- Correctly delete resource in failure cases of CmpInitializeHive. Fixes pool corruption when trying to load an invalid hive file.
svn path=/trunk/; revision=63949
- Bring back PTE frame refcounting when serving a prototype PTE page fault.
- Fix a bug in MiDeletePte where the said PTE frame was not unshared.
- Improve transitional PTEs deletion (will be needed for future work).
Do not always trust the comments stating that "strange RosMm code broke everything"
svn path=/trunk/; revision=63947
- Addendum to r63933: Pass absolute value of the font width to freetype.
CORE-4657 Somehow got lost in the number of attached files.
Spotted by Thomas Faber
svn path=/trunk/; revision=63946
Implement a single-window mode for videos.
Patch by Ricardo Hanke.
TO ALL TRANSLATORS: Update the translations!
CORE-7910 #resolve #comment Committed, thanks!
svn path=/trunk/; revision=63940
- Add a "Play" menu, add keyboard accelerators and add a repeat mode.
Patch by Ricardo Henke.
- Few whitespace fixes.
- Fix french resources. TO ALL TRANSLATORS: Please update the translations!
CORE-7929 #resolve #comment Committed, thanks!
svn path=/trunk/; revision=63938
- Display formatted system messages in case of diverse failues.
- Add an error check for RegisterClassEx() and CreateWindow() in _tWinMain().
Patch by Ricardo Hanke.
CORE-7880 #resolve #comment Committed, thanks!
svn path=/trunk/; revision=63936
- Enable commented freetype call: set font face size and coordinate transformation matrix
Patch by Huw Campbell, reviewed by a gazillion of people.
CORE-4657 #resolve #comment committed in r63933, a million thanks
svn path=/trunk/; revision=63933
Marshall interface pointers during background delete by Huw Campbell
CORE-8460 #resolve #comment Committed by request of Victor. Have fun ^^
svn path=/trunk/; revision=63932
Do not hardcode event logs, but retrieve them from the registry.
Patch by Ricardo Hanke.
CORE-8458 #resolve #comment Committed, thanks!
svn path=/trunk/; revision=63930
- Make a conditional copy of reactos.inf when it is really modified, make reactos.acb depend on it.
Hopefully the last commit so that windows buildbot gets really fixed
svn path=/trunk/; revision=63924
- Improve the "infamous RLE hack" in SetDIBitsToDevice by using a mask bitmap corresponding to the valid RLE data
[GDI32]
- Improve some input checks
svn path=/trunk/; revision=63920
- Do not crash if an invalid PDO is given to IopEditDeviceList
- Always dereference device object when removing it (should be the last reference)
CORE-8294
svn path=/trunk/; revision=63917
Disable ScanNTFSStreams calls because this function exclusively calls BackupRead that is not implemented on ReactOS, and as a consequence makes lots of debug spam when browsing the start (sub-)menus.
svn path=/trunk/; revision=63914
- Remove "VACB chain" optimization for CcCopyRead that was never applied anyway since MAX_RW_LENGTH == VACB_MAPPING_GRANULARITY, so a single read could never exceed one VACB
- Consolidate CcCopyRead, CcCopyWrite and the cached case of CcZeroData into a common worker function
CORE-8410
svn path=/trunk/; revision=63903
- Reference the address file while delivering data to avoid a use after free when an address file is closed during datagram delivery
svn path=/trunk/; revision=63899
Fix a catastrophic bug in S/G DMA. There is a subtle difference between the S/G DMA APIs and the old AllocateAdapterChannel API when it comes to having multiple requests in flight. Callers of (Io)AllocateAdapterChannel CANNOT queue another request until the AdapterControlRoutine is called. S/G DMA allows multiple concurrent DMA requests, but ROS was using IoAllocateAdapterChannel in the S/G API. As a result, the wait block stored in the device object was unexpectedly reinitalized and queued again. This results in a leak of the originally queued request context, potentially performing the new DMA operation twice while dropping the old request, and use after free of the context passed to HalpScatterGatherAdapterControl.
svn path=/trunk/; revision=63898
