's about dialog.
- Add missing DT_NOPREFIX flag to DrawText() calls.
- Let ShowLastWin32Error() show an error message even if the last error code is zero.
- Remove the calls to OpenClipboard() and CloseClipboard() from the functions in winutils.c and let the caller open and close the clipboard instead.
Patch by Ricardo Hanke
CORE-10657
svn path=/trunk/; revision=70328
Check RtlCreateUnicodeString in SetupCreateSingleDirectory.
Patch by Victor Martinez.
CORE-10647 #resolve #comment Thanks a lot!
svn path=/trunk/; revision=70321
- Don't disable the user profile creation hack until it's actually fixed (NtUnloadKey(NULL) doesn't do anything and I don't know what's actually supposed to go there). Should avoid some user profile corruption after crashes in 3rd stage (since 2nd stage will have flushed the registry).
CORE-10381
svn path=/trunk/; revision=70320
- Don't go past the scroll range
- Reset the grid/cell info when a new font is selected
- Should fix some of the issues Vort listed in CORE-10518
svn path=/trunk/; revision=70319
Remove do { } while (0) constructs from PSEH3 macros, so that a break from within a try block does what it is supposed to do.
svn path=/trunk/; revision=70316
- Fix setting the correct font in certain code paths
- non-eisting glyphs should return 0xffff, not 0x001f
- Don't set last error on failure, GDI APIs don't do that
- Remove GreGetGlyphIndicesW, it was mostly a copy of NtGdiGetGlyphIndicesW and it wasn't used
- Fixes the GetGlyphIndices tests and fixes showing invalid fonts in charmap
svn path=/trunk/; revision=70315
- Don't take a reference on the device object in ScsiClassClaimDevice since it's not going to be released before device removal. Inspired by classpnp. Fixes ejecting mass storage devices.
CORE-8911 #resolve
svn path=/trunk/; revision=70310
- Get rid of the global SetupData variable. Allocate the setup data struct from heap instead and pass its pointer to all wizard pages.
- Keep the setup data pointer in all wizard pages where it is needed.
- Keep the handle to the unattended.inf file open as long as the setup wizard exists.
svn path=/trunk/; revision=70309
Properly return unicast addresses in GetAdaptersAddresses() by properly return a struct sockaddr_in.
This fixes a bunch of failing tests in iphlpapi_apitest, and it should fix any application relying on this API to know interfaces addresses.
svn path=/trunk/; revision=70292
Add another check that shows that data returned by GetAdaptersAddresses() are incomplete and thus broken...
"This isn't the type you're looking for." would say an old man
svn path=/trunk/; revision=70290
Arrange large and small icons after changing the view.
Patch by Carlo Bramini.
#CORE-10350 #resolve #comment Thanks a lot!
svn path=/trunk/; revision=70288
Little Improvements:
* Duplicate code removement regarding the simple list pages (and now everyone has pageup/down + selection via char)
* Duplicate code removement regarding the partition type string
* Document the pages
Patch my Michael Fritscher
#CORE-10012 #resolve #comment Thanks a lot!
svn path=/trunk/; revision=70283
DispTdiQueryIpHwAddress(): use the correct function to get the interface associated with an IP address.
Also, because local links aren't in the ARP table (obviously), handle them without the mechanism described previously
This fixes arpinging oneself.
svn path=/trunk/; revision=70277
- Fix a bug IRPFinish(): do not touch the IRP if there's a pending return. Thanks to Thomas for his help!
- Implement DispTdiQueryIpHwAddress(): the current implementation is not optimal, but works. It first reset the ARP table entry corresponding the to target host, if it exists. Then, it sends the ARP probe and wait till the ARP table is updated to return the HW link address.
Could be improved, but it would require more important work in the driver, which is not a good solution when we're to release ;-).
Obviously, this fixes arping not working in ReactOS: https://twitter.com/HeisSpiter/status/673088515647930368
And this fixes a few iphlpapi_apitest tests.
svn path=/trunk/; revision=70269
Implement client-side only of console font APIs:
- documented: GetConsoleFontSize, GetCurrentConsoleFont;
- undocumented: GetConsoleFontInfo, GetNumberOfConsoleFonts, SetConsoleFont (see http://cboard.cprogramming.com/windows-programming/102187-console-font-size.html for a working usage example of these APIs).
(Also related to CORE-7931)
svn path=/trunk/; revision=70258
"add Everything Search Engine to ReactOS downloader" by Alexander Rechitskiy
(Silently remove some BOMs I created...)
CORE-8309 #resolve #comment Added, thx ^^
svn path=/trunk/; revision=70250
rapps-database does not update automatically because of broken build scripts by reactosfanboy
CORE-10633 #resolve #comment Fixed by removing the copies completely and updating the ones in rapps. Not the best solution, but it works for the update task.
svn path=/trunk/; revision=70249
Fix DLL control themeing by Jared Smudde
CORE-10563 #resolve #comment Hm, yeah, you are right. Somehow I really forgot that one when I did that manifest fun in the past ^^ Thx
svn path=/trunk/; revision=70245
Add/Update Traditional Chinese (zh-TW) translation in ReactOS. [Wave 2/4] by Henry Tang Ih
CORE-10615 #resolve #comment Committed, thx ^^
svn path=/trunk/; revision=70242
Update RosBE version to 2.1.3 by Hermes Belusca
CORE-10603 #resolve #comment Fixed. Does not work in ROS anyway, but fixed ^^
svn path=/trunk/; revision=70239
Missing italian translations in ReactOS by Carlo Bramini
CORE-9725 #comment Applied/Committed lautus and services patches
svn path=/trunk/; revision=70238
- Hackfix the existing charmap for the 0.4 release (just in case I don't get time to finish the new one). This fixes showing invalid glyphs and also removes the advanced section as it was unimplemented.
svn path=/trunk/; revision=70228
- Add support for canonical/string verbs when using CMINVOKECOMMANDINFO. This fixes switching views and adding a new folder when using the button in the common dialogs. This really needs to go in a CDefFolderMenu class (which likely inherits the context menu class), but we don't use that design.
- Partially implement CDefaultContextMenu::GetCommandString 'while I was there' It's completely untested though.
CORE-10089
svn path=/trunk/; revision=70215
Implement an arping tool. Some portions of code are just plain copy/paste from ping tool code (booh, it's bad!).
It was designed and tested on W2K3.
It's in RosApps for two major reasons: such a tool doesn't exist by default in Windows 2K3.
And it doesn't work in ReactOS.
svn path=/trunk/; revision=70212
Add a new apitest for iphlpapi.dll.
It includes tests for testing the SendARP() and its underlying IOCTL.
Tested against W2K3.
Obviously, most of these tests fail on ReactOS.
svn path=/trunk/; revision=70210
- Properly acquire process rundown
- Reference job object when assigning a process to it
- Don't assert when creating a child process whose parent is in a job
Fixes crashes in kernel32_winetest:process
svn path=/trunk/; revision=70177
- Don't cache IFilterMapper2 interface in FilterGraph2. Fixes use after free when playing videos with MPC HC.
- Fix a reference leak
CORE-7671
svn path=/trunk/; revision=70118
- Allow the dialog to be resized, which in turn will grow / shrink the grid view and the fonts inside it. This means users can make the grid larger if they don't have their glasses on or want to see all the fonts in more detail. We no longer need a large font window popping out from the grid when you click on a cell (which I was never a fan of).
- Don't allow scrolling previous to row 0 or after the last row
svn path=/trunk/; revision=70101
- Implement drawing fonts onto the gridview
- Implement scrolling through the grid via the scrollbar
- When a new font is selected, setup that font for drawing. This includes setting up the DC for painting, getting the valid glyphs for that font (this fixes the bug in the existing viewer which has lots of squares for glyphs), setting up scrollbar info, etc.
- Make the cell object more responsible for its own painting
CORE-10518
svn path=/trunk/; revision=70098
- Use the correct attributes when overwriting a read-only file in cabinet_copy_file. Will be sent upstream after the 1.8 release.
CORE-10534 #resolve
svn path=/trunk/; revision=70095
- Patch by Bruno Jesus : Don't add scroll bars to mdiclient if the window style doesn't allow it. This fixes wine bug 9435.
- Patch by Alex Henrie : Avoid double initialization. Use the window class icon if the MDI child doesn't have its own.
- Patch by Michael Stefaniuc : IDI_WINLOGO is already an int resource (PVS-Studio).
- Patch by Frédéric Delanoy : Use BOOL type where appropriate.
- Wine sync port.
svn path=/trunk/; revision=70068
- Return the unique ID from BaseSrvGetTempFile instead of an NTSTATUS. Fixes GetTempFile when using kernel32.dll from Windows.
svn path=/trunk/; revision=70038
GetDriveTypeW: Use neither CreateFileW nor InternalOpenDirW to open the drive, because both functions call SetLastError upon failure. This can cause failing checks in lines 23, 44 and 59 of the kernel32_apitest:GetDriveType test. Use NtOpenFile instead.
svn path=/trunk/; revision=70037
- Implement TCPSendIoctl() to issue IOCTLs to network stack. We should use global handles to \Device\Ip & such, but as iphlpapi doesn't have such feature yet, a hack is here to workaround.
- Implement SendARP() using the TCPSendIoctl() function. So far, it doesn't work, as tcpip.sys is missing appropriate code path.
svn path=/trunk/; revision=70025
-Enable some controls that shouldn't be disabled -> Spotted by Jared
-Yesterday I heard the whispered word "alphabet", correctly order file/fuction names
svn path=/trunk/; revision=70001
- Add document icon to wordpad.
[SHELL32]
- Add FriendlyTypeName support to HCR_MapTypeToValueW as well.
- Sync HCR_GetClassNameA ReactOS additions into HCR_GetClassNameW.
[NOTEPAD][SHIMGVW]
- Add localized FriendlyTypeName file type strings.
[SHELL32][HIVECLS]
Update the FriendlyTypeName file type strings.
This commit is an addendum to revision range r32966-r32969 and r32973.
svn path=/trunk/; revision=69999
- Patch by Alexandre Julliard : Add workaround for removed SC_TASKLIST system menu item. Word 95 assumes that the item exists.
- Wine sync port.
svn path=/trunk/; revision=69959
GetAdapters: Don't stop looking for the dns servers if an interface hasn't an entry.
This is not an error condtion, the loopback doesn't have one.
svn path=/trunk/; revision=69944
-Update some resource IDs to match with Win2k3
-Comment and reorder the resources for better overview
Part 3/3 CORE-10030
svn path=/trunk/; revision=69940
- Acquire the PEB lock before reading Peb->ProcessParameters->Environment in RtlQueryEnvironmentVariable_U. Patch by Mark Jansen.
CORE-10381
svn path=/trunk/; revision=69939
- Fix MB_GetString: the input parameter is the ID of the button string to be retrieved, ranging from 0 to 10 inclusive (this corresponds to the IDOK ... IDCONTINUE range, minus one).
- Add the missing "Close" button string.
svn path=/trunk/; revision=69937
MessageBoxIndirect fixes by Carlo Bramini:
- Implemented loading of text and caption if they are detected as resource IDs.
- Use direct resource string pointers with LoadStringW and use the returned string length.
- Dramatically improve the implementation of ID and resource string assignments, by using a little look up table. This removes some ugly, difficult to maintain copy-paste code.
- Fix the scaling of logical coordinates by making it aware of rounding, i.e. the size of the controls is now calculated correctly.
CORE-10352 #resolve #comment Thank you for the patch! :D
svn path=/trunk/; revision=69935
- Start to implement the charmap required for the 0.4 release
- We now have a basic dialog with the grid in place
CORE-10518
svn path=/trunk/; revision=69931
NetWkstaGetInfo: Retrieve the primary domain name instead of the account domain name as the workgroup/domain name. System CPL shows the proper workgroup name.
svn path=/trunk/; revision=69927
- Initialize DeviceRelations->Count in FdoQueryBusRelations
- Don't fake success for unsupported PNP minor functions, it makes the kernel very unhappy
Powered by Special Pool
svn path=/trunk/; revision=69900
- Addendum to r69867: free device info pointer at the end of acpi_bus_add to avoid use after free.
Powered by Special Pool
svn path=/trunk/; revision=69899
VfatGetAllInformation: Return as much information as possible, but do not fail if the name string does not fit into the buffer.
svn path=/trunk/; revision=69891
- Change SetDIBitsToDeviceFromClipboard to correctly use the requested format instead of always using CF_DIBV5
- Enable/Disable the menu point "Save as..." dynamically
- Remove unneeded import libraries
- Add WS_EX_ACCEPTFILES window style and handle WM_DROPFILES
- Add some defines and structures for handling clipboard files (work in progress)
- Add Open and Save dialogs
Patch by Ricardo Hanke. CORE-10520
svn path=/trunk/; revision=69890
- Add a test verifying that opening a file using different parent directories still yields handles to the same file
CORE-10483
svn path=/trunk/; revision=69877
- CDefView: When we have to flicker don't use the worst possible color to erase the window but the background color of the list view. At least now some flickering will go unnoticed.
svn path=/trunk/; revision=69873
- Remove the HPET hack, which is finally no longer needed (presumably thanks to the ACPI_ACQUIRE_GLOBAL_LOCK/ACPI_RELEASE_GLOBAL_LOCK fixes).
CORE-10509
svn path=/trunk/; revision=69872
- Update to version 20141107
- Get rid of the ROS diffs in actypes.h by making use of the new ACPI_USE_SYSTEM_INTTYPES define
svn path=/trunk/; revision=69866
- Fix GCC versions of ACPI_ACQUIRE_GLOBAL_LOCK/ACPI_RELEASE_GLOBAL_LOCK. It now generates the same code as the original
CORE-10509
svn path=/trunk/; revision=69863
- Apply the same fix as r69849, but for the Forward tooltip. Also it would be nice to use ATL strings to retrieve the resources & not hardcoding an extra NULL character inside the resource string.
- Add a missing string resource in romanian RC file.
svn path=/trunk/; revision=69850
-Fix blurry icons.
-Make the icons sharper and clearer in all size.
-Modified the design of some icons.
-Brand new Fax icons!
-Remove unneeded icon sizes
Things I have changed from the original patch:
-Fix 2 copypasta and 1 double definition errors
-Changed the icon back to the floppy one in folders.cpp and drvdefext.cpp because this would have resulted an a regression for floppy drives
svn path=/trunk/; revision=69847
- GetAllUsersProfileDirectoryW: Fail if lpcchSize is NULL.
- GetDefaultUserProfileDirectoryW: Fail if lpcchSize is NULL.
- GetUserProfileDirectoryW: Fail if hToken is NULL.
svn path=/trunk/; revision=69841
Start factoring out all code related to console-specific stuff (while leaving everything else in place). This is WIP (and the corresponding .c files are included where they are needed instead of being compiled, will be fixed later).
svn path=/trunk/; revision=69838
- GetProfilesDirectoryA: In case of success, return the length of the profiles directory.
- GetProfilesDirectoryW: Set ERROR_INSUFFICIENT_BUFFER, if lpProfilesDir is NULL.
svn path=/trunk/; revision=69837
CreateEnvironmentBlock: Set or overwrite the SystemRoot, SystemDrive and USERPROFILE environment variables of the new environment block.
svn path=/trunk/; revision=69834
- Simplify and restructure thread send messages, use look aside list, message signals and global message link for sorting trouble messages.
- See CORE-9210 for more information. Related to and fixed CORE-9718 and CORE-9975. Could have fixed CORE-9695.
svn path=/trunk/; revision=69793
- IEnumString::Next's pceltFetched parameter is not optional, so provide it. Fixes exception when using Emerge Desktop
svn path=/trunk/; revision=69792
Fix GetAdaptersAdresses operational status.
ReactOS is now seen online.
CORE-9178 #resolve
Please retest apps with connectivity problems.
svn path=/trunk/; revision=69791
Only scale characters with the world-to-device transform, when the graphics mode is GM_ADVANCED. Fixes Petzold's WhatSize testapp.
CORE-7154 #resolve
svn path=/trunk/; revision=69782
Adds a simple CLIP utility, as found on Windows, to ReactOS.
Like the Windows version, it can just redirect the output of another console program or a file to the clipboard.
By Ricardo Hanke, with minor corrections by me.
CORE-10411
svn path=/trunk/; revision=69775
Fix a nasty copypasta bug: When the DC brush was selected into a DC, it was painted with the pen color instead of the brush color. Should fix Snappy Driver installer and potentially a lot of other graphical issues.
CORE-9424 #resolve
svn path=/trunk/; revision=69774
- Return the appropriate status code when encountering file system corruption
- When encountering corruption, print a message by default instead of breaking into the debugger
svn path=/trunk/; revision=69768
- Reimplement the SYSTEM.INI/WIN.INI tabs (they should be seen as 2 different instances of the same window; this would be cleaner to do in C++) with all the expected functionality working.
- Add TreeView helpers for moving/deleting branches.
- Add TreeView helpers for 3-state checkboxes, with support for Windows 2k3 and Vista+ with and without manifest, depending on the comctl32.dll version used by msconfig at runtime. In particular, take advantage of the TVS_EX_PARTIALCHECKBOXES extended style introduced on Windows Vista+.
- Just initialize once the uxtheme function pointers.
- Create the msconfig dialog box hidden and center it on screen before showing it (removes the flickering caused by the quick move).
- Use the correct 16x16 icon for the dialog title icon, adapted from a patch by Jared Smudde with suggestions by Ismael Ferreras Morezuelas.
CORE-9333
svn path=/trunk/; revision=69761
Check the size of RLE bitmaps, while decompressing. Fixes possible buffer overrun.
Patch by Kamil Hornicek
CORE-8735 #resolve
svn path=/trunk/; revision=69760
Restore saved DCs, when changing the DC owner. Otherwise there could be saved DCs with references to process owned resources, causing GDI object leaks.
Fixes assertion when closing iWisoft Video Converter
Also don't copy lSaveDepth and hdcSafe in DC_vCopyState, those must be handled manually.
CORE-10435 #resolve
svn path=/trunk/; revision=69759
Fix COMCTL32_VERSION definition.
[COMCTL32]
The comctl32 we currently have in trunk cannot pretend to be version 6+ (it was currently reporting version 6.82, aka something for Windows 20? ) Fix it and make it report 5.82 instead.
CORE-10457
svn path=/trunk/; revision=69755
Comment out for now in MCB the assert hit when attempting to write on an extX volume.
Aleksey, could you have a look at it?
CORE-10407 #comment Can you retry with r69750?
svn path=/trunk/; revision=69750
Improve the oplocks stubs:
- Always return success when FsRtlCheckOplock() is called: we deny oplock creation at FsRtlOplockFsctrl(), so the oplocks can only be fine.
- Always return TRUE when FsRtlOplockIsFastIoPossible() is called: we deny oplock creation at FsRtlOplockFsctrl(), so no oplock can prevent fastio from happening.
This enables (for real :-)) write support in extX FSD. Not 100% perfect though, as it hits an assert in MCB (but that's yet another story).
Full implementation to come later on.
CORE-10407
CORE-10442
svn path=/trunk/; revision=69742
- Initialize SepRmDbLock and create the system and anonymous logon sessions in Phase 0, right before the system process token is created.
- Implement functions to reference and dereference a logon session.
- Reference a logon session in SepCreateToken and SepDuplicateToken.
- Dereference a logon session in SepDeleteToken.
svn path=/trunk/; revision=69735
- Do not always print the usage when an error occurs. Print it only when a syntactic error happens.
- FAT volume labels are constituted of 11 chars (not 8!) which should be printable ASCII (in particular, >= 0x20) and in uppercase.
svn path=/trunk/; revision=69732
- Properly parse SymbolicLinkName (in particular, don't assume it's null-terminated) in IoSetDeviceInterface. Fixes IoDeviceInterface test failures and subsequent crashes due to memory corruption
CORE-9456
svn path=/trunk/; revision=69729
- Explicitely use UNICODE versions of OpenSCManager and OpenService.
- When OpenSCManager or OpenService fail, they return a NULL handle (not INVALID_HANDLE_VALUE).
svn path=/trunk/; revision=69719
- Port the Services page in C++ and implement hide/show MS/ReactOS services & enable/disable. WIP.
- Fix resource ID for the "delete" button in the FreeLdr page.
- Revert the freeldr "advanced" dialog that I erroneously removed in r69638.
CORE-9333
svn path=/trunk/; revision=69717
- Move the message typedefs into a shared header file.
- Implement LsapRmCreateLogonSession and LsapRmDeleteLogonSession. Call LsapRmCreateLogonSession in LsapCreateLogonSession.
svn path=/trunk/; revision=69714
- Remove a few todo comments that are now completed. Also remove numbers to make it simpler to remove one entry from the todo list.
svn path=/trunk/; revision=69713
- Implement CompareIDs for the folders in the shell namespace. The previous implementation was a generic solution that just compared names.
- CDefView: Implement using CompareIDs method from the IShellFolder. Remove the previous implementation that was hardcoded in the CDefView and worked mostly for file items.
svn path=/trunk/; revision=69712
- co_UserCreateWindowEx: Fail when a wrong handle was passed for the parent window. "Fixes" a crash down the road due to missing null checks in other places (null checks are needed because desktop windows don't have parents).
svn path=/trunk/; revision=69709
- Fix exception specifier (Ged or Thomas, can you double-check those please? :) )
- ATL can be used in code not using the C++11 standard aka. fix build when double '>' characters are present, for example in template<foo bar<stuff>> .
svn path=/trunk/; revision=69702
- Flush the CPU instruction cache after the ATL window proc thunk code is patched into memory.
- Implement some (not all) CAtlList methods that I will need later on.
- Implement some (not all) CStringT and CSimpleStringT methods & operators that I will need later on.
svn path=/trunk/; revision=69701
- Add the SRM code from Timos kernel-fun branch.
- Move its lsasrv code to a separate file.
Thank you very much, Timo!
svn path=/trunk/; revision=69697
Import missing parameters for Ext2 driver.
This fixes extX volumes being only available in read-only on ReactOS.
CORE-10272 #comment Fixed with r69686
svn path=/trunk/; revision=69686
Fix NTDLL implementation of mbstowcs() and wcstombs() so that they return length in caracters and not in bytes.
This fixes last failing *to* CRT apitests
CORE-10390
svn path=/trunk/; revision=69683
Import _mbstowcs_l() from Wine and call it in mbstowcs().
This fixes crashes when calling the msvcrt, crtdll implementation of mbstowcs() with no output string.
Thus, it fixes a few crashing apitests
The NTDLL version is still broken in some way, need to investigate why.
CORE-10390 #resolve #comment Fixed with 69682. Thanks for the report!
svn path=/trunk/; revision=69682
- In UserAttachThreadInput, copy the CaretInfo structure instead of saving a pointer that may be freed right after
CORE-10395 #resolve
svn path=/trunk/; revision=69677
- Force all refreshes to go through the CDeviceManager method instead of calling CDeviceView directly.
- Fixes missing checked radio item in the view menu on startup
svn path=/trunk/; revision=69659
- When refreshing the view, reselect and expand the previously selected node in the tree. This allows to switch from 'by type' to 'by connection' and see where the selected device sits in the tree hierarchy (and vice versa)
svn path=/trunk/; revision=69658
- Automatically refresh the view whenever there is a hardware change on the machine
- Don't force a refresh when we enable/disable devices, this is now be picked up automatically
svn path=/trunk/; revision=69649
- Add DoDeleteAsync method that takes the data object and the fMask parameter from IContextMenu in order to distinguish if we need a permanent delete.
- CDefView: Properly add CMIC_MASK_SHIFT_DOWN and CMIC_MASK_CONTROL_DOWN flags when calling Invoke.
- CDefaultContextMenu: use DoDeleteAsync and pass fmask so as to make Shift+Delete shortcut to permanently delete files.
Note: GetKeyState sometimes returns wrong results but this doesn't concern the shell.
CORE-4365
svn path=/trunk/; revision=69647
- Fix pool leak in failure case (CID 1322140) and don't treat NTSTATUS values as Win32 errors in NtGdiGetGlyphIndicesW.
svn path=/trunk/; revision=69641
- Reenable the FREELDR.INI/BOOT.INI page. When boot items are added in the list, recompute its new contents width and adjust it so that we can horizontally scroll its contents (this is the only common control which does not support automatic width computation when you add contents in it...).
- Save the original FREELDR.INI/BOOT.INI in a default path (I took the one used by windows' msconfig for its INI files).
CORE-9333
svn path=/trunk/; revision=69638
- Implement the "file extract" dialog and enable it in the general page. The extraction functionality proper will be implemented later on.
- Remove an unneeded resource string.
CORE-9333
svn path=/trunk/; revision=69632
- Enable the specstrings.h in shlwapi header (since we have it)
- Add missing BIF_SHAREABLE define (which exists in the 2k3 PSDK), even if it's not used at all in our code (and in wine either); I use it in msconfig for a shell browse dialog (because I want it to behave as I expect also on windows).
svn path=/trunk/; revision=69631
- Add some file utility functions, they will be used later for managing the startup programs.
- Half-plement the general page controls.
- Simplify some code; fix file headers; remove the old toolspage.c (replaced by its c++ equivalent).
CORE-9333
svn path=/trunk/; revision=69630
- Sanitize allocation lifetime and fix memory leaks in ScmConvertToBootPathName. CID 1102363
- Fix leak in failure case of REnumServicesStatusExA. CID 716334
svn path=/trunk/; revision=69629
- Patch by Joachim Priesner : Hide horizontal Listbox scroll bar if no horizontal extent is set. See CORE-1101.
- Patch by Alex Henrie : Simplify if-else statement. Scroll listbox item completely into view on LB_SETCURSEL.
svn path=/trunk/; revision=69625
- Patch by Dmitry Timoshkov : Always call BeginPaint/EndPaint when handling a WM_PAINT message for the button.
- Patch by Alex Henrie : Remove always-false condition.
svn path=/trunk/; revision=69624
- Add missing shlobj.h changes from the previous commit
- Correctly define ITEMIDLIST types, and add initial support for STRICT_TYPED_ITEMIDS
CORE-8441 #resolve
svn path=/trunk/; revision=69621
- Fixup a few missing languages. Hopefully I got them all this time.
- Add a new line to the end of the resources to keep msvc happy
svn path=/trunk/; revision=69617
- Implement updating the device from the toolbar
- Add a message before uninstalling a device. Dedicated to Jared Smudde.
svn path=/trunk/; revision=69614
- GlobalMemoryStatusEx: Correctly compute ullTotalPageFile which should be in bytes, instead of in number of pages. Patch by contributor "kkat". CORE-10361
- GlobalMemoryStatusEx: Fail if the stored length in the data buffer is not what is expected by the API (required by the spec, see MSDN; on the contrary, GlobalMemoryStatus does not require that.)
- GlobalMemoryStatus: Correctly round up the reported total/available memory values, in case they are bigger than ~= 2GB.
svn path=/trunk/; revision=69608
- Fix duplicated resources.
- Implement the Tools tab. It reads the list of available tools (and their localized names & description) from an internal XML file. It can also load an external user-defined XML list of tools; see KB906569 for a description of what it is.
- XML stuff is implemented using COM. It is far easier to use COM with C++ than with C, hence the tools tab code and the XML parser are compiled as C++ whereas everything else is compiled in C. Therefore at the moment msconfig is hybrid C/C++; this is done using CMake magic.
CORE-9333
svn path=/trunk/; revision=69604
Don't keep spining forever in CcRosFlushDirtyPages() when flushing a dirty VACB fails on a RO volume.
This can be triggered with extX volumes and can prevent shutdown from succeeding
svn path=/trunk/; revision=69602
Modify the script to take into account that now, host-tools are no longer required.
Because it needs RosBE 2.1.2 (update guys!) still provide a fallback method to allow using old build method with host-tools. In case you need it, just call configure.sh with-host-tools
svn path=/trunk/; revision=69601
- Reserve some more space for ntdll so that we're compatible with Windows's version (base 0x7c910000). Should fix VMware Hybrid builder
svn path=/trunk/; revision=69587
Implement ExtX support in FS_REC, this allows it to autoload ext2fs.sys
Starting with this revision, browsing ExtX volumes in ReactOS works without any other modifications.
Thanks to Peter Hater for his initial work and to Thomas for his reviews
svn path=/trunk/; revision=69577
Mark (old) Linux partitions (notably extX volumes) as recognized by ReactOS in IsRecognizedPartition()
This is only enabled when building ReactOS.Otherwise, it matches Windows definition.
svn path=/trunk/; revision=69575
- Delete the old devmgmt and replace it with the new one.
- The new one is a temp container for devmgr.dll and simply calls DeviceManager_ExecuteW. Eventually mmc.exe will replace this module
svn path=/trunk/; revision=69547
- Kill off the old devmgr and replace it with the new one
- The new one now contains the (rewritten) device manager code, is now a C++ library and makes use of the recently improved ATL
svn path=/trunk/; revision=69545
Isolate KD a bit by replacing Rtl* memory routines with internal versions. Lets one put breakpoints inside kernel memcpy/memset without making KD very, very sad. Fix MmDbgCopyMemory to also not use RtlCopyMemory -- there is no need for it since it only performs 1-to-8 byte copies anyway. Minor fixes in the print/prompt routines.
svn path=/trunk/; revision=69539
Fix mapping of working set list page -- it should be local.
Get rid of MI_MAKE_LOCAL_PAGE and use ValidKernelPdeLocal/ValidKernelPteLocal consistently. This is what they exist for!
We should now hopefully not be creating any global pages when we don't want them.
Next step; fix flushing.
Bonus: Use the right lock in MiSessionLeader
svn path=/trunk/; revision=69532
Add super-complicated handling of global pages to KeFlushCurrentTb (pretty much the same code which has been in HalpFlushTLB for the past ~6 years). This should be all that is required to make this feature work (everything else being in place already), and *seems* to work fine but is disabled under a switch until tested thoroughly.
Global pages, an important optimization that allows for not flushing the whole x86 TLB every time CR3 is changed (typically on context switch to a new process, or during process attach/detach), relies on us doing extra work whenever we do alter a global page. This is likely where any bugs will have to be flushed out!
Fixup Ki386EnableGlobalPage while we are at it -- disable/restore interrupts properly, and verify PGE-bit isn't set (nothing should have touched it before this routine, which is responsible for initializing it, so we shouldn't have to disable it). Fix, but disable, the CPU-sync spin as well as there should be no particular reason to do this for PGE-enabling during initialization (no other processor will be messing with PTEs at this stage, as compared to a call to KeFlushEntireTb).
Everyone, repeat after me: Global pages are awesome!
svn path=/trunk/; revision=69528
Merge DbgBreakPointWithStatus and RtlpBreakWithStatusInstruction together as one function (the latter is just a label for KD), and add new macro necessary for this (MASM very much wants "::" on a global label inside a PROC local scope). Timo, you are awesome.
Bonus: Complement HEX()'s awesomeness with the other explicit radix specifiers.
svn path=/trunk/; revision=69515
- Increase height of the animation control in the file operation progress dialog to avoid cut-off. Patch by Jared Smudde.
CORE-10326 #resolve
svn path=/trunk/; revision=69510
- Pass the correct DeviceObject to PoRequestPowerIrp's callback
- Use the IRP stack to store callback parameters instead of a pool allocation as shown by the test
svn path=/trunk/; revision=69503
Add SHA-1 digests to RAPPS DB packages by Ismael Ferreras Morasuelas (swyter)
CORE-10319 #resolve #comment Errr... sorry, I was just told that we already verify downloaded stuff in rapps... committed and thx for the work you spent in it...
/me hides
svn path=/trunk/; revision=69498
- Separate system-defined from user-defined window properties. The values stored by win32k/user32 must not interfere with those stored by applications. Fixes various applications (e.g. PeaZip) randomly getting their window properties overwritten by the system
CORE-3897 #resolve
svn path=/trunk/; revision=69494
- Add the SystemLCID to the GlobalData variable and use it in the Advanced page.
- Add a stub for storing font linking settings.
CORE-10075
svn path=/trunk/; revision=69486
- Remove some DECLARE_RETURN()/RETURN()
- Add asserts for the user lock
- Annotate window property functions and use consistent parameter names
svn path=/trunk/; revision=69485
Fix a FIXME: properly reset timer when re-using a queued IRP for reading.
Also fix a possible race condition between time out DPC and IRP re-use.
Thanks to Thomas for his help with DPC & timers.
CORE-10245
svn path=/trunk/; revision=69483
Implement asynchronous reading from mailslot.
Patch by Nikita Pechenkin
Adjustements, style fixing by myself.
CORE-10245 #resolve #comment Modified patch committed with r69475. I have tested kernel32:mailslot and your provided test, both are passing 100%. Thanks!
svn path=/trunk/; revision=69475
- "Fix" SHChangeNotifyRegister to report file system changes. I am not really sure if these changes are correct but really make it work now.
CORE-1171
svn path=/trunk/; revision=69469
- Correctly handle NULL ppidl in EnumerateNext as required by Skip() callers
- Don't reenumerate NT directories every time EnumerateNext is called from Skip()
CORE-10311 #resolve
svn path=/trunk/; revision=69462
- Implement BiInitializeAndValidateHive, BiLoadHive. Leveraging our existing cmlib and the previously committed code, we are now able to read the BCD file, parse it as a registry hive, and lookup the BCD Objects subkey! The BCD Data Store is now available to the Boot Manager (which also validates that reading from ETFS works).
svn path=/trunk/; revision=69454
- The Objects key is in the root, not under Description.
[MKHIVE]:
- Make it *actually* build a BCD hive. It was doing nothing before, because it turns out that irrespective of what appear to be dynamic, command-line driven behavior modes, all of mkhive's logic is deeply tied together through C-level changes and knowledge. (for example, that's why calling it with a single .inf, it still wants to build 6 hives for you). There were also some copy/pasta bugs in the BCD handling code, now fixed.
svn path=/trunk/; revision=69453
- Fix multiple bugs in ETFS code (confusion between file offset vs. disk offset)
- Implement EtfsGetInformation, EtfsSetInformation, and fix ETFS_FILE definition to make this easy.
- Implement EtfsRead.
- Fix multiple bugs in file I/O code (swapped/reversed validation checks)
- Make BlStatusPrint call EfiPrintf on debug builds, even without BD.
- Add some additional error logging.
svn path=/trunk/; revision=69452
- Improve the internal locale storage. Replace the array of strings by a customized struct.
- Implement the locale customization properly. The customization property sheet modifies the internal locale storage rather than the registry. Now, cancelling changes works properly.
CORE-10074
svn path=/trunk/; revision=69449
- Whoever wrote mkhive ran into the problem I just did, and realized hey-i-dont-fucking-want-to-use-that-almost-copy-pasted-but-not-shared-code-so-why-dont-i-use-the-good-kernel-code-instead... but instead of taking the time to create a fucking library like I just did (cause that would be hard man, and fixing headers n' shit), just hard-coded the paths to the kernel files directly like a lazy shit. Fix mkhive to just rely on cmlib, which now has the files shared there.
svn path=/trunk/; revision=69445
- Move cmindex.c and cmname.c to cmlib, because bootmgr/winload use the kernel's cm implementation, unlike our freeldr which uses an almost copy-pasted-but-not-shared-because-hey-why-not-have-to-fix-the-bugs-in-two-places-instead-of-just-one implementation.
- Minor fixes to headers to allow those two files to be used/built outside of ntoskrnl.
svn path=/trunk/; revision=69444
Usability fixes (part 2):
- Properly keep the mouse show state across screenbuffer switches.
- Use unicode strings for the disk menu items.
See r69428.
svn path=/trunk/; revision=69431
Convert COLOR16 values to byte range before calling XLATEOBJ_iXlate.
This fixes gradient colors in some applications, as for example in Code::Blocks.
Patch by Vort. This is a fix for code written in revision r8117, so old !!
CORE-8652 #comment Committed in r69430, thanks guys!
svn path=/trunk/; revision=69430
Usability fixes:
- Update the menu each time the user chooses to show or hide the mouse pointer.
- Keep the mouse show state across screenbuffer switches (text/graphics) (half-hackish at the moment; will be elegantly fixed later on).
- Update the floppy disk menu items each time the user mounts / ejects a disk; display the disk image file name currently mounted (CHAR strings used at the moment, WCHAR conversion will follow soon).
svn path=/trunk/; revision=69428
Reading/writing to memory should #SS(0) when the segment is the stack segment,
and #GP(0) otherwise. It should never generate #NP.
svn path=/trunk/; revision=69427
- Temporarily hack the boot sequence to floppy, then hard disk. See r69421; will be improved in the future.
- Complete the BIOS 16-bit bootstrap code.
svn path=/trunk/; revision=69422
- Add a basic boot sequence functionality (read from CMOS); will be improved in the future.
- Print a "FATAL BOOT FAILURE" error message when INT 18h is called.
- Fail startup if we cannot mount the available hard disk images.
- Improve some diagnostic error messages.
svn path=/trunk/; revision=69421
- Don't forget to push the error code when the exception handler is a task gate.
- Use SS0/ESP0, SS1/ESP1, or SS2/ESP2 for ring 0/1/2 code task CALLs.
svn path=/trunk/; revision=69420
Properly implement support for IOCTL_DISK_GET_DRIVE_GEOMETRY, IOCTL_DISK_GET_DRIVE_GEOMETRY_EX, IOCTL_CDROM_GET_DRIVE_GEOMETRY_EX, IOCTL_DISK_GET_LENGTH_INFO in CdRomDeviceControl
CORE-10287 #resolve #comment Fixed with r69418
svn path=/trunk/; revision=69418
Don't bugcheck on unimplemented FsRtlOplockFsctrl(), FsRtlCurrentBatchOplock(), FsRtlCheckOplock()
Properly implement FsRtlInitializeOplock() (which is no-op on Windows 2k3)
CORE-10284 #resolve #comment Resolved with r69412
svn path=/trunk/; revision=69412
- Add floppy image automount support from the registry, as it was done for HDDs in r69390 (it's really copy-paste & adaptation of the existing code).
- Increase HDD number support from 1 to 4 (maximum) for INT 13h.
svn path=/trunk/; revision=69411
- Don't support IOCTL_DISK_GET_DRIVE_GEOMETRY_EX, IOCTL_CDROM_GET_DRIVE_GEOMETRY_EX for now
- Properly implement IRP completion for IOCTL_DISK_GET_LENGTH_INFO, IOCTL_DISK_GET_DRIVE_GEOMETRY
svn path=/trunk/; revision=69409
Fix a division-by-zero problem by storing the mouse driver's Width and Height
in DWORDs, thus avoiding the integer overflow which could make them zero.
svn path=/trunk/; revision=69408
- Always use the descriptor type, and never the limit, to identify
legacy vs. modern TSSes.
- Don't forget to write back the modified TSS link.
svn path=/trunk/; revision=69407
Read global settings from the register (in HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\NTVDM ; note that this is a ReactOS-only configuration key, because no equivalent functionality exists on windows), and currently initialize:
- the BIOS and ROM files to be used;
- the hard disks to be mounted at startup (they cannot be unmounted at runtime, contrary to the floppies) (maximum 4 hard disks).
svn path=/trunk/; revision=69390
Major bugfixing session for RtlGenerate8dot3Name()... Up to know, AllowExtendedCharacters was totally ignored and any upcased char, as long as it was not in the forbidden list was accepted. This was bringing numerous problems (cf: apitest :-)).
So, now, the following fixes were brought:
- AllowExtendedCharacters isn't ignored any longer. If it's given, the unicode char is upcased, if it's not, as previously, the ansi char is upcased
- Be more strict in the allowed chars in name: only allow "graph" chars. This avoids generating broken names that aren't allowed in FAT
- In case no char could be added in the name, then, for the generation of the name checksum and use it as shortname
- When writing the checksum, don't use reversed order, but direct order, to match Windows behavior (as exposed with tests)
- When writing the checksum, use when possible bit operations instead of numerical operations for performances reasons (NFC)
- Rewrite RtlpGetCheckSum() with an algorithm that produces the exact same checksums than Windows 2003.
This whole commit fixes all apitests related (direct visible effect). As a bonus, it also fixes short names generation in FAT: we no longer produce broken volumes with international file names.
This also fixes less visible issues: we were producing unicode strings with null char in their middle (as exposed in tests), not sure about how all functions could handle this: likely not that good, this could have lead to various memory issues & corruptions.
CORE-10223 #resolve #comment Fixed with r69389
svn path=/trunk/; revision=69389
Add more tests for RtlGenerate8dot3Name().
Not sure how many more are failing in ReactOS: they all pass in my WC (teasing :-))
CORE-10223
svn path=/trunk/; revision=69388
- Start implementing a user menu for mounting/unmounting floppy disks at runtime. Menu state refresh & proper localization remain to be done.
- Add a temporary "Sleep(INFINITE)" where the VdmShutdown function is susceptible to trigger again a VDM cleanup in case it was called again in parallel. For diagnosing CORE-10182, see also r69366.
svn path=/trunk/; revision=69375
- Don't duplicate the code for getting the name of a drive. CDrivesFolder already does it (and it does it better). Make the drive properties call it to get the name it needs.
CORE-10258
svn path=/trunk/; revision=69373
- Improve a change from r27016. Running the control panel applets in the context of explorer is not correct. We should really use ShellExecute here.
svn path=/trunk/; revision=69370
- Fix SH_ParseGuidDisplayName which was completely broken when I committed it.
- Use GUIDFromStringW instead of SHCLSIDFromStringW in _ILCreateGuidFromStrW. The latter needs the string to terminate right after the guid.
svn path=/trunk/; revision=69369
Add basic disk support with mounting/unmounting images in NTVDM:
- basic disk controller (at the moment this is just a collection of helper functions. A real HW emulation will come later on).
- INT 13h services for the BIOS.
At the moment, the images to be mounted are hardcoded in disk.c. Please see disk.c for examples of how to use the function. This will be reworked to allow user choice without having to recompile NTVDM.
CORE-10262 #resolve
svn path=/trunk/; revision=69365
Misc trap/interrupt fixes:
- Stub legacy KiGetTickCountHandler (INT 2A) to fail instead of breakpointing. We have made it so far without it, and if someone finds something that actually uses this instead of Win32's GetTickCount we can add it then. Raises the bar of making ReactOS breakpoint from 2 bytes to 4.
- Bugcheck with the right status code in case Kei386EoiHelper is called (which means you are trying to load a Windows HAL which we totally do not support right now).
- Use Ke* instead of Kf* in C code for IRQL routines.
svn path=/trunk/; revision=69362
- Don't hardcode magic values of characters if they have C escape codes.
- Print an extra newline on call to the Ctrl-C handler, as it is done on real DOS.
- Use a do{}while loop which loops while we don't have read a CR character.
svn path=/trunk/; revision=69361
DosReadFile echoes the line feed when it encounters it.
A backspace should delete the '^' sign for special characters.
svn path=/trunk/; revision=69360
Fix DOS character device I/O. Implement CON line buffering.
Make sure INT 21h functions 01h, 06h, 07h, 08h, 0Ah and 3Fh work as expected
for CON input.
svn path=/trunk/; revision=69356
[EX] Remove "KeGetCurrentProcess" and just use the EPROCESS Pcb member for type correctness instead of inventing a new macro to hide the casting.
[PS] Remove stray INIT_FUNCTION (doesn't do much on a prototype).
[IO] Define _IO_DEBUG_ to 0 again after someone forgot to leave it on 4 years ago (only bloated the code since IopTraceLevel was 0 anyway).
svn path=/trunk/; revision=69354
Add more tests for IoCreateFile(), this time purely dealing with symlinks reparse points.
They won't run on ReactOS (obviously) as we don't have RW NTFS support.
On Windows, they show interesting results....
With Windows 2003, creating the symlink is allowed and works, but then Windows is totally unable to deal with it and keeps complaining about an unhandled reparse tag.
The only way to open it is to open the reparse point itself. Not that working symlinks... Not sure which part is not able to handle the said reparse tag. NTFS? Io?
With Windows Vista+, it just works fine. Symlink is created & functionnal.
Broken logic!
svn path=/trunk/; revision=69353
Add a test for CORE-10188 (which was fixed by Thomas in r69236).
It is based on Nikita Pechenkin's patch with a few modifications by me to avoid race condition on start (and avoid flappy test) and to match more closely our coding style
ROSTESTS-190 #resolve #comment Committed in r69351. Thanks!
svn path=/trunk/; revision=69351
- Use ROS-style paths for include files.
- Fixup its CMakeFiles for C++11 compilation under GCC (done the way we did for other C++ projects).
- Add a VS project for DEVMGR to make Ged happy.
[ATL]
Start C++ fixups:
- Add 'placement new' operator.
- Add some missing functions (that were not caught by MSVC compilation because we atm. accept function usage without the existence of its actual code...).
More to follow...
svn path=/trunk/; revision=69348
[NTOSKRNL]: Add a helper function such that user-mode cannot pass certain kernel-only object attributes to ObOpenObjectByPointer.
[NTOSKRNL]: Make \Device\PhysicalMemory OBJ_KERNEL_EXCLUSIVE
[NTOSKRNL]: Deny access to user-mode when opening a handle to an object with OBJ_KERNEL_EXCLUSIVE
[NTOSKRNL]: Fix NtOpenProcess, NtOpenThread, NtOpenThreadTokenEx, NtOpenProcessTokenEx to use the helper function. Should fix Rtl using OBJ_KERNEL_HANDLE in user-mode.
Thanks to ThFabba for finding out these bugs, and the hbelusca for adding a missing check to NtOpenProcessTokenEx.
svn path=/trunk/; revision=69340
- Fix ZF and CF flags setting in function INT 2Fh, AX=1214h.
- Use CreateEventW explicitely.
- Set the VdmTaskEvent event at startup. This allows user input by default (even if we don't run our own DOS) in NTVDM. Fixes keyboard input not working when testing boot&usage of MSDOS 6.
svn path=/trunk/; revision=69339
Aside question: we have a "nmidebug" driver in \drivers\base\. In it is a NMI callback "NmiDbgCallback". I was wondering what "((void(*)())&KiBugCheckData[4])();" should do, according to the surrounding code, since in some conditions this code path is actually run: http://i.imgur.com/TUsEr5p.jpg
svn path=/trunk/; revision=69337
- Properly free handles in driver (spotted by Thomas)
- Remove dead code in driver (spotted by Thomas)
- Send an IOCTL to stop on symlinks to allow more precise tests
svn path=/trunk/; revision=69336
- Fix devmgmt.exe to correctly build and link against devmgr.dll
- This is a temp exe that just calls DeviceManager_ExecuteW. This will later be replaced by mmc.exe and devmgmt.msc
svn path=/trunk/; revision=69334
- Add support for CAtlString and all its supporting base classes.
- It doesn't do much more than assign wchat_t arrarys and call LoadString, but it's a full base to start adding more methods as people decide to.
svn path=/trunk/; revision=69332
- Add a few helper functions: SH_ParseGuidDisplayName, SHELL32_SetNameOfGuidItem, SHELL32_GetDetailsOfGuidItem.
- SHELL32_SetNameOfGuidItem is based on a patch for CORE-5995
- Use the helpers in CControlPanelFolder, CDesktopFolder and CDrivesFolder. Implements getting the comment and renaming registry items in these folders and share some more code.
svn path=/trunk/; revision=69328
[NTOSKRNL]: Actually apply the foreground priority boost to threads in a foreground process.
[NTOSKRNL]: Correctly handle the case where the quantum length and fixed/variable flag are set to "default", instead of falling back into the server case.
Thanks to WINSRV doing the right thing, ReactOS console apps now receive the priority separation boost. However, my tests show that it doesn't last as long as it should on Windows (quantums too short? bug in priority decrement?). Also, since Win32k.sys doesn't tell the kernel about foreground GUI apps, they don't get the boost. Someone needs to add a NtSetInformationProcess call in whatever win32k function determines active focus.
Dedicated to ThFabba who needs to write some tests ;-)
svn path=/trunk/; revision=69324
- Do not hardcode the number of XMS memory blocks that can map the HMA. It should be function of the MAX_ADDRESS value of our emulator.
- Require that MAX_ADDRESS should be greater or equal than 1 MB.
svn path=/trunk/; revision=69323
Extend the IoCreateFile tests with two parts tests.
First part properly checks for opening/reparse handling on mount points (and ideally on symlinks, see below).
The other part checks for proper handling of the IO_STOP_ON_SYMLINK parameter.
Regarding symlink, I'm clearly not convinced that what I did is correct (hence the failure on both W2K3 & ROS). It seems to me that symlink resolution is up to the FSD and not to IO which only handles mount point traversal.
If someone (Alex?) can confirm/infirm and give more insight, that's more than welcome.
svn path=/trunk/; revision=69322
- Build the properties as C++
- Rearrange the folder structure
- The whole thing now builds and runs from VS (with a few hacks). CMake next
svn path=/trunk/; revision=69320
- Add a helper HCR_GetClassName function to load the name of a reg folder to a STRRET
- Further simplify CAdminToolsFolder, CFontsFolder and CMyDocsFolder
svn path=/trunk/; revision=69319
Implement a save routine for the GeoID. Stores the current GeoID in the default user profile too, if required.
CORE-10172 #resolve
svn path=/trunk/; revision=69309
- Unicodify the CPL. Get rid of <tchar.h>.
- Pass a pointer to all getter and setter functions of the customization property sheet pages.
- Remove the "Property of" text from the property sheets.
- Convert "SendMessage(GetDlgItem(..." to "SendDlgItemMessage(...".
svn path=/trunk/; revision=69308
- Print a message if EISA DMA support is detected. Detection should probably not even be attempted unless the machine supports EISA; the print will show whether that's correct.
CORE-9895
svn path=/trunk/; revision=69302
Save and restore the general purpose and segment registers around Int32Call
calls in the BIOS. Also, make sure CF doesn't get set by the called interrupt
procedure.
svn path=/trunk/; revision=69297
- Add missing OBJ_KERNEL_HANDLE
- OBJ_KERNEL_HANDLE in user mode has no effect... no reason to make a define around it
CORE-10207
svn path=/trunk/; revision=69296
Fix CVE-2015-1769:
- When opening remote databases, don't attempt to migrate w2k database to current format.
- Also, when opening/creating the remote database, don't allow following symlinks anymore.
In case a symlink would be detected, an error message will be printed.
Commit dedicated to Alex :-).
CORE-10216 #resolve #comment Fixed with r69295
svn path=/trunk/; revision=69295
Our ReconcileThisDatabaseWithMasterWorker() routine will expect two parameters: the device extension + the device to which reconcile remote database.
So, fix our code accordingly
svn path=/trunk/; revision=69291
Implement the IOCTL_MOUNTMGR_QUERY_DOS_VOLUME_PATHS:
- Implement MountMgrQueryVolumePaths()
- Implement MountMgrValidateBackPointer()
- Implement MountMgrQueryDosVolumePaths()
- Rename a struct var to reflect its real usage
svn path=/trunk/; revision=69284
- Fix a activation drawing issue, if the window is not activated do not draw it active. Used testmenu program from CORE-1573 to test it. If this creates a new problem please open a Jira report.
svn path=/trunk/; revision=69263
- Support loading executables with the old EXE signature ('ZM' instead of 'MZ').
- Fix the EXE loader so that:
* we can load programs of >= 1Mb size, by exploiting the fact that on real DOS, the computed exe size (which uses the number of 512-byte blocks, stored in the EXE header) is stored in a WORD variable.
This has as an effect to truncate the value, and if it is such that the truncated value is less than the size of available memory, the EXE can be loaded (partly, of course). Some apps expect this behaviour,
for example, Turbo C++ IDE.
* Simplify the memory limits computation, according to http://www.tavi.co.uk/phobos/exeformat.html , which is the correct algorithm used by DOS (and confirmed by FreeDOS also).
svn path=/trunk/; revision=69261
Correctly call the device strategy/interrupt routines. Should fix a bunch of weird errors. This also fixes the NTVDM crash when Command&Conquer terminates, however it doesn't fix the DOS/4GW exception it encounters.
CORE-9678
svn path=/trunk/; revision=69246
- Use a NULL timeout for infinite waits instead of waiting for 100 ns. CORE-10188 #resolve
- Wait for available read data in user mode to handle thread termination
- Return STATUS_IO_TIMEOUT also for a zero-length timeout. Fixes Wine tests
- Avoid MmGetSystemAddressForMdl
- Acquiring a mutex is not a UserRequest
svn path=/trunk/; revision=69236
Fix the file header description. Consistently use .endm for closing a macro started with .macro.
[NTVDM]
- Add "_clean" targets to generated asm16 targets.
- Fix CMake dependency generation in a somewhat hackish way: I don't know how to cleanly specify the location of generated files for custom targets. And it is also hackish because in a CMake custom command which depends on the COMMAND.COM target, I also need to explicitely say that the command depends on the output file created by the COMMAND.COM target...
- Add COMMAND.COM to the CD files, even if NTVDM doesn't use an external command.com for now. CORE-5221
svn path=/trunk/; revision=69234
Implement the IOCTL IOCTL_MOUNTMGR_VOLUME_MOUNT_POINT_DELETED:
- Implement MountMgrVolumeMountPointDeleteed()
- Rename a struct var to reflect its real usage
svn path=/trunk/; revision=69232
- Fix memory leak in IopDetectResourceConflict. While we're at it use pool tagging, kernel handles and avoid INVALID_HANDLE_VALUE.
svn path=/trunk/; revision=69227
- Fix the GCC macro for structure declaration, so that referencing an offset to the structure doesn't fail anymore (ie. doing "mov dx, offset MyStruct" doesn't assemble to "mov dx, 0x100").
This fixes NTVDM not being able to start any DOS app. Reported by Aleksander Rechitskiy.
CORE-9677 CORE-9679 #resolve
CORE-9678
[NTVDM]
- Fix newline in DPRINT.
svn path=/trunk/; revision=69211
- Add support for DOS VDM reentry, i.e. the fact that a DOS app can start a 32-bit app, which in turn can start another DOS app, ad infinitum...
CORE-9711 CORE-9773 #resolve
- Add a small COMMAND.COM which is, for now, completely included in NTVDM. This COMMAND.COM is needed in order to support reentrancy. The fact that I chose to put it inside NTVDM is that any user can use instead his/her own COMMAND.COM, while retaining the possibility to perform VDM reentrancy (on Windows, if you remove the COMMAND.COM in windows\system32 and replace it with your own, you will break VDM reentrancy on windows' ntvdm).
CORE-5221 #resolve #comment I choose for the moment an internal COMMAND.COM, but you can recompile NTVDM to use it externally instead.
Global remarks:
- Quite a few DPRINTs were added for diagnostic purposes (since DOS reentrancy is a new feature), to be sure everything behaves as expected when being used with a large panel of applications. They will be removed when everything is OK.
- Support for current directories and 16/32-bit environment translation (in ntvdm + basevdm-side) remain to be implemented.
Other changes:
- Improve a bit the VDM shutdown code by gathering it at one place (there's still room for other improvements).
- Add suppport for properly pausing/resuming NTVDM.
- Bufferize some console events before dispatching.
Have fun ;^)
svn path=/trunk/; revision=69204
GetNextVDMCommand function:
- Avoid using deep nested levels of indentation.
- Fix the error check conditions (+ status value) for the CsrClientCallServer call to make it compatible with what windows' basesrv can return to us.
[BASESRV]
- Remove some useless "NTAPI" from functions.
- Factor out some code used for creating and destroying console records, and for the destruction of the pair of event handles.
- Use "IsListEmpty" when needed.
- Fix BaseSrvFillCommandInfo so that it returns the needed lengths (even if the user-given lengths are zero) and set the expected status error in case the buffer lenghts are too small (not STATUS_BUFFER_TOO_SMALL but STATUS_INVALID_PARAMETER because it's what kernel32 error check expects, for windows compat; see above).
- Fix initialization of new DOS records: no need to create a new one if there is one already free (status VDM_READY).
- Fix the loop that searches for a valid DOS record containing command information available as well as the conditions in which case our caller needs to wait in GetNextVDMCommand call.
NOTE: The internal state flags of the DOS records have normally nothing to do with the VDM state flags we report back to the user.
CORE-8247 CORE-9711 CORE-9773
svn path=/trunk/; revision=69201
- Read the current user locale when the CPL is created.
- Modify the user locale and update the examples when the user selects a new language.
- Save the user locale only if the user changed the settings.
CORE-10074
svn path=/trunk/; revision=69199
- Changing one "exit" to "exit /b" was not nice from me in the past.
ARM build now prepares host tools nicely again and you can build em. ARM build itself dies very early due to ..\..\boot\freeldr\freeldr\arch\arm\boot.S(9) : fatal error C1083: Cannot open include file: 'ksarm.h': No such file or directory
svn path=/trunk/; revision=69191
- Additional El Torito support. We now parse the EFI, BOOT directories, and find/open the BCD file, and get a handle to it!
svn path=/trunk/; revision=69189
- Disabling C4214 (nonstandard extension used : bit field types other than int), a somewhat overzealous warning for a "nonstandard" extension we rely on all over.
- Defining inactive DPRINT and friends for MSVC using __noop (no not __nop, __noop. See https://msdn.microsoft.com/en-us/library/s6btaxcs.aspx) to get at a whole bunch of C4127 (conditional expression is constant).
- Remove the last of the exported DriverEntry from dxg and dxgthk. There is no need in anyw way to export them -- dxg exports DxDdStartupDxGraphics to Win32k for initialization purposes and dxgthk is just a forwarder. Linking to a DriverEntry is to begin with is a pretty bad idea. Say goodbye to the last of LNK4216.
svn path=/trunk/; revision=69187
- Hey Arch, your buddies gave you ETFS support for native EFI boot, of course reading sector 0 is all zeroes, you're not on FAT, you idiot! ETFS has data at sector 16 and 17.
- Implement beginning of ETFS support. We correctly mount the reactOS boot cd as ETFS and open the root directory entry!
svn path=/trunk/; revision=69181
- Hackplement Layered Window Support from wine.
- See CORE-1576 for more information. Three days work, ReactOS needs more support for playing with GDI bits.
svn path=/trunk/; revision=69179
- Implement Block I/O read operations & related functions. Code needs cleanup, TBD.
- Unfortunately, I wish I could say it works, but it seems like all we get back are zeroes (from EFI itself). I suspect we are reading from the wrong device -- will have to investigate for a bit.
svn path=/trunk/; revision=69178
- Add some more reserved mapping tests for special pool tags and unaligned addresses
- Validate MDL pages in MmMdl test
svn path=/trunk/; revision=69169
- Add BCD creation.
- Add BCD to bootcd, and also to EFISYS.BIN. Verified the BCD is now present on the EFI partition.
svn path=/trunk/; revision=69156
- Fix more device I/O bugs.
- Silence some dbgprints.
- Implement DiskOpen and DiskClose.
- Refactor block cache in its own file.
- We successfully open our boot device (which appears as a floppy, as expected with the new UEFI boot)!
svn path=/trunk/; revision=69153
- Wrap some KDBG-only stuff into KDBG-define so we don't have KdSystemDebugControl being called needlessly on KD-builds.
svn path=/trunk/; revision=69148
- Reorder some more driver exports to alphabetical order.
- Remove NtGdiFlushUserBatch and NtUserQueryUserCounters from syscall list (unused, and not compatible).
- Move and add some debug-only syscalls to the end of the syscall list.
- Fix parameter count of UserValidateTimerCallback, UserValidateHandleSecure, and UserChangeDisplaySettings to match what is compatible, by removing unused or ignored parameters. NtGdiOpenDCW, NtUserEnumDisplayMonitors and NtUserSetCursorIconData still have the wrong numbers.
- Add "stray" NtGdiMakeObjectXferable and "NtDxEngGetRedirectionBitmap" syscall stubs.
- The syscalls should now be compatible with 2K3 SP2 in their order, but not in their argument count (or, most likely, their actual parameters). This means a Windows 2k3 gdi32 or user32 should crash somewhat later now that half of the syscalls' order is not in disarray.
svn path=/trunk/; revision=69147
[NDK] Add define guard for HalEndSystemInterrupt because it is no longer compatible with Windows, and add back the old definition. Fix NtSetSystemEnvironmentValueEx prototype.
[KERNEL32] Export SetFileCompletionNotificationModes (MSDN and headers documents it as Vista-only, but this is not entirely correct).
svn path=/trunk/; revision=69144
- 1.5KLOC code dump of current device block I/O handling. Totally untested and broken, just a checkpoint for now.
- Implement most of Hash Table Utility Functions.
- Implement parts of Block Allocator.
- Finish Implementation of Generic Table Utility Functions.
- Fix some EFI device emulation code.
- Temporarily disable freeing from heap while I figure out what's corrupting it.
svn path=/trunk/; revision=69142
* Allow customizing the 8-letter volume label from the FAT header.
* Make the efisys.bin have EFIBOOT as a label.
* Improve a bit the help text.
svn path=/trunk/; revision=69140
Thanks gigaherz for the "fatten" utility!, and others for testing.
[CDMAKE]
- Add multi-boot CD support, following El-Torito specification, such that we can the usual ISO boot sector on BIOS-based PCs, and the UEFI loader on UEFI-based PCs.
- Load segment should be stored in little endian.
- Fix the computation of the sector count (count in 512 byte sectors and rounded up).
- Rework the command-line options to make them more compatible with CDIMAGE / OSCDIMG.
CORE-10120
[BOOTDATA]
- Activate the UEFI boot support for our ISOs.
svn path=/trunk/; revision=69139
* Change the number of FAT copies stored by the formatting code to 2.
* Implement /BOOT command, to apply a boot sector to the image (FAT12/16 only, for now).
* Make use of the command above to finally get the generated efisys.bin loading in 7zip as a floppy.
svn path=/trunk/; revision=69135
In a failed attempt at getting 7-zip to recognize the fat images,
* Updated the fatfs backing library to v0.11a,
* Changed the default cluster size to 1 for floppy-sized images (< 4096 sectors) and 8 otherwise (4096 was an invalid value), and
* Changed the fatfs formatting code to generate FAT12 filesystems more windows-like.
I see no reason to discard the changes, so here they are.
svn path=/trunk/; revision=69131
Let's try not to leak memory...
So, close what's been opened in case of an error
And in a general way, close files we opened
CORE-10140 #resolve #comment Fixed with r69127
svn path=/trunk/; revision=69127
* Switch to using our existing host-tools header for types. Thanks Colin for pointing me to the right header (it was discrete so I missed it when looking).
svn path=/trunk/; revision=69112
* Fix folder creation and external file access.
* Add fatten as a host-tool.
* Add a target for creating efisys.bin from the bootmgfw
TODO: Make the name of the boot*.efi depend on the platform (ia32/x64/arm).
TODO: Add efisys as a dependency to the bootcd, when we need to make use of the efisys.bin file for the iso (waiting for hbelusca's cdmake work).
svn path=/trunk/; revision=69109
* Remove pointless dependency on BOOL and CHAR, which helps with portability.
* Disable reentrancy, since all the operations are serialized by the commandline.
* Improve a few comments.
svn path=/trunk/; revision=69106
* Fix the integer types for non-windows platforms (assume they have stdint.h instead of assuming embedded).
* Fix file I/O, which was just barely working enough to pass the one test I did.
svn path=/trunk/; revision=69105
- EarlyPrint -> EfiPrintf, as it seems BL actually has this after all :) Change all callers to conform to it.
- Add common exit/failure path in BmMain. Now the fun begins.
svn path=/trunk/; revision=69103
- Last round of display fixes. GOP works now -- fill out the frame buffer with gray as a test (tested in VirtualBox).
- Now to fix the heap.
svn path=/trunk/; revision=69097
- Add normaliz.dll to the build
- Move the appropriate functions from kernel32_vista to normaliz
- Import IsNormalizedString and NormalizeString stubs from Wine
Now Scilab starts
CORE-10124 #resolve
svn path=/trunk/; revision=69093
-State if there are no audio devices installed
-remove an obsolete resource ID
Patch by Jared Smudde CORE-10009 #resolve
svn path=/trunk/; revision=69088
- Generate XDK headers during build. To make changes to wdm.h, ntddk.h, ntifs.h, devioctl.h, ntdef.h or winnt.h, edit the appropriate files in include/xdk (as you should already have been doing for a long time).
CORE-9586 #resolve
svn path=/trunk/; revision=69083
- Fix text console bugs. It works now both when already at 80x25, as well as when using UEFI Shell to switch to 100x31 first. Tested with Virtual Box.
svn path=/trunk/; revision=69078
- Checkpoint commit of about 1000+ lines of text console support code. Right now, it's very disorganized and not tested.
- Write out the rest of the initialization path. Almost everything else is only for Boot Debugging, BitLocker, TPM, PXE...
svn path=/trunk/; revision=69076
- Fix the ASSERT in KeConnectInterrupt for debug single-processor builds (but keep the old one for multiprocessor build to make it less confusing if it is hit).
svn path=/trunk/; revision=69075
- I/O Library initialization.
- A few more random initialization routines. Skip TPM, BitLocker, Network and non-FAT filesystems for now.
svn path=/trunk/; revision=69073
- Fix some heap bugs. It (works?) now, at least for one allocation.
- Fix GCC support for Phase 1 Architectural Initialization.
- Implement Time measurement for Hyper-V compatible hypervisors (tested with Virtual Box 5.0!) as well as non-hypervisors (also tested)
svn path=/trunk/; revision=69071
- Add heap implementation, continuing from my work last year. Seems to be pretty broken, next steps are fixing it.
- Implement block allocator initialization (needs heap).
- Implement a few other missing initialization paths.
svn path=/trunk/; revision=69068
Change tabs to spaces, but only for the files I write. The rest come directly from the fatfs package and are best left untouched.
svn path=/trunk/; revision=69065
Introduce a tool I wrote some years ago, that helps create and manage a FAT disk image from the command-line.
It is built using parts of the isofs embedded library, with win32 bindings for file access. Should be helpful in order to create the EFI boot images needed for EFI-bootable isos.
Because the code is not really portable (yet), it's currently not included in the build process.
svn path=/trunk/; revision=69063
- Convert msconfig to a fully property-sheet-oriented app.
- Start converting the general page to a property page (serves as a model for the other pages).
- Sync translations for the general page.
- Temporarily remove the other pages from compilation.
svn path=/trunk/; revision=69062
- Implement case in BlMmRemoveRegionFromMdlEx where the descriptor's tail (or entire allocation) is contained in the descriptor -- this covers our BootMgr entry.
- Implement some more initialization functions.
- Start implementing the heap allocator initialization code. Now we need to implement the actual page allocator.
svn path=/trunk/; revision=69059
- Implement algorithm to detect if truncation is needed, which should only happen on IA64 systems I believe (otherwise, the OS and firmware should have the same alignment/page size requirements)
- Implement algorithm to detect if coalescing is needed. This shouldn't happen with firmware memory yet, but might happen once we start carving out stuff.
svn path=/trunk/; revision=69056
- DrvEnableDriver does not need to be exported. Fixes MSVC warnings (LNK4216: Exported entry point _DrvEnableDriver@12)
svn path=/trunk/; revision=69049
- Cleanup some bugs/issues in the memory map parsing code.
- Implement creation of BL/NT-compatible memory descriptors based on UEFI descriptors.
- Implement other remaining parts of the page allocator initialization routine.
- Last part missing is to implement routines for removing from a memory list, and for handling the boot manager's own descriptor.
svn path=/trunk/; revision=69047
- Use the right settings to make our CD UEFI compatible. The Internet says this is supposed to be 10000x harder, but it seems to work fine, at least in Virtual Box.
- Implement an EarlyPrint routine for early debugging/testing.
- Using the above, fix a bunch of stupid bugs. The previously committed code now seems to work (so far).
svn path=/trunk/; revision=69043
Add support for converting the EFI file path as well. The first 1000 lines of many have been written. Time to test on Virtual Box.
svn path=/trunk/; revision=69039
- Rename INIT_FUNCTION to INIT_SECTION and put it at the top of functions for consistency with __declspec declarations (it is actually possible to do this via __declspec in newer MSVC versions).
svn path=/trunk/; revision=69037
- Define some core boot library structures -- reverse engineered to be compatible with MS. This is important if we want to load MS' own binaries for testing (such as MemTest.Efi)
- Define basic BCD structures
- Start implementing EFI->Windows Boot Library conversion routines. Last part remaining is EFI path conversion.
- Fix linking with RTL, by providing some stubs for now.
svn path=/trunk/; revision=69033
- Use MmExpansionLock to synchronize changes to MmProcessList. Should fix random crashes in MmDeleteProcessAddressSpace.
CORE-10111 #resolve
svn path=/trunk/; revision=69031
- Associate a temporary VS project (will be removed later on); change the name of the target for now...
- Allow only one instance of MSConfig running on the system (we use a semaphore, compatible with MS' msconfig, so that if you try to run ROS msconfig on windows, you cannot run also windows' msconfig, and vice versa).
- Introduce some utility functions (for strings and other stuff), will be used later on.
CORE-9333
svn path=/trunk/; revision=69024
Fix rand_s implementation so that it doesn't leak resources, so that it doesn't dereference null pointer, so that it matches MSDN documentation (https://msdn.microsoft.com/fr-fr/library/sxtz2fa8.aspx)
CID 716561
CID 716668
svn path=/trunk/; revision=69005
Fix a nasty typo in QueryPointsFromMemory() which was like breaking the behavior of IOCTL_MOUNTMGR_QUERY_POINTS, IOCTL_MOUNTMGR_DELETE_POINTS
CID 1321853
svn path=/trunk/; revision=69002
Fix uninitialized address freeing in case of error in MupGetProviderInformation(), InitializeProvider()
CID 1322227
CID 1322240
svn path=/trunk/; revision=68990
- Don't try to free an uninitialized pointer in VfatSetRenameInformation. CID 1250303
- Actually use the dynamically calculated HashTableSize. CID 1237119/1237120/1237121
svn path=/trunk/; revision=68986
- Sort symbol entries with line number information before those without. Fixes backtraces showing ":0" as file/line for the first line of a function
svn path=/trunk/; revision=68980
Compile the THISCALL calling convention support only for x86. On x64, we are automatically in THISCALL.
Also I suggest we turn this file (and maybe also /lib/sdk/crt/except/i386/cpp.s) into a "msvc-thiscall.c" as it is already done in some other DLLs (like in ole32, riched20).
Fixes build in x64.
svn path=/trunk/; revision=68975
- Update all values in "HKCU\Control Panel\International" when a new language is selected.
- Update the language examples after a language change.
- Disable all settings that should not be affected by a language change.
Missing:
- Top-level windows do not get notification yet.
- Language settings customization should not work yet.
CORE-10074
svn path=/trunk/; revision=68966
- PspMapSystemDll and IopDisplayLoadingMessage are called from non-initialization code, so don't put them in .INIT. CmpHiveRootSecurityDescriptor is called from initialization code, but too late in boot, so don't discard it either.
- Trying to enable this for HAL as well leads to a weird crash very early in boot. Investigation is ongoing. Until then we will just have to keep the huge PCI Class/Vendor tables in memory...
- Enable it for Win32k though as it seems to work fine there.
- Remove INIT_FUNCTION from function prototypes, it serves no purpose there.
See issue 10097 for more details.
svn path=/trunk/; revision=68937
With the newly introduced mechanism (in r68829), on attributes list loop, there shouldn't be any AttributeList attribute anymore.
So, don't attempt to dump them.
svn path=/trunk/; revision=68933
Finally, use the correct implementation for RtlAssert().
In testbots, the behavior will be the following:
On prompt, sysreg will always issue a 'break once' instruction. In umode, where kdbg isn't instructed to handle breakpoint, this will lead to application being killed. With a lot of noise. This matches previous behavior.
In kmode, where kdbg handles breakpoints, sysreg2 will ask for a bt and will reboot the VM. As it was done previously.
The testbots already have the newest available sysreg2 revision.
This was made possible thanks to the help of Stefan and Thomas.
It was also made possible thanks to the commits:
- in trunk: r68899, r68905, r68917
- in project-tools: r2074, r2232
CORE-10088 #resolve #comment Fixed with r68918
svn path=/trunk/; revision=68918
- CNewMenu: Get the folder path from the pidl passed in the Initialize method. No need to query the defview for that.
- Dont use a reactos only interface to create a new directory. Prefer to do it with less code using exported functions.
- Add a new SelectNewItem method that is used when creating a new file or directory.
- Apply same fix for browse for folder dialog.
svn path=/trunk/; revision=68915
- Move NtFlushInstructionCache from sysinfo.c to virtual.c where it fits better. Likewise, move it from kefuncs to mmfuncs in NDK, and fix function arguments (ULONG -> SIZE_T).
- Re-enable TRAP_DEBUG, adding back critical checks in the trap code. Checks can be improved but it is better than potentially silently messing up system state.
- Move remaining RtlPrefetchMemoryNonTemporal code into kernel. Stubbed for non-x86.
- By Hermes suggestion, override ASSERT to NT_ASSERT only for MSVC builds as that is where the main benefit is.
svn path=/trunk/; revision=68907
Fix headers inclusion on MacOSX.
This fixes hhpcomp build on MacOSX
Patch by Nicolas Rouvière
CORE-10084 #resolve #comment Committed in r68906. Thanks!
svn path=/trunk/; revision=68906
Fix KdpPrompt():
- Don't attempt to return in the middle of the loop. It leads to keep the mouse disabled instead of reenabling it.
- Also, return a string length that actually matches what was written to the output buffer.
This fixes invalid behavior on RtlAssert when used with #if 0 moved to #if 1, or when Windows NTDLL is used in ReactOS.
CORE-10082 #resolve #comment Fixed with r68905
CORE-10083 #resolve #comment Fixed with r68905
svn path=/trunk/; revision=68905
- Move Menu to server side. See CORE-7797 and CORE-8299.
- This was for speed while moving windows about the desktop and fixed test results too.
svn path=/trunk/; revision=68904
- Remove the timer's list entry instead of the list head in NtSetTimer and NtCancelTimer. Fixes list entry corruption on VMware Hybrid builder
CORE-10081 #resolve
svn path=/trunk/; revision=68901
- Correctly reallocate buffer in NetWkstaUserGetInfo. Fixes one of the many exceptions when running netapi32 tests with DPH.
svn path=/trunk/; revision=68897
- Rename SHELL32_BindToChild to SHELL32_BindToFS. Do not let it be used for guid items any more. Split SHELL32_GetCLSIDForDirectoryout of it and call it only when needed.
- Fix callers to use SHELL32_BindToGuidItem for guid items.
- Fix a bug in CFSFolder which marked folder items as files when a binding context was used.
svn path=/trunk/; revision=68885
- All shell folders must have the IPersistFolder. Hence their Initialize method is supposed to succeed for them to actually work.
svn path=/trunk/; revision=68884
- Make SHELL32_GetDisplayNameOfChild return the result via a STRRET. Fix all callers accordingly.
- Avoid any heap allocations if a shell folder needs to forward the GetDisplayNameOf call to a child.
svn path=/trunk/; revision=68877
- Remove SHELL32_GetItemAttributes for good. It is never a good idea to have one function that handles many types of pidls.
svn path=/trunk/; revision=68873
- When binding to a drive item try to initialize the new CFSFolder using IPersistFolder3. That allows us to get the path of the drive right from the pidl instead of getting a new reference to the CDrivesFolder and using GetDisplayNameOf for the pidl.
svn path=/trunk/; revision=68865
- Fix swapped parameters of KeSetIntervalProfile
- Calculate the correct bucket pointer in KiParseProfileList, and actually increment the counter
CORE-10066
svn path=/trunk/; revision=68859
Improve debugging experience when pressing Tab+k, at least with WinDBG, by breaking with a suitable control code (adopt the SysRq one even if we don't use SysRq key for breaking, but instead, Tab+k).
svn path=/trunk/; revision=68855
Addendum to r68851:
- Reset the state of the worker thread *after* having captured the data, not before.
- Fix the explanations given in comments.
svn path=/trunk/; revision=68854
- Remove superfluous dereference in NtCreateProfile
- Allocate the right size for KPROFILE object
- Don't leak the KPROFILE when deleting an EPROFILE
- Avoid some casts
CORE-10066
svn path=/trunk/; revision=68852
Implement the WinDBG Debugger Worker Thread, which is created inside a target process whenever one uses the ".process /i <addr>" (invasive debugging) command using WinDBG in kernel-mode debugging.
See https://www.microsoftpressstore.com/articles/article.aspx?p=2201303&seqNum=2 section "Switching the Current Process Context" (from the book "Inside Windows Debugging" by Tarik Soulami, pp.101-102) for more details.
The names of the variables must be kept as they are, because WinDBG (more exactly, dbgeng.dll) searches for them within the kernel symbols and retrieves their addresses (since those variables are not publicly exported).
svn path=/trunk/; revision=68851
Make its version information compatible with the one of NTOSKRNL (and accessorily its resource file). Addendum to r25937.
svn path=/trunk/; revision=68850
- Only return the flags that the caller asked for in SHELL32_GetFSItemAttributes. Allows us to get rid of a hell lot of calls to CFSFolder::BindToChild which lead to many calls to FindFirstFile.
svn path=/trunk/; revision=68849
Fix x64 compilation:
- Some KDBG symbols for win32k debugging are defined only in x86 platform when _WINKD_ is disabled.
- KdSetOwedBreakpoints is defined only if _WINKD_ is enabled, addendum for x64 to r68842.
svn path=/trunk/; revision=68848
- Avoid statically importing FindFirstStreamW/FindNextStreamW so that our cmd.exe can work on WinXP.
- Accept lowercase /r switch
svn path=/trunk/; revision=68847
- In DirList, initialize stInfo.ptrHead of the dummy node to avoid invalid memory access. Fixes regression in cmd:batch
svn path=/trunk/; revision=68846
- Sync control.c from wine and disable a large part of it that is not needed. Also update the definition of CPlApplet to match wine.
- Make CControlPanel use the new definition of CPlApplet which lets us simplify loading the applets.
- Make CControlPanel to only use unicode and use the new helper functions.
- Simplify the code for the context menu of cpl items.
svn path=/trunk/; revision=68845
- Fix the condition check when setting twice (or more) the same breakpoint.
- Implement support for deferred breakpoints. For more information, see: http://www.osronline.com/article.cfm?article=541 (which also exposes an interesting problem about them).
svn path=/trunk/; revision=68841
It seems that our disk.sys returns "unknown" media type when performing IOCTL_DISK_GET_DRIVE_GEOMETRY on an extended partition.
Not sure how Windows handles this (if any volunteer for a test :-)), in the meantime, forcibly sector size to 512B to allow NTFS driver opening volumes on extended partitions
svn path=/trunk/; revision=68835
The desktop.ini file is only meant to be read if the folder has the R (read-only) or S (system) attributes.
If this causes your favorite folder to stop showing the icon, use "attrib +r <foldername>" or edit it from the properties dialog.
When we add the desktop.ini files, we will also need to add the R attribute to their respective folders.
Followup to CORE-9002
svn path=/trunk/; revision=68834
Totally rewrite the way MFT records attributes are handled.
Up to now, we were having really similar loops, only looking at the resident part of the attribute list, not really caring about how the loop was going.
This was leading to some issues:
- In case the attribute we were looking for was stored in the non-resident part of the attribute list, we would miss it (excepted in the case of FindAttribute() which was properly browsing the whole attribute list).
- In the specific case of FindAttribute(), one would have been able to setup a broken MFT record with the resident attribute list pointing on the non resident attribute list which itself would point to the resident attribute list. In such case, the driver would loop forever caught on the loop, allocating tones of memory. It was possible to trigger this by user space, from a non-privileged user, just by browsing the right directory entry.
- In the case of the other loops (non FindAttribute()), another issue (other than missing attributes) was present, one would have been able to setup a broken MFT record with an attribute of null-length. This would have caused the driver to loop forever on the attribute list. This could be triggered from usermode too. And could be triggered by a non-privileged user.
This commit introduces a new set of functions for attributes browsing: FindFirstAttribute(), FindNextAttribute(), FindCloseAttribute(). It allows safely browsing attributes and handles broken cases. It also performs reading of the attribute list when present and makes sure there's only one read. This method should be the only one to use to browse the attributes.
The whole NTFS code base has been converted to use this newly set of functions. This really simplifies the implementation of FindAttribute(), and prevent unsafe code duplication.
CORE-10037 #resolve #comment Fixed with r68829
svn path=/trunk/; revision=68829
* Fix loading icon info from dekstop.ini. Also took the chance to remove some unnecessary convolution. The old code tried to load the info, but didn't actually make use of the returned string at all.
CORE-9002 #resolve #comment Icon loading should work now. Adding the default desktop.ini files in the right folders will be a followup issue.
svn path=/trunk/; revision=68828
- Towards a complete fix of VGA console attach/detach: a lot of progress is made but few little things need to be checked/adjusted.
- Fix text scroll (at least in text modes) by doing guest-to-guest moves, using the EmulatorCopyMemory function introduced in r68826.
- Fix the computation of the screen rows/columns values stored in the BDA, in graphics modes (in addition to the change of r68090), by dividing the graphics resolution by the character heights & widths. Correct values are indeed needed by some programs, eg. QBasic (I also add a CharacterWidth; normally it should be computable using the CRTC registers, and is always == 8 or 9, but still...).
svn path=/trunk/; revision=68827
Introduce a function for doing guest-to-guest memory copies (given here in a temporary implementation until a proper one is developed). It will be used in my next commit.
svn path=/trunk/; revision=68826
Properly set StreamNameLength (with the length of extra data we append).
This fixes truncated output from FindFirstStreamW/FindNextStreamW
svn path=/trunk/; revision=68822
- Correctly sync ConsoleFramebuffer when we attach/detach VGA from the console (fix memory corruptions introduced by r68720).
- Fix few comments in the process.
svn path=/trunk/; revision=68812
- Move some #defines where they belong; use ARRAYSIZE(...) macro where appropriate.
- Use enumeration for the environment name types.
- Fix BaseCreateVDMEnvironment so that we only convert the paths of TEMP/TMP, PATH, WINDIR and SYSTEMROOT env values to their short form, as done on Windows. To that purpose use the BaseGetEnvNameType_U helper.
- Fix the loop that computed the needed new environment size.
Needed for CORE-9773.
svn path=/trunk/; revision=68806
Simplify the hackssign application usage.
You don't need anymore to specify which VM type you use. The client will attempt to discover it by looking at available shared folders providers.
If it recognizes one, it will perform the assignation.
No Guest Additions/Tools installed will make the tool refuse to assign a drive letter (obviously).
So basically, you just need to issue some: hackssign assign z ReactOS (for example)
svn path=/trunk/; revision=68802
Stack hacks!
Let's introduce hackssign. This application (and its associated driver) are here to allow
users to assign drive letter to their VMware/VBox shared folders.
It relies on two components: a client to communicate the instructions and a driver to execute such instructions.
Do not execute this application nor its driver outside ReactOS. There are barely no security checks nor sanety checks.
You could do substantial damages.
So, let's go back to shared folders state in ReactOS nowadays...
VMware
------
Configure your shared folders
Install VMware Tools
Reboot
From here, you can access your share from cmd or explorer with UNC path (for instance: \\vmware-host\Shared Folders\ReactOS)
Go to C:\ReactOS\bin (or whatever, depending on your setup)
Execute hackssign assign z ReactOS vmware (for instance)
z is the letter you want to give, and ReactOS is the name of your share (you gave it in VMware configuration)
You're done, you can access your share with the Z drive letter
VirtualBox
----------
Configure your shared folders
Install VBox Guest Additions
Reboot
Go to C:\ReactOS\bin (or whatever, depending on your setup)
Execute rosvboxmgmt start
From here, you can access your share from cmd or explorer with UNC path (for instance: \\vboxsvr\ReactOS)
Execute hackssign assign z ReactOS virtualbox (for instance)
z is the letter you want to give, and ReactOS is the name of your share (you gave it in VMware configuration)
You're done, you can access your share with the Z drive letter
Note that you can delete an assignement by performing a hackssign delete z (or whatever drive you want to remove).
In such case, the share remains available with its UNC path.
svn path=/trunk/; revision=68794
[OLEAUT32]
[DDK]
[PSDK]
[STLPORT]
[CPPRT]
Port the vs2015 fixes to also work on x64
[HHCOMP]
Fix compilation with VS in x64 mode
svn path=/trunk/; revision=68791
- Hackplement support for parsing UNC paths. We are not able to enumerate UNC shares yet. However until it is properly implemented we will be able to enter a UNC patch in the addressbar or start explorer with a path that points to a UNC share.
svn path=/trunk/; revision=68787
* Use lazy enumeration of the nt objects and registry keys, as intended for the IEnumIDList interface.
* Display of the NT Object creation date has been lost in the process, but it was only really used by symlinks.
* Symlink target display doesn't appear to work at the moment, will fix later.
svn path=/trunk/; revision=68783
Add the "rosvboxmgmt" tool.
It's purpose is to allow to interact with the VBoxSF driver as we don't have working MPR yet.
In order to have UNC path available, all you need to do so far is rosvboxmgmt start. All the rest isn't mandatory.
svn path=/trunk/; revision=68778
Implement the TokenImpersonationLevel case of SeQueryInformationToken().
This fixes VBoxSF driver not being able to expose shared folders in ReactOS
svn path=/trunk/; revision=68777
- Add support for specifying a parent window handle for the screensaver configuration dialog, adapted from patch by Timo Kreuzer, see CORE-5718.
- Fix failure return values from the helper functions.
- Add TranslateMessage call in the message pump.
- Get the correct left/top coordinates for the screensaver, being multi-screen aware.
- Use Win-compatible window names & styles for the screensaver preview dialog & screensaver window, compatible MS' scrnsave.lib.
svn path=/trunk/; revision=68776
- Move the properties dialog for the recycle bin in its own file. The implementation of the recycle bin is overly complicated and keeping with it a dialog for editing some registry entries makes it worse.
svn path=/trunk/; revision=68775
- Create a new sublibrary for shell32 that will contain the implementation of CDesktopBrowser and desktop dde support. Link it to both shell32 and rshell. Also add a win8+ hack.
svn path=/trunk/; revision=68769
- Take an extra reference on CDefView during the lifetime of the view window, and correctly handle recursive WM_DESTROY messages
CORE-9932
svn path=/trunk/; revision=68761
Rewrite FsRtlRegisterUncProvider() and FsRtlDeregisterUncProvider() so that they can support MUP in case it's required
They also support DFS
svn path=/trunk/; revision=68759
Changes in preparation for CORE-9773
- Split DosLoadExecutable into DosLoadExecutable & DosLoadExecutableInternal, the latter one can be used for small programs embedded in ntvdm.exe (will be used for integrated command.com).
- Factorize common loading code in DosLoadExecutable(Internal); just before starting a program, set the CPU registers to the default values DOS use; see comments in the code for more details (some programs expect those values).
- Set a correct termination type (hiword in the returned ErrorLevel) when terminating programs (the value is different for TSRs and for regular programs).
CORE-9773 CORE-9729 CORE-9711 CORE-8247
svn path=/trunk/; revision=68746
May thou rest in peace!
(...or rather live on in MSPAINT_NEW's commit history)
[MSPAINT_NEW]
I hereby pronounce thee MSPAINT!
svn path=/trunk/; revision=68741
- CDesktopFolder: Create two interal fs folders that wiill eventually take care of all file support for the desktop folder.
- Implement some methods using the internal fs folders.
svn path=/trunk/; revision=68736
Add/Update Chinese (zh-CN) translation by Henry Tang Ih
CORE-9003 #resolve #comment Committed v7 Patch, as the author told me, this resolves this report. Thx once more for translating ROS to Chinese ^^
svn path=/trunk/; revision=68732
- Use SHELL32_GetGuidItemAttributes, SHELL32_GetFSItemAttributes, SHELL32_BindToGuidItem and SHELL32_GetDisplayNameOfGUIDItem in CControlPanelFolder, CDrivesFolder and CFSFolder instead of using functions that accept all sorts of pidl types.
svn path=/trunk/; revision=68730
- Split SHELL32_GetItemAttributes to SHELL32_GetGuidItemAttributes and SHELL32_GetFSItemAttributes.
- Add a couple helper functions that will reduce code duplication in folders with reg items: SHELL32_BindToGuidItem, SHELL32_GetDisplayNameOfGUIDItem
svn path=/trunk/; revision=68729
- CAdminToolsFolder: return proper attributes when no items are specified. (This is fine for reg folders but doesn't work for fs folders).
- CFontsFolder: fix GetAttributesOf as above. Fix Initialize method to return success. All shell folders should support this method.
- CNetFolder: Return the name of the folder when asked.
- CPrinterFolder: It doesn't have subfolders so it doesn't make sense to implement BindToObject.
svn path=/trunk/; revision=68728
Change console processes priority when the console gains or loses focus. FIXME: Win2k3 backtraces show that we should call NtUserSetInformationProcess(unimplemented in ROS) to set some win32 process foreground flags.
CORE-9911
svn path=/trunk/; revision=68716
- Don't use hardcoded structure types in sizeof's where possible.
- Fix the parameters of NtSetInformationProcess calls in CsrSetBackground/ForegroundPriority.
CORE-9911
svn path=/trunk/; revision=68715
- When we have few R/W IO port handlers, just register them directly instead of using a wrap-up function.
- Remove hardcoded bit flag values.
svn path=/trunk/; revision=68710
- Add some missing MouseResetCounters calls when enabling/disabling mouse reporting, and an ACKnowledge signal when we set mouse defaults.
- When we have few R/W IO port handlers, just register them directly instead of using a wrap-up function.
svn path=/trunk/; revision=68709
- When we have few R/W IO port handlers, just register them directly instead of using a wrap-up function.
- Start implementing the PS/2 keyboard command handler.
- Report scancodes only if the keyboard is enabled.
See r68415.
CORE-9920
svn path=/trunk/; revision=68708
Make sure to attach to the specified process before dereferencing ClientInfo, which is a user mode structure.
CORE-l0017 #resolve
svn path=/trunk/; revision=68702
There will come the day... the day I will find you, inventer of the crappy BOM b***sh*t and you will suffer. Same goes to MS notepad using that crap!
svn path=/trunk/; revision=68688
- Ignore IMAGE_SCN_TYPE_NOLOAD when loading image sections, as shown by ntdll_apitest:NtMapViewOfSection. This avoids errors when trying to relocate .rossym sections (since they were not loaded but Ldr tried to change their protection)
CORE-8384
svn path=/trunk/; revision=68678
- When paging in data for an image section, first flush outstanding cached data since paging reads are noncached. Fixes failure to create section in ntdll_apitest:NtMapViewOfSection/Test_NoLoadSection
- Gracefully handle out-of-pool situation
CORE-8384
svn path=/trunk/; revision=68677
New programs for rapps by Petr Akhlamov
CORE-9152 #comment Partial sync skipping DFend and QBittorrent AND WinFF because there is no direct link mirror. Additionally I updated the stuff to more recent versions AND fixed the D/L links....
svn path=/trunk/; revision=68657
Adding New Apps for rapps and Removing Some Unneedful Spaces in aklepad.txt by Henry Tang Ih
CORE-9953 #comment, Patch committed incl. German translation and some fixes. Still missing: Unlocker and GNU Octave
svn path=/trunk/; revision=68655
- Use newer implementation of RtlActivateActivationContextUnsafeFast(): its time finally has come (see r52615 and r52612). It returns proper return value instead of pushing STATUS_SUCCESS into PRTL_ACTIVATION_CONTEXT_STACK_FRAME type and its "Trying to activate improper activation context" warning pops up all the time (its disabled in trunk to prevent debug log spam).
svn path=/trunk/; revision=68654
Adding New Apps for rapps and Removing Some Unneedful Spaces in aklepad.txt by Erdem Ersoy
CORE-9953 #comment Added Agat, Audacity, Gens, HexEdit, Shed, UltraISO and VBA-SDL because of the prove in image form that it works. Rest ist still not committed.
+ Removed the UTF BOMs again!
+ Added German Translations
+ Corrected the Coding to UTF-8
svn path=/trunk/; revision=68648
Change Trash to Recycle Bin by Lee Schroeder
CORE-8391 #resolve #comment Committed the big one with changes to all RC files, because it's the right way to do that. All these strings are originated from the english resource and NEVER were translated up to now, so we keep these strings in sync. Second argument. Trash is not the right word here (YES, a non native dares to say this ^^) We use the Windows way here, unless someone gives me a REALLY good reason not to ^^
svn path=/trunk/; revision=68645
- In AccpOpenNamedObject, immediately save the last error in case of OpenService failure instead of calling another function in between.
CORE-9990 #resolve
svn path=/trunk/; revision=68636
Add winspy++ to rapps by Stefano Toncich
CORE-9659 #resolve #comment Committed with swyter's suggestions and a german translation included.
svn path=/trunk/; revision=68635
- Enable relocation debug prints again. While it's not a critical error anymore when they happen it's still critical for debugging issues.
svn path=/trunk/; revision=68626
-include the testdata in all targets
-load it from the correct path even when the application is using a different working directory
svn path=/trunk/; revision=68623
VfatWrite (but not VfatRead) is supposed to return STATUS_INVALID_USER_BUFFER
if an error occurred while locking the buffer.
[KERNEL32]
Reading 0 bytes using ReadFile doesn't always succeed on Windows.
svn path=/trunk/; revision=68620
In CcCopyData, Buffer can be NULL during read/write operations.
[FASTFAT]
Use SEH to catch exceptions thrown by MmProbeAndLockPages.
Lock the user buffer before changing the file allocation size.
svn path=/trunk/; revision=68619
- Load sfc_os.dll because applications have a tendency to create remote threads in winlogon with sfc_os function pointers, assuming that it is loaded. See https://bitsum.com/aboutwfp.asp
CORE-9598 #resolve
svn path=/trunk/; revision=68615
PS2:
- By default all the PS/2 ports are disabled. They become enabled by the BIOS at the POST step.
- Similarly it is the BIOS POST that sets up the PS/2 controller configuration byte.
- Synchronize the value of bit 2 "System flag" and bit 4 "Keyboard enable flag" in the status register, according to what is set in the controller configuration register. What is the "keyboard enable flag" ? See http://www.os2museum.com/wp/the-dos-4-0-shell-mouse-mystery/ for more details...
HW MOUSE:
- Resetting the mouse sends also an ACKnowledge byte too...
BIOS32:
- Fix the reported number of bytes in the BIOS configuration table.
- Enable the PS/2 ports in the POST.
- Implement the "Pointing Device BIOS Interface" INT 15h, AH=C2h, AL=00h...09h based on VBox OSE & SeaBIOS; we should make our PS/2 mouse driver use it.
The real call to INT 15h, AH=C2h function is still disabled because our mouse driver doesn't react well with it, when we run some applications like MS Diagnostics.
PS2 MOUSE DRV:
- Update copyright notice;
- Remove 2 useless functions;
- I've put in comments in the code the places where calls to the BIOS ps/2 mouse interface are needed.
svn path=/trunk/; revision=68613
Irp->UserBuffer being NULL doesn't indicate any error. It could be that the
caller really wants the result stored at address NULL (which can be valid,
and is valid by default for programs like NTVDM).
svn path=/trunk/; revision=68607
- Copy-paste error fix in the initialization of the collating table.
- Fix DosCombineFreeBlocks so that it doesn't say "DOS memory arena corrupted" because while combining blocks, the current block became the last one and we tried to merge it with a next block which, by definition, cannot exist (see comment in the code).
- Remove a spurious call to DosCombineFreeBlocks in DosResizeMemory which was forgotten in r68596.
- Check the validity of the encountered MCBs during the loop for linking or unlinking the UMB chain to the main arena chain.
- Fix the size of the created UMB blocks.
svn path=/trunk/; revision=68606
- Use a static buffer for read or writes of "few" bytes (and allocate one for larger operations).
- Writes of zero length, in DOS, truncates or extends the file to the current position of the file pointer. This fixes some "corruption" in some files (so far, only .INI files) in Windows 3.11 installation.
svn path=/trunk/; revision=68604
- Implement OEM bitmaps sizes for server information. If more is need, it will be added. This will be plugged in later.
svn path=/trunk/; revision=68603
- Implement function 4Dh "Get All Handles Number of Pages", 54h "Get Handle Directory / Search for Named Handle / Get Total Number of Handles"
- Fix the return error of few functions.
- Start handling the "system handle" (special handle 0; see the specs for more details); WIP.
svn path=/trunk/; revision=68601
- Increase the maximum number of EMS handles to the maximum permitted by the specification (ie. 255 handles)
- Zero out EMS handles names.
- EMS status are always returned in the AH register!
- Implement functions 4Bh "Get Number of Opened Handles" and 4Ch "Get Handle Number of Pages"
- Remove an hardcoded value.
- Cosmetic change: rename EMS_STATUS_OK to EMS_STATUS_SUCCESS to comply with the name convention used also in XMS, etc...
svn path=/trunk/; revision=68600
- Add some level of "Advanced debugging" (see ntvdm.h) which one can adjust to enable/disable debugging features inside NTVDM (this can be useful as long as NTVDM is under heavy bugfixing. When it will be more perfect, this stuff will be removed).
- Add the possibility to load option ROMs at a given segment. Currently their list should be specified from inside ntvdm.c (in the BiosInitialize call), but I plan to make it available from a registry option (or via command-line for NTVDM-standalone mode).
- Start to separate the initialization of "static" BIOS data (stuff that lives in ROM space) and initialization of "dynamic" BIOS data (eg. initializing the interrupt vector table, the BIOS data area at segment 40h, ...) so that we may be able to reuse part of our code to be able to more-or-less run external (16-bit) BIOSes, or the Windows NTVDM BIOS that uses BOPs to run some of its stuff in ntvdm in 32-bit (as we do for our 32-bit BIOS, except that *all* of our bios is 32-bit, not just some parts). Also, some file reorganization will be in order there soon...
- Add video BIOS version information in memory so that tools such as Microsoft Diagnostics can correctly recognize our video BIOS (btw, we try to emulate the functionality of Cirrus' CL-GD5434).
- Correctly put video BIOS ROM header (+ checksum) in memory so that it is recognized as such by diagnostics tools.
- During BIOS POST, scan for ROMs starting segment 0xC000 (where video ROMs reside).
- Store statically the BIOS configuration table.
- Fix INT 16h, AH=12h "Get extended shift states" so that it correctly returns the state of right Ctrl and Alt keys.
- Fix bit-setting state; report that our keyboard is 101/102 enhanced keyboard.
- Correctly set the error return values (AH=86h and CF set) when a function of INT 15h is unsupported.
- Implement INT 15h, AH=C9h "Get CPU Type and Mask Revision"; INT 1Ah, AH=02h "Get Real-Time Clock Time" and Ah=04h "Get Real-Time Clock Date" by reading the CMOS.
- Implement CMOS century register support.
svn path=/trunk/; revision=68598
Remove broken check. Broken cases are already handled by upper if case.
This fixes the regression that leaded some valid ISOs not to be read anymore by ReactOS
CORE-9254
svn path=/trunk/; revision=68595
- Don't accept 64 bit PE files on x86 (will cause use of uninitialized variable ImageBase). CORE-9955
- Remove a pointless check in MmMapViewOfSection
svn path=/trunk/; revision=68593
Fix build on MSVC by rewriting the code in *STANDARD* C++ !! (and not into some strange idiom called "GCC-C++"). I suggest also to write a proper class for dynamically-allocated (resource) strings instead of either having the static arrays of hardcoded sizes, or being tempted to use non-standard constructs as the one I just saw.
(and btw, instead of defining a new "SIZEOF()" macro, there is one which already exists in the PSDK called "ARRAYSIZE()" which just does the correct job).
svn path=/trunk/; revision=68587
EMS:
- Introduce (and use) helpers for checking validity of EMS handles.
- Do not hardcode the EMS page frame segment, but instead allow it to be changed (for now, it is still set to a default value; in the future, via some configuration file or via the registry). This is done by acquiring a UMB block (see after).
XMS:
- Implement functions 0x01 "Request HMA" and 0x02 "Release HMA".
- Fix the return value of functions 0x04 "Global disable A20" and 0x08 "Query free Extended Memory"; simplify code of function 0x0B "Move EMB".
- Halfplement function 0x0F "Reallocate Extended Memory Block" in the simple case of size reduction (size expansion is left to the programmer as an exercise :PP )
- Rewrite the UMB provider support (functions 0x10, 0x11, 0x12) by calling the Upper Memory Area manager helpers (see after) (this is closer to reality: UMBs are either provided by XMS driver itself, or by an EMS driver which hooks into the XMS driver chain -- as it is done with MS' himem+EMM386; sometimes all that stuff is contained inside one driver only --) instead of calling back into DOS. This is the DOS which calls XMS for getting the UMB blocks and initializing them! (and not the other way around as it was done in r68001!).
NTVDM:
- Introduce an "Upper Memory Area manager" which maintains a list of areas of upper memory (>= A000:0000 and <= FFFF:000F) that can be used as RAM blocks. It is intended to work closely with the NTVDM memory manager and be used by XMS for getting possible free UMBs, and by VDDs for implementing the VDDInclude/ExcludeMem APIs (which adds/remove blocks in/from the UMB pool; those are unaccessible to DOS if those APIs are called after NTVDM have been started, but are accessible by XMS).
DOS:
- Add a helper function for detecting early DOS arena corruptions (for debugging purposes only).
- Make the DOS memory manager really UMB-compatible. This means:
* not hardcoding the start of the UMB chain;
* getting all the available UMB blocks from XMS and initializing them, marking the reserved blocks as read-only (with a correct header; reserved blocks are eg. VGA memory area, ROM blocks...).
There is room for improvements obviously (see the FIXMEs in the code). Used documentation is mentioned in comments in the code.
This commit should fix quite some apps, as well as it fixes corruptions of loaded ROMs in upper memory: that's how I came into working on fixing the UMB support. In other words, during those two last weeks, I was like in: http://i.imgur.com/zoWpqEB.gifv
CORE-9969 #resolve
svn path=/trunk/; revision=68586
- Remove some hardcoded values.
- Reshuffle again DOS initialization to better reflect how it is done in the real world (see "Advanced MS-DOS Programming" by Ray Duncan, Chapter 2 "MS-DOS in Operation"): split DOS data stuff into DOS BIOS data, and DOS kernel data areas.
In preparation for an upcoming commit.
svn path=/trunk/; revision=68585
- use optimal buffer size to load null terminated strings using LoadString
- add printing-related menu entries and dialogs (no printing functionality yet)
svn path=/trunk/; revision=68584
* Rework the WIN7_COMPAT_MODE into actually just disabling the incompatible functions, and introduce WIN7_DEBUG_MODE which does the rest of the "hacks" I used to ease development.
svn path=/trunk/; revision=68580
- When directly submitting the results to the web service, make sure to URL-encode the test log, or & signs will terminate the log and make Testman see the test as crashed (and have all other kinds of possible consequences)
svn path=/trunk/; revision=68578
- Skip user32:msg_paint test_paint_messages if running on Windows in non-interactive mode. This should allow WHS-Testbot to get a bit further
ROSTESTS-185
ONLINE-441 ROSTESTS-184
svn path=/trunk/; revision=68575
- Add support for intsafe signed Add functions
- Guard signed math in ENABLE_INTSAFE_SIGNED_FUNCTIONS
CORE-9947 #resolve
svn path=/trunk/; revision=68573
NtQuerySystemInformation:
- Windows Vista and later: Check the information class before probing the buffer.
- Pre-Vista Windows: Check the information class after probing the buffer.
- Pre-Vista Windows: Fixed the aligment check for invalid information classes.
- Pre-Vista Windows: The SystemFlagsInformation and SystemKernelDebuggerInformation classes return no required length in case of an error.
This fixes all ntdll_apitest:NtSystemInformation fails.
svn path=/trunk/; revision=68571
- move bitmap and DC initialization to ImageModel
- remove unused global variables
- remove copy-pasta bug from SelectionModel::DrawSelection
svn path=/trunk/; revision=68565
Show shutdown dialog at the logon screen if the user wants to shutdown. Patch by Jared Smudde.
CORE-9919 #resolve #comment Thanks a lot!
svn path=/trunk/; revision=68562
Add a case for VSSolution generation with VS14. It DOES NOT WORK with the cmake version we use currently, but it's better to have the error rather than generating ninja files without warning.
svn path=/trunk/; revision=68551
- Make <cmath> functions available in top-level namespace. Fixes stlport build with VS2015.
- Remove unused __cmath_power and __pow_helper
svn path=/trunk/; revision=68485
- Add ellipsis to the caption text if the window size is reduced. See CORE-9474, it can use more tweaks.
- Also after all this, DefWndNCPaint can soon be moved into Win32k, draw.c in already in. This could help with caption redrawing while moving the window.
svn path=/trunk/; revision=68476
- Improve painting messages, separated functions and dual functions for isolating drawing issues. Next thing to do is fix Set Window Position; Still no SWP_FRAMECHANGED support.
- See CORE-7447 for test results.
svn path=/trunk/; revision=68474
- Use the partition list code from usetup. Scan all harddisks at startup and keep the list alive as long as diskpart is running.
- Use the partiton list in the 'select disk' and 'list disk' commands.
- Implement the 'list disk' and 'list partition' commmands.
- Add required strings.
svn path=/trunk/; revision=68473
- Skip ole32:dragdrop test_DoDragDrop if running on Windows in non-interactive mode. This should allow WHS-Testbot to get a bit further
ROSTESTS-182
ONLINE-441 ROSTESTS-181
svn path=/trunk/; revision=68441
- Remove an unused but set variable. Fixes compilation with FEATURE_UNIX_FILENAME_COMPLETION defined. Patch by Lee Schroeder.
CORE-9913 #resolve
svn path=/trunk/; revision=68420
- Move some #defines into their correct headers.
- Fix some DPRINTs.
- Add missing DOS SYSVars values (memory-related), confirmed by DOS undocumented & FreeDOS.
- Implement INT 21h, AX=5700h and 5701h "Get/Set File last-written date & time".
- Implement INT 2Fh, AH=13h "Set Disk Interrupt Handler" chain support, which is an obscure functionality (to make story short: allows inserting disk filter drivers): see the comments inside the code (dos.c/dos.h) for more information.
svn path=/trunk/; revision=68418
We don't support any PS/2 keyboard command for now. Just return ACKnowledge. This unblocks some programs that want to initialize the keyboard by sending keyboard commands and then performing polling on the port until "valid" data comes out. (unblocks e.g. DOS Mugen game, amongst others...).
svn path=/trunk/; revision=68415
Pasting a larger-than-current-canvas image now DOES resize the canvas.
(adapted from a patch by Gian Sass)
CORE-9674 #resolve #comment Fixed in r68398
svn path=/trunk/; revision=68398
- Whitespace fixes.
- Improve some DPRINTs.
- hardcoded_values--;
- Fix DisplayMessage() for NULL-terminated strings.
- Free the allocated block of memory in the mouse driver cleanup procedure.
- Properly inherit the parent process environment block if needed.
- Return the correct last error if DosCopyEnvironmentBlock fails.
- Big ifs turned into switch.
- Implement some INT 2Fh DOS multiplex functions (AX=1200h "Installation check", AX=1203h "Get DOS data segment", AX=1214h "Compare FAR pointers", AX=122Fh "Set DOS version to return"); remove the DPRINT for AX=1680h "Release Current Virtual Machine Time-slice".
- Stubplement INT 2Ah "DOS critical sections / network" (it just monitors calls to itself at the moment).
- Use the boot drive root path as the current directory if we failed retrieving a valid one.
- Offsetize the DOS_SYSVARS members (so that I don't have to compute by head the offsets, especially when some of the members is a structure from which I don't know in advance its size); add some C_ASSERTs on the size of DOS standard structures.
- Move the NullDriverRoutine buffer outside the sysvars buffer (because it has nothing to do inside it).
- In ConDrvReadInput: do not echo the first part of an extended character.
- Add basic support for changing the reported DosVersion.
svn path=/trunk/; revision=68395
Enable the 'Edit...' and 'Remove' buttons on the Advanced IP Settings dialog only if the 'IP addresses' and 'Dafault gateways' lists are not empty.
svn path=/trunk/; revision=68392
Show an error message if formatting a partition fails.
Based on patch by mna.
CORE-7969 #resolve #comment Thanks a lot!
svn path=/trunk/; revision=68388
Since CsrClientCallServer returns the value of ApiMessage.Status, then just use ApiMessage.Status instead of declaring another Status variable for querying the return value of CsrClientCallServer (continuation of r62460).
svn path=/trunk/; revision=68384
- Add the IAtlMemMgr interface
- Add a basic implementation of CWin32Heap which just implements the base IAtlMemMgr
- Required for CAtlString / CAtlStringMgr (coming later)
svn path=/trunk/; revision=68381
- Use the correct HINSTANCE in the dll. Fixes loading strings for tooltips, etc.
- Rename CMainWindow to CDeviceManager
- Fixup the decelerations and add missing SAL decoration
svn path=/trunk/; revision=68378
- Add few remaining members to the SYSVARS list.
- Implement retrieving the (DOS) boot drive (by using the host boot drive letter from %SYSTEMDRIVE%).
Should fix Win3.11 and MSD 2.0+ boot drive retrieval.
svn path=/trunk/; revision=68377
- Rearrange the file structure, move the properties/resource/misc stuff to a ./properties folder
- Add a CMakeLists.txt file to combine both old and new code. Should now build with the msvc RosBe (doesn't yet run)
svn path=/trunk/; revision=68371
I know this *_new thing is a bit overused, but there are a lot of changes coming to devmgr and rather than take it out of the build for a few days, it's easier to do it this way
svn path=/trunk/; revision=68369
- last few fixes to the device manager code to get it building with rosbe (msvc)
- Next steps are to merge it with the existing devmgr code (mostly done in my WC) and get it running via mmc.exe & devmgmt.msc
svn path=/trunk/; revision=68367
Implement basic DOS country info (some stuff is not implemented yet, but the basic functionality is there). This is needed by Microsoft Diagnostics 2.0+ (amongst others)...
See also r66039.
CORE-9903 #resolve
svn path=/trunk/; revision=68363
- Unambiguously specify top-level namespace for Win32 API functions in preparation for upcoming ATL changes. Based on a patch by Benedikt Freisen.
CORE-9675
svn path=/trunk/; revision=68359
Make the keyboard layouts selection not dependent on the selected language.
Patch by joao.jeronimo.89.
CORE-9630 #resolve #comment Thanks a lot!
svn path=/trunk/; revision=68354
- Show that PAGE_NOCACHE also works for image and data sections
- Show that *MapViewOfSection returns user space addresses
CORE-9808
svn path=/trunk/; revision=68352
Continue streams integration:
- NtfsGetFileSize() will now return the file size depending on the open data stream
- NtfsGetFileSize() will also optionally return allocated size
- In NtfsMakeFCBFromDirEntry(), when initializing FSRTL_COMMON_FCB_HEADER entry, use data stream sizes
svn path=/trunk/; revision=68348
- Implement INT 10h, AH=13h "Write String" function.
- Fix VidBiosPrintCharacter to use the colour attribute only in graphics mode, or when explicitely needed.
- Fix some special cases where specifying a BIOS mode page of 0xFF means "use current page".
- Use a macro to check whether a given mode number specifies a text mode, and use it where needed.
svn path=/trunk/; revision=68347
- Add a separate class for the root node
- Move creating the action menu items to a generic function so we can use it for both the context menu and the main menu
- Create the main action menu as a dynamic popup menu. It doesn't work quite the same as the menus via the resources, so if any menu expects know how to fix this, be my guest
svn path=/trunk/; revision=68342
- Add description of the 'L' key to the partition selection page.
- Make the 'ENTER' key work for logical partitions too. Selected unpartitioned space will be converted to a logical partition.
Patch by Gerhard Gruber (sparhawk).
CORE-9641 #resolve
svn path=/trunk/; revision=68339
- Implement dynamic context menu for device and class nodes
- Fix CanUninstall to return correct results
- Strings hardcoded for now, will fix soon
svn path=/trunk/; revision=68334
Fix user buffer handling for FSCTL_GET_NTFS_VOLUME_DATA. It fixes NTFSInfo.
We're just lacking a few information in the output data now
CORE-8725
svn path=/trunk/; revision=68327
Time to free some Sysinternals tool: ntfsinfo.
To make it short, this tool dumps various information about a NTFS volume and its reserved meta-data files
Our version comes with three advantages compared to the Russinovich's tool:
- It's FLOSS
- It works properly on NT5+ (developed with W2K3 & W7) whereas R's cannot display meta-data files information
- It will open a volume by default if none provided
One issue so far: it doesn't work properly on ReactOS!
CORE-8725
svn path=/trunk/; revision=68326
- Skip msi:msi test if running on Windows in non-interactive mode. This should allow WHS-Testbot to get a bit further
ROSTESTS-180
ONLINE-441 ROSTESTS-179
svn path=/trunk/; revision=68319
Add CmpDestroySecurityCache() and CmpDestroyHiveViewList() stubs and call them in CmpDestroyHive and CmUnloadKey().
CORE-6492 #resolve #comment Thank you Hermes! This is exactly what I needed!
svn path=/trunk/; revision=68313
Validate the stream we're trying to open properly exists.
This check might have to be moved later on when the driver switches to RW
svn path=/trunk/; revision=68309
Always gather the size of the unnamed stream for directory display.
This fixes file size display when there are several data streams available
svn path=/trunk/; revision=68308
- Do not forget to initialize LoadDll->NamePointer in DbgkMapViewOfSection!!
- Just copy the NamePointer (pointer to pointer to file name) in DbgUiConvertStateChangeStructure. See the description of the lpImageName member of the LOAD_DLL_DEBUG_INFO structure in https://msdn.microsoft.com/en-us/library/windows/desktop/ms680351(v=vs.85).aspx for more details.
This fixes some debugging stuff with GDB (see the below-mentioned report).
Adapted from a patch by andy-123 (whom I don't remember his name^^),
CORE-7019 #resolve #comment Fixed in r68306
CORE-8622 #comment An updated fix was committed in r68306
svn path=/trunk/; revision=68306
HvFree:
- Free the BaseBlock.
- Do not free the hive itself because the HHIVE is part of the CMHIVE structure and freeing it here could be an unpleasant surprise.
[NTOSKRNL]
CmpDestroyHive: Free the CMHIVE after calling HvFree.
CmUnloadKey: Free the hive storage, the hive locks and the CMHIVE properly.
svn path=/trunk/; revision=68303
Implement support for NTFS $DATA streams:
- The driver is now able to read various streams for a same file, using the same syntax as Windows.
- This fixes to read (in general) files with multiple streams where reading unnamed stream was leading to read beyond file end
- Also fix reading small files which are smaller than a sector
For demo, see: http://www.heisspiter.net/~Pierre/rostests/NTFS_Streams.png
svn path=/trunk/; revision=68302
While handling IRP_MN_QUERY_DIRECTORY, in case there is no $FILE_NAME attribute, dump file attributes & MFT ID.
This fixes ASSERT failure when browsing some Win7 volumes
svn path=/trunk/; revision=68290
Don't attempt to open the unnamed stream from $DATA attribute on file lookup. It was unused & leaking.
It fixes opening files that don't have unnamed stream (even though we cannot read there contents yet as we only support unnamed stream)
svn path=/trunk/; revision=68285
- Add a stub implementation for IActiveIMMMessagePumpOwner.
- Use #ifdef __REACTOS__ to mark a ROS-diff
CORE-4953 #resolve
svn path=/trunk/; revision=68275
- Mark the page as being on the standby list before dereferencing it in MmFreePagesFromMdl. Since the PFN is marked as deleted, MiDecrementReferenceCount will correctly insert it into the free list, but this is required to satisfy MiDecrementReferenceCount's sanity check.
CORE-9877 #resolve
svn path=/trunk/; revision=68273
- Add support for selecting any device when refreshing the view. This allows us to re-select a device after it's been enabled instead of collapsing the tree and losing track of what you did.
- Plug some memory leaks
- HeapAlloc -> new
svn path=/trunk/; revision=68272
Don't attempt to read the unnamed stream of the $DATA attribute of each file matching:
- We don't need such attribute later on
- In case there's no such unammed stream, finding the file fails, whereas the file really exist
This fixes (at least here) browsing Win7 NTFS system drives, where our driver was always failing on pagefile.sys
svn path=/trunk/; revision=68268
- Add basic functionality for enabling and disabling devices. Make it accessible via the toolbar and main menu
- Implement checking whether a device has a problem and if a device can be uninstalled
- Implement getting, setting and removing device flags in the install params
- Fix creating and destroying the (currently empty) context menu
svn path=/trunk/; revision=68265
- ReAllocateLayoutBuffer: Write all new partition table entries to the disk.
- UpdateDiskLayout:
- Create link entries for logical partitions.
- Wipe unused logical partiton table entries.
svn path=/trunk/; revision=68257
- Fix a GCC warning (error on Macs) concerning the printf "%Iu" specifier.
- To make sure data included by bin2c, that need to be interpreted as a (long) string, is correctly NULL-terminated, introduce an output format called "STR": this creates a long character array using the "\xXX" format.
- To circumvent a silly limitation of MSVC (all versions?) about the maximum length (65535 bytes) of a char string (aka. char MyArray[] = "...my_long_string...";), introduced yet another output format called "BINSTR" which creates an array of bytes that is NULL-terminated (aka. char MyArray[] = {0xde, 0xad, 0xbe, 0xef, 0x00};).
- Finally, the output format "BIN" is the regular bin2c format (simple array of bytes).
[HAL_PCIDATA][FREELDR_INSTALL]
Use the above-described functionality (for HAL_PCIDATA, use BINSTR format because the PciVendors data is otherwise too large to be stored in "char_string" format).
svn path=/trunk/; revision=68255
- Reset FunctionDescriptorCount to zero in USBCCGP_LegacyEnum since the function can be called multiple times
CORE-9226
svn path=/trunk/; revision=68250
- Flush cout after each write as already done by Colin in r66855, per ROSTESTS-158
- Use DbgPrint instead of OutputDebugStringA again because the latter only calls the former anyway
Fixes test summary lines not being recognized by Testman, especially on VMware.
A little summary so we don't keep going back and forth with this function:
1) Only writing complete lines is required to that the output doesn't mix with debug output from other components. See r55618
2) OutputDebugStringA splits its input into 512-byte-sized blocks with no regard for line breaks, so using it with strings larger than 512 bytes breaks (1).
3) OutputDebugStringA eventually calls DbgPrint("%s", string) anyway so using it with chunks smaller than 512 bytes is not an optimization
As a result, yes this function MUST split up the lines itself, this can't be optimized or simplified away! kthxbye
ROSTESTS-178 #resolve
svn path=/trunk/; revision=68246
In case of directory enumeration, validate the record earlier to really prevent any potentiel buffer overflow
CORE-9254
svn path=/trunk/; revision=68244
Revamp r68233:
- Don't duplicate code, implement checks in a helper function
- When checking name content, do it earlier for better performances
- Add extra checks to prevent a potential buffer overflow in case of Joliet names with illformed entries
CORE-9254
svn path=/trunk/; revision=68239
UpdateDiskLayout:
- Resize the layout buffer if necessary.
- Start writing logical partitions to the disk.
- Disable the code that wipes unused table entries. May be enabled later.
Right now, you can only create one logical partition because the link entries are not created yet.
svn path=/trunk/; revision=68235
Add a few sanity checks to our handling of ISO 9660 / Joliet:
- Null name entries are forbidden
- Degenerated entries (mistmaching sizes) reveal broken layout
In both cases, the lookup for a file is aborted and the CD considered as corrupted by the FSD.
explorer displays empty CDs then.
CORE-9254 #resolve #comment Fixed with r68233
svn path=/trunk/; revision=68233
- Add basic implementations of RtlApplicationVerifierStop and RtlCopyMappedMemory for the benefit of some Windows dlls
CORE-9857 CORE-9858 #resolve
svn path=/trunk/; revision=68232
UpdateDiskLayout:
- Set the proper HiddenSectors value.
- Fix counting the PartitonaNumber.
UpdatePartitionNumbers() has become obsolete. Remove it.
svn path=/trunk/; revision=68226
- Add a test for SystemFlagsInformation that demonstrates the parameter checks performed by NtQuerySystemInformation/NtSetSystemInformation
CORE-9849
svn path=/trunk/; revision=68225
As shown by the non commited yet tests from Thomas, we have to check for privilege in NtSetSystemInformation - SystemFlagsInformation
svn path=/trunk/; revision=68224
- Probe input buffer for read in NtSetSystemInformation() and call the helpers function in a PSEH block
- For the NtSetSytemInformation - SystemSessionCreate specific case, as we return session ID, probe for write directly there.
svn path=/trunk/; revision=68223
Don't trust the user!
Probe buffers in NtSetSystemInformation - SystemSessionCreate and in NtSetSystemInformation - SystemSessionDetach
svn path=/trunk/; revision=68221
Begin upgrading the VGA emulation to SuperVGA emulation
(Cirrus Logic CL-GD5343 compatible).
This is still a WIP. So far only the VCLK and MCLK have been implemented.
svn path=/trunk/; revision=68216
- Fix assertion failure in 2nd stage by setting s_IsTerminated when an instance does not yet exist.
- Rename CComCreatorSingleton to CComCreatorCentralInstance since it's not /technically/ a singleton
svn path=/trunk/; revision=68215
- Also make My Computer a singleton, as shown by the CMyComputer test. With r68210 and this commit, SHFileOperation on a large number of files is approximately 3 times as fast.
CORE-9839 #resolve
svn path=/trunk/; revision=68212
- Introduce a CComCreatorSingleton class and use it for CDesktopFolder. Multiple calls to SHGetDesktopFolder will now return the same instance, which provides a massive speedup in many shell operations.
CORE-9839
svn path=/trunk/; revision=68210
- Add a test showing that repeated calls to SHGetDesktopFolder return pointers to the same object, and that its CreateViewObject method instead creates new objects.
CORE-9839
svn path=/trunk/; revision=68200
- Use a separate heap allocation for the thread parameters to ScServiceMainStub, since the thread can live longer than the ACTIVE_SERVICE structure
CORE-9235
svn path=/trunk/; revision=68196
- Restructure co_UserProcessHotKeys to correctly process modifier-only hotkeys. Fixes start menu opening after Win+X shortcuts, and probably several other hotkey-related issues. Thanks to reactosfanboy.
CORE-9728 #resolve
svn path=/trunk/; revision=68190
- Add tooltips for the toolbar
- Fix the 'disable' button not showing for devices that be disabled
- Some code cleanup / noise
svn path=/trunk/; revision=68185
- Add support for WM_ENTERMENULOOP/WM_EXITMENULOOP and update the status bar style to allow hints to be drawn
- Fixup the menu hints
svn path=/trunk/; revision=68184
- In 'Devices by Connection, if a device has a problem, expand the treeview to to show that problem device.
- Add a missing break in WM_COMMAND to stop the app from closing
- Fix clearing the lists
svn path=/trunk/; revision=68182
- Make class and device nodes polymorphic and move the code into separate files
- Add an abstract base class which we attach to the listview nodes
- Add double clicking of nodes
svn path=/trunk/; revision=68180
- Set the correct length in FsdGetFsVolumeInformation/FsdGetFsAttributeInformation, and write back partial data when returning STATUS_BUFFER_OVERFLOW
- Replace unnecessary checks with ASSERTs
CORE-9820 #resolve
svn path=/trunk/; revision=68175
- Move the devmgmt code to devmgr where it belongs and start to rewrite it (devmgmt_new wasn't a good design). It's not part of devmgr yet, I'll merge it and add it to the build when it's more complete.
- Add support for caching devices to speed up switching device views
- start to add dynamic context support so we can enable/disable, update and uninstall devices depending on its state.
- WIP
svn path=/trunk/; revision=68165
- Simplify PicIRQComplete helper function.
- Set up temporary stack for the BIOS.
- Initialize the "User Data Area" that is found at 0050:xxxx (see http://helppc.netcore2k.net/table/memory-map for more details).
- Some INT --> UINT
- Simplify few macros.
- Simplify DOS initialization; add few functions that will be useful later on.
- Remove trailing whitespace.
svn path=/trunk/; revision=68159
Improve the recovery console:
- Add help texts.
- Implement the partinfo command which displays the partition table of the chosen drive.
svn path=/trunk/; revision=68158
Fix broken filesystem detection.
- Treat FAT32 filesystems like FAT filesystems because we do not provide a FAT32 entry in the filesystemlist.
- HACK: Do not try to check an unidentified filesystem.
CORE-9815 #resolve
svn path=/trunk/; revision=68156
PATCH: Get configure and host-tools building on arm again by Steven Edwards from dev-ml
+ adding Timo's portable sqrt variant to where it's needed to fix win32k build.
+ disabled the assert in sqrt.c for now.
svn path=/trunk/; revision=68145
Fix compilation for AMD64 and ARM.
- It's no good idea to use the x86 ASM optimizations on any architecture. Luckily Alter made plain C++ ones for that case, too. Thus modify the #ifdefs a bit to use em.
- Fix a few directly used functions instead of the defined architecture fixed ones of em.
- After some discussion with Thomas Faber the ULONG vs UINT32 problems were resolved by using the same platform defines for all three architectures.
BUILDS? yes WORKS? Trololololo ^^
svn path=/trunk/; revision=68144
Complete NtQuerySystemInformation - SystemNumaProcessorMap stub so that it is possible to query the highest node number (and only that!)
CORE-9680
svn path=/trunk/; revision=68142
Improve bin2c by saving in a #define the size of the generated array, and specify this size in the declaration of the array so that doing sizeof(...) becomes possible. Use "unsigned char" as the (portable) type of the contents of the array.
[HAL]
Add PCHAR casts where needed to take into account the previous modifications.
svn path=/trunk/; revision=68139
Implement GetNumaHighestNodeNumber()
Implement GetNumaNodeProcessorMask()
Implement GetNumaProcessorNode()
Implement GetNumaAvailableMemoryNode()
They won't work yet though, given that the kernel mode counterpart is still unimplemented.
CORE-9680
svn path=/trunk/; revision=68136
- Stubplement the font generator functions for text-mode.
- Store in plane 2+ the font ONLY in text mode (guilty r68085). Fixes spurious graphics corruption in gfx modes.
CORE-9791
CORE-9794 #resolve #comment Should be fixed in r68134.
svn path=/trunk/; revision=68134
- Immediately finish a zero-length request in TdiSendDatagram. This may not be completely correct (should possibly get forwarded to tcpip instead) but is better than asserting.
CORE-9810
svn path=/trunk/; revision=68131
Addendum to 68126. The loader should return STATUS_INVALID_IMAGE_PROTECT
if e_lfanew is 0, or if the DOS stub appears too large due to a bad
e_lfanew. (Note: All plain MZ executables have an invalid e_lfanew)
svn path=/trunk/; revision=68127
- Set VadType = VadImageMap for image section mappings
- Rewrite MmGetFileNameForAddress and NtAreMappedFilesTheSame. Take into account that the first address must be an image section mapping
- Fixes a few kernel32_winetest:virtual tests
CORE-9736 #resolve
svn path=/trunk/; revision=68123
- Implement mandatory filesystem selection, formatting and filesystem checks for all new partitons.
- Implement optional filesystem selection, formatting and filesystem checks for formatted boot and install partitions.
- Enable installing ReactOS on primary partitions other than the first one. Usetup will install Freeloader on the first (or active) partition (aka boot partition) and install ReactOS on the chosen partition (aka install partition).
svn path=/trunk/; revision=68112
- Some INT --> UINT fixes.
- Fix the third value of MemorySizes according to the specification, confirmed by Aleksander.
svn path=/trunk/; revision=68099
Fix IMUL. MINSHORT and MAXSHORT aren't good constants since they don't have
a (SHORT) cast, and (LONG)((SHORT)MINSHORT) is uglier than just using -32768.
svn path=/trunk/; revision=68078
Fix INT 21h, AH=55h. It's supposed to set the current PSP too.
Fix the order of registers in DOS_REGISTER_STATE to make it more compatible.
svn path=/trunk/; revision=68066
Revert r67603, apparently DosTerminateProcess should just always pop the registers.
Make our DOS_REGISTERS_STATE structure more compatible.
svn path=/trunk/; revision=68065
Stubplement INT 25h and INT 26h. These interrupt calls are really weird since
they don't IRET on return, so the default interrupt stub messes up the stack.
svn path=/trunk/; revision=68054
Programs don't expect the XMS function AH = 07h to access the keyboard
controller, since the specification says it's done in a hardware-independent
way. So, reimplement this function by calling EmulatorGetA20() instead.
Also, for the sake of simplicity, use EmulatorSetA20 instead of PCAT_A20Control
to change the A20 state.
svn path=/trunk/; revision=68029
- Implement the DOS idle interrupt.
- Link to the parent's environment block by default in DosCreatePsp.
- Fix INT 21h/36h (Get Free Space) to report 0xFFFF if the number
of clusters is too high to fit in a word.
svn path=/trunk/; revision=68028
opgroups.c: We shouldn't restore SP in the opcode handler as that might
interfere with triple faults. Also, in this case it's also wrong because
the exception procedure already restores the stack pointer.
common.c: Fix a comment, no code changes.
svn path=/trunk/; revision=68027
Missing italian translations in ReactOS by Carlo Bramini
CORE-9725 #comment WineD3D, ieframe and shortcuts patches committed. thx ^^
See issue 9725 for more details.
svn path=/trunk/; revision=68011
Add/Update Chinese (zh-CN) translation by Henry Tang Ih
CORE-9003 #comment Forgot one file...
See issue 9003 for more details.
svn path=/trunk/; revision=68010
Implement XMS functions 10h, 11h and 12h by forwarding them to the
VDM DOS memory manager.
CORE-9651 CORE-9741 #resolve #comment Fixed in revision 68001.
svn path=/trunk/; revision=68001
Add/Update Chinese (zh-CN) translation by Henry Tang Ih
CORE-9003 #resolve #comment Trunk patch applied, thx ^^
See issue 9003 for more details.
svn path=/trunk/; revision=67993
- Move the fastio code to a separate file.
- Add Pierres IRP-Queue code and use it for IRP_MJ_QUERY_VOLUME_INFORMATION, IRP_MJ_SET_VOLUME_INFORMATION, IRP_MJ_QUERY_INFORMATION, IRP_MJ_SET_INFORMATION, IRP_MJ_DEVICE_CONTROL and IRP_MJ_FILE_SYSTEM_CONTROL requests.
svn path=/trunk/; revision=67988
Directly implement NtfsOpenFileById() for special files from the MFT where we cannot moonwalk there path.
It concerns any file with ID < 0x10.
This function can only cope with file which ID is < 0xc. Windows Internals reports that 0xc - 0xf entries are unused.
It would be interesting to double check returrned status in case of error, and also what Windows returns when attempting weird opening of suchs entries (overwrite, create for unused, etc).
svn path=/trunk/; revision=67983
At the request of Christoph von Wittich, bring devmgmt_new into the build.
* Created CMakeLists.txt and added to parent folder.
* Removed some references to CAtlString that were not needed and are not implemented in our headers.
* Fixed build with gcc.
* Spread all the changes done to the english resources to the other languages. Many languages now contain untranslated strings that will need to be fixed.
Note that I made no effort to fix any bugs or improve any features. The app comes as-is, except it now builds with rosbe.
svn path=/trunk/; revision=67976
Update LibreOffice version to 4.4.3 and add zh-CN translation in RAPPS/RAPPS_NEW by Henry Tang Ih
CORE-9748 #resolve #comment Committed, thx
svn path=/trunk/; revision=67966
- When a DOS application is started by a process that has no console,
the ConsoleHandle in the VDM_CONSOLE_RECORD is initially NULL.
Fix that by making an initial VDM_GET_FIRST_COMMAND call in NTVDM
and setting the ConsoleHandle to the handle of NTVDM's console in
BaseSrvGetNextVDMCommand.
- Fix a bug in BaseSrvGetVDMExitCode (VDM_READY means the task has exited).
svn path=/trunk/; revision=67961
Implement a portable version of sqrt. It's using an optimized combination of different algorithms, resulting in code that calculates the result in the maximum possible (when internally using double) precision with only 11 multiplications and a single division.
svn path=/trunk/; revision=67956
- Add FdoQueryBusRelations to enumerate the 'attached' raw port.
- Create the PDO for the raw port (\Device\ParallelX) and let the symbolic link (\DosDevices\LPTX) point to it.
- Move the open, close and write code from the FDO to the PDO.
svn path=/trunk/; revision=67936
Now that the Fast486 FPU is feature-complete, tell programs that we have a builtin FPU.
(This commit may cause regressions if there are bugs in the FPU)
svn path=/trunk/; revision=67934
Implement FPTAN by calculating the sine and then dividing it by
the approximated cosine: sqrt(1 - sin(x) ^ 2).
This method turned out to be faster and more precise than the
Maclaurin series for tan(x), which converges very slowly.
svn path=/trunk/; revision=67932
- Don't attempt to moon walk path name for reserved MFT entries.
- Don't attempt to dereference file name if it doesn't exist
svn path=/trunk/; revision=67918
- Use UserRefObjectCo in IntNotifyWinEvent to avoid a reference leak in case the call-out does not return
- Sanitize list walk
svn path=/trunk/; revision=67907
Add very long tests for tunnel cache.
So far, they are skipped because they are really long (they require more than 15s to complete).
Unfortunately, for unknown reason now, they are failing in ReactOS (whereas they shouldn't!).
An idea would be to see whether we can set an expiry of 1s for tunnel cache just for dbgtests and query the value from registry, so that it could help reduce test duration on ReactOS (but not on w2k3).
ROSTESTS-177
svn path=/trunk/; revision=67901
Add simple tests regarding renaming and tunnel cache behavior for short names in user mode.
Longer tests should be added to show what happens in case of cache expiry.
All pass on w2k3 but not on ReactOS.
CORE-3875
svn path=/trunk/; revision=67897
Fix the implementation of FSIN and FCOS:
First of all, 4 steps for a Maclaurin series is not enough.
Second, the angle should be normalized to the [-pi/2, pi/2] range.
svn path=/trunk/; revision=67894
- Move the code to copy and rename hardware profiles from the dialog procedures into separate functions.
- Use the dialogs only to retrieve the destination profile name.
- Do not copy or rename a hardware profile if it source and destination name is the same.
- Replace a hard.-coded string by a resource string.
svn path=/trunk/; revision=67889
Short commit message: implementation of (names) tunnel cache in file system RTL.
This is in the vast majority work done by Johannes Anderwald, I've just reviewed, fixed a few things, and implemented last bits.
Thanks to Johannes for his initial implementation (and huge work!).
Dedicated to Hervé's secret plans ;-).
CORE-7272
CORE-3875
svn path=/trunk/; revision=67886
Don't attempt to allocate null-size buffer when copying empty string.
Note to next generations students: test your code and DON'T LEAK MEMORY!!!!
It still leaks atm, will have a look later on.
Note to Mr Bragin, beloved ReactOS Project Coordinator: did you actually reviewed your students work?
CORE-7272
svn path=/trunk/; revision=67885
- Add a debugging assert that will save your lot of time if you attempt to manually queue a wrong IRP. You cannot queue an IRP which is set to be later queued again. This is a broken behavior.
- Remove a bit of code duplication
svn path=/trunk/; revision=67880
- Update the CrtModeControl byte in the BDA when we change video modes.
- Implement INT 10h, AX=1003h "Toggle Intensity/Blinking Bit".
- Partially implement INT 10h, AH=1Bh "Functionality/State Information (VGA)".
svn path=/trunk/; revision=67879
- Only free the IRP context in case the request is not to be queued
- Stubplement request queueing support in NtfsDispatch()
svn path=/trunk/; revision=67877
Split NtfsFsdDispatch() in two.
Now, NtfsFsdDispatch() will only take care of allocating the IRP context and complete the IRP in case it failed (that fixes a null-pointer dereference, just for the record).
NtfsDispatch() will really dispatch the IRP to the internal functions. So that it can be called either directly from NtfsFsdDispatch() or by a queued IRP (to be implemented).
svn path=/trunk/; revision=67875
Use a better algorithm for the base 2 logarithm.
Put the shared code for FPREM and FPREM1 in an inline function used by both.
svn path=/trunk/; revision=67855
Due to some kind of mixup, I thought the series for log(x) was actually for log10(x).
Fix Fast486FpuCalculateLogBase2 by dividing the result with log(2) and not log10(2).
svn path=/trunk/; revision=67841
Fix a copy-paste bug. Adjust the number when subtracting in Fast486FpuAdd. Don't overflow
the mantissa in Fast486FpuToSingle/Fast486FpuToDouble without correcting the exponent.
svn path=/trunk/; revision=67838
INI files for setting parameters? Lets get to the past Win 3.11 standards and make a configure dialog for that which reads and writes the stuff to the registry.
svn path=/trunk/; revision=67833
Followup to r67830: When no one handles a message, properly return that it was not handled.
This way all messages are handled correctly and not just NCCREATE.
Also took the chance to short-cut NCDESTROY, since the icon handlers have no need to receive this message.
CORE-9592 #resolve
svn path=/trunk/; revision=67831
do not handle WM_NCCREATE in ProcessWindowMessage
otherwise it get passed to ProcessIconMessage which will flag it as handled in case g_NumIcons is 0
which then will result in ATL not calling DefWndProc anymore
CORE-9592
svn path=/trunk/; revision=67830
Fix Fast486FpuArithmeticOperation. It doesn't really matter for the operation
itself which operand is the source/destination, because the FPU actually
always performs the operation on ST0 and something else. So, FDIV/FDIVR were
never really reversed here, it's just that FDIV always divides ST0 by something
and FDIVR always divides something by ST0.
svn path=/trunk/; revision=67828
Fix UnsignedDivMod128 (again).
Fix Fast486FpuAdd to handle numbers whose difference of exponents is greater than
the number of bits in the mantissa.
svn path=/trunk/; revision=67827
The maximum length of hardware profiles names is 79 characters. Use a matching string length and set the length of the edit control accordingly.
svn path=/trunk/; revision=67825
Implement support for FastIO for:
- FastIoQueryBasicInfo
- FastIoQueryStandardInfo
Now, with this commit and the two previous, ReactOS won't attempt to issue an IRP for these query, but will directly go with the FastIO path.
The performance improvement is really visible in 1st stage (at least, here with VBox).
svn path=/trunk/; revision=67824
Fix several division related problems:
- In UnsignedDivMod128, fix the leading zero calculation for 128-bit numbers.
- In Fast486FpuDivide, there's no need to increase the exponent by 64 (adding zeros
at the end of a number after the decimal point changes nothing).
- FDIV/FDIVR were reversed
svn path=/trunk/; revision=67820
- Zero is not a denormalized number.
- In Fast486FpuMultiply, we must account for the fact that the resulting mantissa
also has the "decimal point" at the second position.
svn path=/trunk/; revision=67819
Implement MiGetFileObjectForSectionAddress and use it in MmGetFileNameForAddress and NtAreMappedFilesTheSame. Don't call MmLocateMemoryAreaByAddress from these functions anymore.
svn path=/trunk/; revision=67810
Fix a fixme in VfatGetUserBuffer():
Only request high priority page when in the paging path. Otherwise, request normal priority
svn path=/trunk/; revision=67808
Don't check if a AMR3 owned memory area can be inserted, since we already know it's ok. Delete the related memory area code.
svn path=/trunk/; revision=67798
Add a tiny hack to MiFindEmptyAddressRangeInTree and MiFindEmptyAddressRangeDownTree to support VADs in the kernel VA range. Make MmFindGap use those functions to look for free space.
svn path=/trunk/; revision=67797
Get rid of MiRosTakeOverSharedUserPage. We don't need it anymore, since the user page is manually handled in the page fault handler and since it's outside the VAD region, we don't need to block that range. Now there should be no memory areas left at all in MmDeleteProcessAddressSpace.
svn path=/trunk/; revision=67796
Modify MmLocateMemoryAreaByAddress as well to use the VAD table. Since the page fault handler uses it to find ARM3-Fault pages and the shared user page does have a memory area, but not a VAD, add a check for the shared user page in the old fault handler.
svn path=/trunk/; revision=67795
Implement MiRosCleanupMemoryArea to cleanup memory areas from MmCleanProcessAddressSpace, since later when we remove the old-style memory area links, we will not be able to retrieve those.
svn path=/trunk/; revision=67794
Use the VAD tree to search for MEMORY_AREAS, except when inserting an ARM3 based MA. In this case we still need to use the old tables, since the related VAD is already inserted into the VAD tree.
svn path=/trunk/; revision=67793
- Allocate Memory area before trying to insert it in different pathes of MmCreateMemoryArea. This change is a preparation for upcoming changes.
- Cleanup MmInsertMemoryArea from useless stuff
svn path=/trunk/; revision=67791
Use 1 as base address parameter to MmCreateMemoryArea, when 0 is the actual base and remove FixedAddress parameter in MmCreateMemoryArea
svn path=/trunk/; revision=67790
Made mazescr use screensaver.lib for multi screen support
Add symbol to make things look more... mature in C:\ReactOS
Fixes of my two problems (app does not really terminate and maze is not generated completely) and tidy up (-200 lines of code!) of my mess and the mess we had in there before by David Quintana. Thx man :-D
svn path=/trunk/; revision=67781
- Add coprocessor constants
[NTOSKRNL]
- Handle process in NtFlushInstructionCache and validate parameters
- Use _MoveToCoprocessor instead of GCC inline assembly for ARM
svn path=/trunk/; revision=67770
- Simplify CurrentDirectories array usage.
- We can directly use a pointer to a DOS_DATA structure stored in the guest memory.
- Use the DOS data structures for CurrentDrive (in DOS_DATA) and LastDrive (this is in fact NumLocalDrives in the SYSVARS list).
- Use ANSI versions of GetLocaleInfo where applicable.
svn path=/trunk/; revision=67750
Bring in initiated work on IRPs in NTFS (and complete it).
This simplifies the way IRPs and their context are handled in the driver.
Only VfatDispatchRequest() is responsible of completing IRPs (if required), freeing the IRP context (if required), and queueing the IRP for later completion (if required).
This allows removing duplicated code, hacks, and so on. It might help reducing IRP leaks, or memory leaks.
By default, VfatDispatchRequest() will complete the IRP and delete the IRP context.
In case you don't want it to complete the IRP (because you passed it down, for instance), remove the IRPCONTEXT_COMPLETE flag. See for instance: VfatDeviceControl().
In case you want to queue the IRP (you previously called: VfatQueueRequest()), call the newly introduced VfatMarkIrpContextForQueue() function that will prepare it.
svn path=/trunk/; revision=67743
- Separate C++ exports per architecture, add ARM exports (that again have different mangling)
- Add symbol aliases for ARM C++ exports
- Compile in/out functions only on x86
- Add some missing exports for non-x86
svn path=/trunk/; revision=67730
setjmp is a macro expanding to _setjmp or _setjmpex. In soviet russia and on ARM builds, the compiler will convert invocation of _setjmp back into a call to setjmp. The compiler handles _setjmp and _setjmpex in the same way, generating code where r0 points to the jump_buf and r1 points to the current functions stack frame (sp on function entry). So make setjmp and _setjmp aliases to _setjmpex and make the latter use r1.
svn path=/trunk/; revision=67729
- Add some early "return" after calls to EmulatorTerminate()
- Some VDDs rely on the fact that NTVDM calls ExitProcess on Windows when it is going to quit, so let's do the same.
- Put port 61h (and 62h) support in ppi.c (programmable-peripheral-interface).
svn path=/trunk/; revision=67726
Create the "Hardware Profiles\0000" key as part of the registry setup. This must be done at this point because the "Hardware Profiles\Current" link will only be created, if the "Hardware Profiles\0000" key exists.
svn path=/trunk/; revision=67725
Partly revert previous changes, bringing back copysignf. Add global defines copysignf=_copysignf, since we have that now.
svn path=/trunk/; revision=67723
Implement basic funtions of the parallel port driver. It supports Centronics-Mode only.
You can print files using: copy <filename> lpt1:
svn path=/trunk/; revision=67713
- Remove useless copysignf invocation and the wrong defines for it
- BrokenHacksOnTopOfBrokenCode--;
- Should be sent upstream
svn path=/trunk/; revision=67709
This is evil under amd64, too... so we disable it for now.
[ROSHTTPD]
UINT -> size_t
[NCFTP]
About time to clean this mess. Are we a Unix? No sir!
svn path=/trunk/; revision=67697
Wine guys... if you put all stuff which uses one local variable in a #ifdef __i386__, then god heaven... put that one in a #ifdef, too! Should be sent upstream!
svn path=/trunk/; revision=67696
Fix the XMS driver. AH = 08h is supposed to return the size of the
largest free block in AX, and the total amount of *free* memory
in DX.
svn path=/trunk/; revision=67675
Implement VDD user hooks: they are called when a DOS program is started or terminated, and when the VM enters a blocking state or is resumed.
[NTVDM]
- Do some parameter checks and set last errors in VDD memory helpers.
- Fix a bug in VDDInstallIOHook when installing IO hooks: the IoHandlers pointer MUST NEVER be incremented!!
[TESTVDD]: Update my test VDD with tests for the user hooks.
svn path=/trunk/; revision=67671
Use the SDA for storing the request in DosCallDriver.
Rename some more fields we aren't going to use to "Unused..." in the SDA.
svn path=/trunk/; revision=67651
The LMSW instruction doesn't #GP when the program tries to return to real mode.
Rather, it just doesn't happen (the PE bit remains 1).
svn path=/trunk/; revision=67647
- Fix/add a number of macros based on native definitions
- Add "intrinsics" for GNU ARM assembler
- Move trap macros to asmmacro.S
svn path=/trunk/; revision=67639
- Only set the IOSB status if we complete the IRP
- Fix the assert: the caller might want not to complete the IRP nor to queue it (passed down IRPs)
svn path=/trunk/; revision=67636
Add two more flags (mutually exclusive) for IRP context:
- _COMPLETE will cause the IRP to be completed at the end of the dispatch, with the run-time priority boost set by caller
- _QUEUE will cause the IRP to be queued for delayed execution (not yet implemented)
This allows more flexibility for callers that can set the behavior thanks to the flags.
Default behavior is the previous one: by default the IRP is completed at the end of the dispatch
That one should really come to FastFAT...
svn path=/trunk/; revision=67635
- Move stuff from armddk.h to XDK/DDK
- Fix ARM CONTEXT flags
- Add missing KeGetCurrentIrql() for ARM
- Improve some amd64 based definitions
- Use Native math for Int64Sh**Mod32 on ARM
svn path=/trunk/; revision=67631
- Remove unused HYPERSPACE_BASE (which was broken for x86, too)
- Add PAE values for HYPER_SPACE and HYPER_SPACE_END on x86
- Add missing MM_HAL_VA_START and MM_HAL_VA_END for ARM
svn path=/trunk/; revision=67630
Don't force the run-time priority boost in NtfsFsdDispatch(). Some dispatched MJ functions might want to adjust it (MJ_READ/MJ_WRITE/MJ_CREATE).
The same should be actually done in FastFAT. I believe that would help avoiding several code duplications...
svn path=/trunk/; revision=67624
- Modify the int32 dispatch to clear the CF only on entry, so that we can
track in our handlers whether it was the first time the BOP executed.
- Implement INT 15h, AH = 41h (Wait On External Event).
- Fix INT 15h, AX = E801h - modify the stack CF instead of the handler CF.
- Make INT 16h AH = 00h clear CF when there is a character, so that the BOP doesn't
repeat.
svn path=/trunk/; revision=67612
- Implement XMS function 0x0E (Get Handle Information).
- Fix INT 15h, AX = E820h - setCF can't be used inside an int32 callback for changing
the return carry flag, because it has a different meaning there (it means that the BOP
should be repeated). Instead, we must modify the CF that was pushed on the stack.
svn path=/trunk/; revision=67611
- Enable C4189 (local variable initialized but not referenced) on MSVC builds and make it an error.
CORE-9676 #resolve
svn path=/trunk/; revision=67607
DosTerminateProcess shouldn't blindly attempt to restore the old state,
(for example if the process was started manually by a debugger).
svn path=/trunk/; revision=67603
- Move some interrupt initialization where it belongs and activate few other ones.
- Initialize interruct vector for INT 2Ah (Network check) because some apps directly call it and crash if it contains NULL.
svn path=/trunk/; revision=67595
- Improvement/Usability: when a DOS app is started in a separate console, print its full path as it's done in Windows.
- Stubplement a LoadGlobalSettings function for NTVDM so that it will be able to grab user settings from the registry.
- Move DOS-only related variables AcceptCommands, CommandThread and SessionId to where they belong. I will rework ConsoleCtrlHandler to make it more elegant in the future (DOS-agnostic).
- Remove the useless ENTER-key-up hack.
svn path=/trunk/; revision=67589
Remove notification icons when their attached HWND is no longer valid. Patch by Joachim Henze (reactosfanboy), with some extra newlines by me.
CORE-9537 #resolve
svn path=/trunk/; revision=67583
Implement EMS function AH = 53h (Get/Set Handle Name).
Some programs expect to find the string "EMMXXXX0" at offset 0x0A in
the segment where the EMS interrupt handler is, so move the interrupt
handler BOP to the private device area of the EMS driver.
svn path=/trunk/; revision=67573
- Add nothrow versions of new/delete operators
- Add <cassert>
- Fix a TODO and add missing dependency to <exception> header
svn path=/trunk/; revision=67555
Add some missing german translations
Add some missing manifests
Improve some resources
Add possibility to build an application with one language
svn path=/trunk/; revision=67552
- Implement call gates.
- Fix Fast486InterruptInternal and get rid of ugly hacks.
- Fix IRET (CS loaded too early).
- Add a structure definition for the 16-bit TSS (not supported yet).
svn path=/trunk/; revision=67545
- Implement portable fabsf.c
- Move some ARM stubs to a better location
- Add stubs for __dtoi64, __i64tos, __u64tos, _clearfp
svn path=/trunk/; revision=67537
- Handle failure to allocate VM for thunk code
- Factor out the assembly thumk code, fix amd64 code and implement ARM code
svn path=/trunk/; revision=67535
- Convert tabs to spaces
- Remove excessive local variable indentation for alignment, since that doesn't help to improve readability
- No code changes
svn path=/trunk/; revision=67534
- Fix a typo.
- Use a #define for the device flag.
- In r67526, the following change was committed but not described in the commit message:
- Enable opening devices in DosCreateFile and DosCreateFileEx.
svn path=/trunk/; revision=67527
Update the copyright year (better late than never).
Push the error code inside Fast486InterruptInternal, to make the size of the
pushed value on the stack correct.
Update the CPL in Fast486TaskSwitch.
svn path=/trunk/; revision=67526
Add a few definitions based on (Windows RT 8.1) symbol files, ks386.inc, ksamd64.inc, ksarm.inc and Windows 10 WDK (ntosp.h ftw!)
svn path=/trunk/; revision=67525
- Implement KiCpuId and make use of it
- Get rid of ugly CPUID, RDMSR and WRMSR functions
- remove unused KTS_ constants
svn path=/trunk/; revision=67524
- Implement C versions of RtlInterlockedPushEntrySList, RtlInterlockedPushListSList, RtlInterlockedPopEntrySList and RtlInterlockedFlushSList.
- RtlInterlockedPushListSList is fastcall, not stdcall
svn path=/trunk/; revision=67522
- Improve definition of ExInterlockedCompareExchange64 and interlocked SList functions and improve some annotations based on native header
- Update processor feature constants
svn path=/trunk/; revision=67521
Mouse:
- Remove now useless code in mouse support in BIOS.
- Implement INT 33h, AH=1Bh "Return mouse sensitivity" (returning hardcoded standard values since we don't support custom sensitivities) (AH=13h and 1Ah are marked as UNSUPPORTED).
- INT 33h, AH=1Fh "Disable mouse driver" returns in ES:BX the old INT 33h vector value: implement that.
- Implement INT 33h, AH=21h "Software reset", AH=24h and 4Dh and 6Dh "Software version and mouse info", "pointer to copyright string" and "version 'string'" functionalities.
DOS:
- Initialize in Win2k3-ntvdm-compatible order the DOS drivers: NUL, then CON, then XMS, and then EMS.
- Fix segment/offset inversion usage in INT 21h, AH=5Ch "Lock/Unlock region of file": when using MAKELONG macro to build a far pointer, the first parameter (loword) is the offset, and the second parameter (hiword) is the segment.
svn path=/trunk/; revision=67518
When storing a segment selector, the operand size attribute is only ignored when
writing to memory (where it's treated as if it's always 16-bit).
svn path=/trunk/; revision=67512
Implement INT 27h (Terminate and Stay Resident).
Keep track of the last entry SS:SP in the INT 21h handler.
Restore the stack in DosTerminateProcess.
The number of bytes to keep resident applies only to the block which holds the PSP,
other blocks are not freed.
svn path=/trunk/; revision=67507
Separate the process-related code from the DOS kernel.
Enable starting processes from other processes in STANDALONE mode.
Implement INT 21h, AH = 55h and INT 21h, AH = 26h (Create/Clone PSP).
Implement overlay loading.
svn path=/trunk/; revision=67498
- Arch, don't assume the driver you see in the backtrace is one that needs fixing. CDFS is the one who's a file system driver here...
svn path=/trunk/; revision=67496
[BUILD]
msvc.cmake: Disable thread-local static initialization.
CMakeLists.txt: Disable PCH for VS2015.
configure.cmd: Make it aware of cl.exe version 19.x
[CPPRT]
Add alias for the new variants of the delete operators.
[BROWSEUI]
[MFIFS]
[FRAMEDYN]
[NDIS]
[DDK]
[PSDK]
[STLPORT]
Add explicit declarations of the new delete operators for those modules that don't use the WITH_RUNTIME option.
[WIDL]
[WPP]
Do not alias the snprintf family of functions to the _snprintf variants, since VS14 already declares them internally.
svn path=/trunk/; revision=67483
When a segment isn't present, the exception should always be #NP (unless
it was the stack segment, in which case the exception should be #SS).
svn path=/trunk/; revision=67479
The fix in r67477 was incorrect. When loading an executable high, we need to
allocate all of the memory and copy the program itself to the highest part of it.
I know it sounds illogical but that's how it works...
svn path=/trunk/; revision=67478
Also check for BlockData == 0 in RosResizeMemory. This, however, wasn't a crash since we
were lucky enough to cast the segment to a WORD before the access.
svn path=/trunk/; revision=67476
Continue the work converting the UI to using ATL-wrapped windows. CORE-9593
[RSHELL]
[EXPLORER]
Fixup two class declarations to work with the rosctrls.h changes.
svn path=/trunk/; revision=67474
- Move definition of TIMER_TABLE_ENTRY from ntoskrnl to NDK
- Add definitions of KTIMER_TABLE, KDPC_LIST, SYNCH_COUNTERS
svn path=/trunk/; revision=67471
Use portable definitions for service table constants. (Moved from ARM specific definition in ntoskrnl internal header, with a fixed value for TABLE_NUMBER_BITS, which is 1, not 3)
svn path=/trunk/; revision=67470
- Consistently return STATUS_INVALID_IMAGE_PROTECT from PeFmtCreateSection if the PE signature is missing. Patch by V.
CORE-9618 #resolve
svn path=/trunk/; revision=67464
In DosAllocateMemory, when the "last fit" allocation strategy is selected, split
the block so that the last part of it is used.
svn path=/trunk/; revision=67462
Initial work on a lightweight layout engine for ATL::CWindow.
Although a lot of work is still required making the classes more generic and layout-aware,
but it is already able to replicate the original rapps layout without explicit positioning of the windows.
Committing mostly for backup/history purposes.
CORE-9593
svn path=/trunk/; revision=67460
Remove the useless wrappers BiosPeekCharacter and BiosGetCharacter. They're also
a source of bugs since 0xFFFF could legitimately appear in the buffer.
svn path=/trunk/; revision=67459
- In INT 21h, AH = 0Ah, the final carriage return is not counted.
- Implement XMS function 0Bh (Move Extended Memory Block).
- The driver must preserve the contents of unlocked Extended Memory Blocks (EMBs),
so move the bitmap setting/clearing code into XmsAlloc and XmsFree.
- Make XmsAlloc optimize memory blocks by moving them around, except it can't
move locked blocks.
svn path=/trunk/; revision=67457
Implement INT 21h, AH = 52h (Get List of Lists - SYSVARS).
Rewrite the SFT code to keep the SFT in guest memory.
Implement the NUL device driver.
svn path=/trunk/; revision=67444
Fix RegQueryInfoKey
- Return a value in lpcClass if lpClass is NULL
- Return correct error codes as shown by tests
- Correctly name parameters -- most of these are NOT byte counts
CORE-6976
svn path=/trunk/; revision=67425
- Add a test for RegQueryInfoKey. This serves as a placeholder to be replaced with the upstream version once accepted -- however it tests against Win 2003 behavior, so it can validate changes to our advapi32.
CORE-6976
svn path=/trunk/; revision=67424
Remove hack in PrivMoveFileIdentityW and properly handle failing call to BasepNotifyTrackingService().
Fixes regression brought by r67355
svn path=/trunk/; revision=67423
- Add a test for NtGdiAddFontResourceW. This goes in gdi32_apitest until we fix w32knapi to be Testman-compatible (ROSTESTS-167). Patch by Víctor Martínez Calvo.
ROSTESTS-166 #resolve
svn path=/trunk/; revision=67391
- Don't use the connection's name as the tab title in the Connection Properties dialog. Patch by Lee Schroeder.
CORE-9603 #resolve
svn path=/trunk/; revision=67390
- Avoid printing an unterminated string in registry test. Powered by DPH.
- Add a ROS hack to avoid a buffer overflow when one of our DPRINTs is returned by WaitForDebugEvent (e.g. when using DPH)
svn path=/trunk/; revision=67382
When the VGA memory map is A0000-BFFFF, mirror the bottom 64 KB into the top.
This should take care of all VGA addressing problems.
svn path=/trunk/; revision=67377
* Turn the files into .CPP, to ease future changes. All the code remains class-free for now, with one exception.
* Adapt the download progress dialog implementation to ATL/C++, which is a best match for COM classes.
* Add an ATL CModule class to the process, necessary for
CORE-9593
svn path=/trunk/; revision=67376
Store the return address in INT 22h as a far pointer.
The amount of memory stored in the Env variable is EnvSize, not sizeof(Env)
since Env is a pointer.
svn path=/trunk/; revision=67369
In DosCreateProcess, allocate space for the environment block dynamically and expand
it if needed, just like we do in CommandThreadProc.
In GetNextVDMCommand, remove the check that checks whether VDMState is one of
VDM_NOT_LOADED, VDM_NOT_READY or VDM_READY - that check makes no sense whatsoever,
since those aren't input values for that structure field. Their bit masks do correspond
to valid input fields, but even then the check makes no sense.
svn path=/trunk/; revision=67363
- Report A20 line status in the PS/2 controller output port.
- Properly implement XMS functions 3 and 5 (Global Enable/Disable A20 line), 4 and 6 (Local Enable/Disable A20 line) and 7 (Get A20 line status) using flag+counter and PS/2 I/O calls.
- Fix XMS driver version report.
svn path=/trunk/; revision=67361
- Halfplement BasepGetComputerNameFromNtPath() which allows querying the computer name given a handle & an NT path
- Implement BasepNotifyTrackingService() which issues a FileTrackingInformation on file move
svn path=/trunk/; revision=67355
Implement the XMS functions AH = 03h and AH = 04h (global enable/disable A20).
Make the GetHandleRecord function in the EMS driver inline.
CORE-8277 #comment Please retest.
svn path=/trunk/; revision=67352
- Fix copy-paste error in the initialization of DriverHeader.
- Use BOP_... syntax for BOP #define names (to comply with the rest of the code).
svn path=/trunk/; revision=67351
- Move the EMS code from the BIOS to the DOS driver where it belongs.
- Expand the DOS device API with a new function, DosCreateDeviceEx, which
will allow 32-bit DOS driver to reserve private memory.
- For each DOS device, create an entry in guest memory so that 16-bit code
can call 32-bit DOS drivers directly.
- Implement an XMS driver stub that uses the above.
- Arch, that's not how the DOS driver strategy routine works, you need to
give it the request in ES:BX which it will store somewhere, and then call
the interrupt routine.
svn path=/trunk/; revision=67339
- Implement BasepOpenFileForMove() used by PrivMoveFileIdentityW(), with its Windows bugs
- Implement PrivMoveFileIdentityW()
Not sure yet why Windows implementation exposes a sharing violation during our API test. Our absence of sharing violation looks more legit...
To be investigated.
svn path=/trunk/; revision=67335
Make the total amount of EMS pages configurable.
Allocate memory for EMS instead of sharing a part of the guest memory.
svn path=/trunk/; revision=67325
Make the GUI version actually work by disabling wmain per ifdef which will be set only at CUI build.
Fix resources to make the text fit nicely and get rid of the RLEXT/LTEXT mess
Add German translation
svn path=/trunk/; revision=67319
- Do not display a message box in CLanStatus::QueryInterface. Fixes hang when running com_apitest:netshell
- Change the order of some interfaces in CNetworkConnections to make com_apitest slightly happier
svn path=/trunk/; revision=67314
- Keep a CLanStatus instance around in CNetworkConnections (as before r65758) and properly manage its lifetime. Fixes Status context menu item for network connections.
CORE-9564
svn path=/trunk/; revision=67311
- Some additions to the previous commit
- Make destructors private
- Use delete to free object allocated using new
- Try to fix some ITEMIDLIST typing
svn path=/trunk/; revision=67310
Fix artifacts from incomplete C++ conversion:
- Add m_ prefix to member variables to avoid shadowing locals
- Use initializer lists in constructors
- Use destructors instead of putting code in Release
- Avoid C-style casts (not exhaustive)
svn path=/trunk/; revision=67307
- Move Network Connections back to Control Panel, where it belongs. Was accidentally moved to My Computer in r66485.
CORE-9495 #resolve
CORE-9564
svn path=/trunk/; revision=67298
Add ARM stubs for _abnormal_termination, _except_handler2, _except_handler3, _global_unwind2, _local_unwind2, __C_specific_handler, __CxxFrameHandler, __rt_srsh
[RTL]
Add ARM stubs for RtlGetCallersAddress, RtlDispatchException, RtlUnwind, RtlInitializeContext
[NTOSKRNL]
Add ARM stubs for READ/WRITE_REGISTER_*
svn path=/trunk/; revision=67294
We have a zero-warnings policy in trunk, accordingly:
- Disable overflow warnings in mciwave
- Remove all -Wno-error directives
- Remove all allow_warnings invocations
- Make allow_warnings defunct, it needs to be reworked, e.g. according to CORE-6959.
svn path=/trunk/; revision=67291
- Rearrange header inclusions. Put almost all NDK stuff in ntvdm.h.
- sizeof(array)/sizeof(array[0]) --> ARRAYSIZE(array) : increase code readability.
- Use Rtl heap functions (other parts of the code already did use them).
svn path=/trunk/; revision=67283
Add manifest to vmwinst, explorer and winlogon.
Remove em from telephon and ntobjshex again. No dialog, no manifest ^^
svn path=/trunk/; revision=67246
Add manifest to telephon.cpl.
[SHELLEXT]
Add manifest to netshell, stobject and ntobjshex.
Translate ntobjshex to German.
svn path=/trunk/; revision=67243
- The auxiliary PS/2 port IRQ is disabled by default.
- In the DOS mouse driver's IRQ handler, just reading the entire packet
is faster than waiting for one IRQ per byte. (real drivers can do that too,
using polling, which is slower on real hardware but faster in NTVDM).
- Increase the speed of the PS/2 IRQ scheduler so that mouse packets don't
fill up the queue completely. And consequently, increase the number of
CPU steps per cycle to avoid "stuck key" bugs.
svn path=/trunk/; revision=67231
- Make some compilers and static analyzers happy by removing a no-op condition. Patch by Samuel Serapion.
CORE-7214 #resolve
svn path=/trunk/; revision=67230
- Fix range check in NtUserSetWindowFNID. Patch by Samuel Serapion.
- Update comment on an old hack
CORE-7212 #resolve
svn path=/trunk/; revision=67228
- Make GetDriveType test succeed if the current directory is the drive root. Spotted by Edijs.
- Use strsafe functions
ROSTESTS-160 #resolve
svn path=/trunk/; revision=67222
- Completely rewrite the timing system. Replace hardcoded callbacks
with dynamic hardware timers.
- Finish implementing the PS/2 mouse.
- Fix the DOS mouse driver.
Inspired by a patch by Stefano Toncich (Tonix) (see CORE-9166).
CORE-9166 #comment A different fix was committed in r67219.
svn path=/trunk/; revision=67219
- Add the client interface code of the server service.
- Implement NetRemoteTOD. This will not fix CORE-5423 because the server side (NetrRemoteTOD) is not implemented yet.
svn path=/trunk/; revision=67197
Refactor the code for __rt_div to allow reusing it for __rt_sdiv64 and __rt_udiv64 and implement those as asm wrappers around worker functions (due to the calling convention they cannot be implemented fully in C)
svn path=/trunk/; revision=67195
Initialize the list of hooks before the call to NtAllocateVirtualMemory, to prevent crashes
in MemCleanup in case it fails.
svn path=/trunk/; revision=67186
Disable some functions in wksta_new.c and re-enable their counterparts in wksta.c. For some reason I cannot get the wine dummy code working in the workstation service. Perhaps it is yet another RPC bug.
CORE-9415
svn path=/trunk/; revision=67182
* Fix comparison length: StrCmpNW takes chars, not bytes.
* While at it, make canonical comparisons use memcmp instead of string comparison, as we only care about equality, not ordering.
Better fix for CORE-9432
svn path=/trunk/; revision=67181
Fix the Interlocked function mess. Implement all Interlocked functions in xdk/interlocked.h as they exist in native. Use this header to generate wdm.h and winnt.h. Remove Interlocked definitions from winbase.h. Fix definition of ExInterlockedAddLargeInteger.
svn path=/trunk/; revision=67173
Fix usage of interlocked[Compare]ExchangePointer. These functions take a PVOID* as first parameter and nothing else. In native headers they are inline functions, not macros.
svn path=/trunk/; revision=67171
- Add missing KeMemoryBarrierWithoutFence, NUMBER_POOL_LOOKASIDE_LISTS, MM_SHARED_USER_DATA_VA, MM_EMPTY_PTE_LIST, MM_EMPTY_LIST, P*E_PER_PAGE, SESSION_POOL_LOOKASIDES, MiPdeToPte for ARM
- Fix KI_USER_SHARED_DATA for ARM based on native DDK
svn path=/trunk/; revision=67168
* Make the info struct pointers const, to ensure we are not modifying the contents from the const idlists.
* Disable checking that the provided PIDL is part of the folder. Fixes enumeration in ros, but feels a lot like hiding a bug.
CORE-9432 #resolve
svn path=/trunk/; revision=67166
Add more recent Setup for Steam (2 years more recent) + localisations for the setup by Radek Liska
CORE-4196 #resolve #comment Added to RAPPS, have fun ^^
svn path=/trunk/; revision=67165
- Fix ARM build of libwine. Patch by Yuntian Zhang. CORE-8672 #resolve
- Don't include non-existant ARM intrinsic headers
svn path=/trunk/; revision=67149
- Remove nsIDOMWindowUtils::SetCriticalDisplayPortForElement, as it's no longer present in Gecko. Fixes stack corruption when exiting IE.
CORE-9513
svn path=/trunk/; revision=67135
Add Revo uninstaller 1.95 by nicoalta01
CORE-8892 #resolve #comment Added, have fun. I added CCleaner, too while bein at it ^^
svn path=/trunk/; revision=67133
- Add missing SEH/Probe in NtGetPlugPlayEvent and IopGetInterfaceDeviceList, and correctly copy the interface list. Patch by Stephan Röger.
CORE-9498 #resolve
svn path=/trunk/; revision=67129
- Fix uninitialized variable use in EnumEnhMetaFile
- Apply Wine commit 88e81025 (gdi32: Fix typo in a condition (PVS-Studio).) by André Hentschel
CORE-9191 #resolve
svn path=/trunk/; revision=67126
- Another addendum to r67107: don't let CleanupAndFreePacket try to free the buffer since it does not have the correct pointer
CORE-8210
svn path=/trunk/; revision=67110
CORE-9060 #comment rapps-db-update-and-enhancement-includes-es-es-l10n-2015-03-20.patch committed + continued your ways in some more entries + made the German translation follow it either + added some missing German translations.
svn path=/trunk/; revision=67100
- Rewrite the progress dialog making it responsible for controlling services. (starting, stopping, etc).
- Fixup the rest of the code to make use of running actions through the progress dialog.
- As a side-effect, this should fix CORE-4585
svn path=/trunk/; revision=67094
LTEXT "Hartă a caracterelor v0.1\nDrept de autor (C) 2007 Ged Murphy (gedmurphy@reactos.org)\nTraducere de Ștefan Fulea (stefan.fulea@mail.md)", IDC_STATIC, 48, 7, 210, 36
LTEXT "Hartă de caractere v0.1\nDrept de autor (C) 2007 Ged Murphy (gedmurphy@reactos.org)\nTraducere de Ștefan Fulea (stefan.fulea@mail.md)", IDC_STATIC, 48, 7, 210, 36
IDS_LICENSE "Această aplicație este publică; fiind permisă modificarea și/sau (re)distribuția sa în termenii Licenței Publice Generale GNU publicată de Free Software Foundation; sau versiunea 2 a Licenței, sau (la alegere) a oricărei versiuni ulterioare.\r\n\r\nAceastă aplicație este distribuită doar în speranța de a fi utilă, FĂRĂ însă NICI O GARANȚIE; nici măcar cu garanția implicită a VANDABILITĂȚII sau a UTILITĂȚII ÎNTR-UN SCOP ANUME. Pentru mai multe detalii consultați Licența Publică Generală GNU.\r\n\r\nPuteți vedea această licență aici:\r\nhttp://www.gnu.org/licenses/gpl.html\r\n\r\nDe asemenea puteți consulta traduceri neoficiale ale acestei licențe aici:\r\nhttp://www.gnu.org/licenses/translations.html"
[Sürücü:][Yol]Kütük adı İçeriği görüntülenecek bir kütük.\n\
Komut\t\t Çıkışı görüntülenecek bir komut.\n\n\
İstemde ""-- Sürdür --"" ile bir sonraki sayfayı göstermek için rastgele bir düğmeye basabilirsiniz.\n"
IDS_CONTINUE " -- Sürdür (%100) -- "
IDS_FILE_ACCESS "%s kütüğüne erişilemiyor."
END
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
