It is now a wrapper around GdiPolyDraw, which implements the main body of the function, while the Nt function only probes and captures the user mode data into safe kernel mode buffers.
Previously some people thought it was a great idea to just wrap the entire implementation in SEH, so when something throws an exception somewhere for whatever reason, we can just catch it here... and LEAK ALL RESOURCES IN THE PROCESS!
TODO: Rewrite all of this. It's broken everywhere.
@I_Kill_Bugs fix
CORE-11787, CORE-17721 and CORE-19721
For function IntExtTextOutW with space character, the x-dimension should be taken into account.
Fixes HexEdit 1.2.1 right side of display window not being cleared.
Account for x-dimension if TA_UPDATECP flag set and 'String' is not NULL.
Clarify 'etx' is ASCII End of Text
Don't free the linked font name if it's already NULL. Since it is allocated only once (when loading the font name from Registry), it should be freed only once too. All subsequent freeing attempts will result in invalid pool freeing.
This fixes the 0xC2 BAD_POOL_CALLER bugcheck when trying to resize window (and draw more text) from soundcloud.com in K-Meleon 76.5.4 Goanna engine and when switching between channels (during redrawing the text) in DiscordMessenger v1.06 Beta.
Addendum to 0f9e889736.
CORE-19681
Fixes a bug in Midori 0.5.11 installer which caused bad rendering
that was making the text hard to read (no background).
In Windows it's visible as FillRect(0x89011a1f, 0x0019f87c, 0xabababab),
and in ReactOS as SetBkMode(0xb5010fce, -1546139919).
After a number of attempts to get the same behavior, I haven't found a way
to fix it without a hack to make the installer look identical to Windows.
At least I added handling for the incorrect value in SetBkMode.
This shouldn't matter anywhere. It may not look the same as on Windows,
but the text will at least be readable.
If East Asian people were unable
to see the Latin characters, it
becomes a barrier to mutual
understanding.
FontLink will break that barrier.
JIRA issue: CORE-9616
JIRA issue: CORE-15480
- Modify font substitutes.
- Unify the lock variables.
- Add FONTLINK and
FONTLINK_CHAIN structures.
- Add FontLink_Create and
FontLink_Destroy functions.
- Add FontLink_Chain_Init,
FontLink_Chain_Free,
FontLink_Chain_LoadReg,
FontLink_Chain_Populate, and
FontLink_Chain_FindGlyph
functions.
- Implement FontLink.
- Add font file DroidSansFallback.ttf
for LiveCD.
Check system integrity by explicitly
releasing the font engine.
JIRA issue: CORE-9616
- Use RTL_STATIC_LIST_HEAD against
g_FontListHead and
g_FontCacheListHead variables and
omit initialization of them.
- Implement FreeFontSupport function.
- Call FreeFontSupport in win32k!
DriverUnload function.
- Fix some usages of
CONTAINING_RECORD macro.
Improve checks in AlphaBlend and TransparentBlt functions: check whether DCSrc is of DCTYPE_INFO also, to fail in that case properly too.
Spotted by PVS-Studio analysis.
Reference: https://pvs-studio.com/en/blog/posts/cpp/1122/.
Support emoji characters and minor characters.
JIRA issue: CORE-19030
NOTE: Not every emoji is supported yet (No ZWJ support). Surrogate pairs only.
NOTE: This feature only works if the appropriate font is selected.
- Add IS_HIGH_SURROGATE and
IS_LOW_SURROGATE macros.
- Add Utf32FromSurrogatePair helper function
to convert a surrogate pair to a UTF-32 code
point.
- Convert the surrogate pairs in the text in
TextIntGetTextExtentPoint, IntGetTextDisposition,
and IntExtTextOutW functions.
Slight performance improvement.
JIRA issue: CORE-15554
- Add bNoTransform variable.
- Check whether mat is identity matrix.
- If so, then don't do FT_Vector_Transform.
Slight performance improvement. STACK_TEXT_BUFFER_SIZE = 100 was too small
for long text. This PR can reduce unnecessary buffer allocations in NtGdiExtTextOutW
function.
JIRA issue: CORE-15554
- Change STACK_TEXT_BUFFER_SIZE to 512.
The code itself isn't wrong but we're entirely missing the logic needed
to make it work. This leads to spontaneous crashes with video drivers and
various unity games.
When I am implementing the IME soft keyboard (#6021 and #6036),
I noticed an issue with PatBlt function.
- Fix the rectangle coordinates when the value was
negative in NtGdiPatBlt function.
- Fix NC_DrawFrame function.
- Fix UserDrawWindowFrame function.
CORE-19334
According to the results of CImage testcase,
the top byte of the GetPixel() return value is
zero if the return is a valid color.
Do bitwise-AND operation if the color value is valid.
CORE-19008
The code was passing 0 instead of SECTION_INHERIT::ViewUnmap (2). 0 isn't even a proper constant to be used here. It worked, because MmMapViewOfSection only compares against ViewShare (1) and treats everything else as ViewUnmap.
bmBits is only used and assigned on output. It points (holds the address)
to the array of DIB bit values. The "Bits" parameter is however a pointer
to a variable that will receive the address of that array.
So it makes no sense to initially assign bmBits to the value of the Bits
parameter...
I intend to port back the combined work of Thomas Faber and Serge Gautherie in context of CORE 14271.
Both developers fixed wrong retval evaluations for SeSinglePrivilegeCheck() and RtlCreateUnicodeString().
Both functions do return a BOOLEAN, and therefore using NTSTATUS() on them is wrong.
Those bugs have been fixed at multiple places. That is long gone.
But Serge fixed his locations a bit more elegantly, without the need for additional variables.
Therefore this addendum adapts a few of Thomas locations to the improved Serge-ified style.
Yes: I intentionally used a space instead of a minus after the mentioned CORE 14271,
as I don't want that pure stylistic addendum to be linked with the initial ticket anymore.
That would be overkill.
Many thanks for Simone Lombardo for pointing to the code needing attention
and providing a working proof-of-concept solution.
CORE-12377
CORE-18084
CORE-13889
CORE-17932
[ENG] Implement DirectDraw management in switch display mode functions
(e.g. resolution change, color depth, display frequency etc.):
- Switch DirectDraw instances between the two PDEVs (the current one and
the new one allocated by ourselves) by calling dxg!DxDdDynamicModeChange
function.
- Suspend them before and resume after the display mode switch, by calling
dxg!DxDdsuspendDirectDraw and dxg!DxDdResumeDirectDraw appropriately.
We currently don't have these functions implemented, but MS DXG has, so
it allows to properly manage DirectDraw PDEVs using this driver, similarly
to Windows.
My analysis confirms that these functions are always called in XP/2k3 on
display mode switch, even when there is no any DirectX app running at the
moment. Analyzing their prototypes show that my guesses are correct.
- Initialize hDev and dhpdev members for EDD_DIRECTDRAW_GLOBAL for newly
created surfaces, switch them during mode change and re-initialize after
it also. They are commonly used by DirectDraw stack.
In addition, enable DirectDraw for old and new PDEVs, by calling
dxg!DxDdEnableDirectDraw function.
[NTDDRAW] Additionally, fix usage of DirectDraw lock count in the PDEVOBJ
structure.
- Enable cDirectDrawDisableLocks member for storing its value, instead of
DxDd_nCount, which is marked as ROS-specific.
- Use it in win32k!DxEngGet/SetHdevData for getting/setting DirectDraw
count appropriately.
My analysis also shows that in Windows, the PDEVOBJ::cDirectDrawDisableLocks
method calls DxEngGetHdevData with type 8, which corresponds to our DxDd_nCount.
So there are no doubts that this member is used there.
- Rename DxEngGetHdevData_dd_count alias of type 8 to DxEngGetHdevData_dd_locks,
to match more accurately an actual member name. Update the enumeration
and fix all code parts appropriately.
All these changes allow to properly change display mode during executing
DirectDraw applications, when they try to switch in full-screen mode.
At least a bugcheck that happened before my changes, does no longer appear.
There are still some games that don't run correctly, as if there is no
3D acceleration (which actually exists). This requires further investigations.
CORE-18221
Load the DirectX graphics kernel driver (dxg.sys) by win32k at WINSRV
initialization time, in NtUserInitialize(). Keep it always loaded in
memory, as on Windows, instead of loading it only by DirectX dlls.
This fixes the problem of acessing this driver: we need only to call
DxDdEnableDirectDraw() and do other stuff when DirectDraw/Direct3D is
required by anything. In other cases, it is called from win32k PDEV
functions when changing display mode (as in Windows). Since it's used
by other things too, it needs to be always loaded.
Otherwise, if it's not loaded, its APIs are not accessible when needed,
and execution fails.
For example, it fixes display mode change problem in VMWare, when a
new mode fails to be applied. Indeed, when it manages DirectDraw stuff,
it calls DXG routines, and therefore fails if dxg.sys isn't loaded
in memory at this moment.
- Implement InitializeGreCSRSS() initialization routine, that initializes
supplemental NTGDI/GRE data once CSRSS and WINSRV are loaded:
* Call DxDdStartupDxGraphics() inside it, which loads dxg.sys.
* Additionally, move fonts and language ID initialization there, from
win32k!DriverEntry. Confirmed by analysis on Windows.
- Call InitializeGreCSRSS() in NtUserInitialize() main initialization routine
(called by WINSRV initialization).
Moved to NTGDI from previously NTUSER place:
Co-authored-by: Hermès Bélusca-Maïto <hermes.belusca-maito@reactos.org>
Currently, LDEVOBJ_bUnloadImage always returns FALSE due to missing functionality in ntoskrnl.
LDEVOBJ_vFreeLDEV contains an assert that driver is correctly unloaded before freeing it.
Prevent this assert by calling LDEVOBJ_vFreeLDEV only if LDEVOBJ_bUnloadImage succeed.
This can be easily triggered if DrvEnableDriver function (called in LDEVOBJ_bEnableDriver) fails.
This allows DevExpress Ribbon Notepad sample to start.
Also add a fat DPRINT1 suggested by Timo Kreuzer,
because there is a lot wrong in the code.
CORE-18091 CORE-18558
