Patch for win32k/eng/gradient.c

[Stephen Hodges]

This fixes the build error for building with optimized = 1 , DBG =1

Stephen Hodges

Cell: [phone]

AIM: theteofscuba

Web: http://www.thespoke.net/MyBlog/stevo/MyBlog.aspx

[WaxDragon]

Well this bugchecks ros on boot for me.

DBG=1
optimized=0

(NTDLL:ldr/utils.c:1205) LdrGetExportByName(): failed to find auxMessage
(NTDLL:ldr/utils.c:1205) LdrGetExportByName(): failed to find mxdMessage
(NTDLL:ldr/utils.c:1205) LdrGetExportByName(): failed to find midMessage
(NTDLL:ldr/utils.c:1205) LdrGetExportByName(): failed to find widMessage
(NTDLL:ldr/utils.c:1205) LdrGetExportByName(): failed to find wodMessage
(io/create.c:90) Parent is a Directory which is neither a file type nor a device
 type
CmdLine: <setup -newsetup>
Loaded 'syssetup'!
(devinst.c:930:SetupDiInstallClassW) Incomplete function.
(devinst.c:930:SetupDiInstallClassW) Incomplete function.
Assertion 'pboDest && psoDest && pco' failed at eng/gradient.c line 532
KeBugCheckWithTf at ke/catch.c:180
Bug detected (code 1e param 0 0 0 0)
  KMODE_EXCEPTION_NOT_HANDLED

Breakpoint Exception: 3(0)
Processor: 0 CS:EIP 8:c0005e34 <ntoskrnl.exe: 5e34 (ke/i386/brkpoint.c:46 (DbgBr
eakPoint))>
cr2 77e85b9b cr3 2bdc000 Proc: c0770728 Pid: 6 <setup.ex> Thrd: c077c7a0 Tid: 2a
DS 10 ES 10 FS 30 GS 10
EAX: 00000048   EBX: 00000010   ECX: 00000000
EDX: 000003f8   EBP: dde37c34   ESI: 0062f428   ESP: dde37bc0
EDI: dde37d74   EFLAGS: 00000286 kESP dde37bc0 kernel stack base dde35000
Frames: <ntoskrnl.exe: 9532a (error.c:853 (RtlAssert))> <win32k.sys: c293 (eng/g
radient.c:534 (IntEngGradientFill))> <win32k.sys: 47fa3 (objects/fillshap.c:1435
 (IntGdiGradientFill))> <win32k.sys: 481e2 (objects/fillshap.c:1521 (NtGdiGradie
ntFill))> <ntoskrnl.exe: 3ad7 (/tmp/ccck20Pr.s:156 (KeUserModeCallback))> <gdi32
.dll: 8656 (misc/win32k.c:175 (NtGdiGradientFill))> <gdi32.dll: bf37
(objects/fillshap.c:137 (GdiGradientFill))> <user32.dll: 3a680
(windows/nonclient.c:1233 (D
rawCaption))> <user32.dll: 38838 (windows/nonclient.c:438 (DefWndNCPaint))> <use
r32.dll: 3919b (windows/nonclient.c:663 (DefWndNCActivate))> <user32.dll: 1ab9b 
(windows/defwnd.c:958 (User32DefWindowProc))> <user32.dll: 1b8f3 (windows/defwnd
.c:1549 (DefWindowProcW))> <user32.dll: 20a31 (windows/dialog.c:1549 (DefDlgProc
W))> <user32.dll: 159fc (windows/message.c:926 (IntCallWindowProcW))> <user32.dl
l: 16a01 (windows/message.c:1830 (User32CallWindowProcFromKernel))> <ntdll.dll: 
bd22 (rtl/callback.c:35 (KiUserCallbackDispatcher))> <0> <user32.dll: 119f8 (mis
c/win32k.c:525 (NtUserShowWindow))> <user32.dll: 183a3 (windows/window.c:1287 (S
howWindow))> <user32.dll: 1f1e0 (windows/dialog.c:831 (DIALOG_CreateIndirect))> 
<user32.dll: 20b3a (windows/dialog.c:1607 (DialogBoxIndirectParamW))> <comctl32.
dll: 3301d (:711 (PROPSHEET_CreateDialog))> <comctl32.dll: 3669f (:2838 (Propert
ySheetW))> <syssetup.dll: 3c77 (wizard.c:1418 (InstallWizard))> <syssetup.dll: 1
a45 (install.c:412 (InstallReactOS))> <setup.exe: 1627 (setup.c:157 (RunNewSetup
))> <setup.exe: 16d7 (setup.c:188 (WinMain))> <setup.exe: 2cdb (:200 (WinMain))>
 <setup.exe: 120d> <setup.exe: 1258> <kernel32.dll: 32359 (:338 (BaseProcessStar
t))> <DEADBEEF>

[WaxDragon]

Clarification: It was a clean install, so it was the second stage
installer that was triggering it.  I checked adding it into a working
install, opening the run dialog triggers the bugcheck. Not boot
related. ;0)

[Royce Mitchell III]

Sorry about that, guys...

I've been adding assertions in win32k where they're needed, and I got a 
bit too carried away with that one. It should be fixed now.

[Royce Mitchell III]

Mike Nordell and I have started a page on the wiki listing issues with 
the 0.2.5 release that downloaders/user should be aware of (if they're 
not deemed to be show-stoppers). If we don't warn our users of the 
issues we know about that they are definitely going to experience, we 
may end up spending a lot more energy covering the ground later.

Anyways, here's the url we hijacked: 
http://mok.lvcm.com/cgi-bin/reactos/roswiki?KnownProblems

--
Royce Mitchell III

[Royce Mitchell III]

http://reactos.com/bugzilla/show_bug.cgi?id=32

fixed... I think?

[Jason Filby]

Hi

Aren't we duplicating reactos.com/bugzilla here?

Cheers
Jason

On Fri, 24 Dec 2004 16:51:55 -0600, Royce Mitchell III [e-mail] wrote:
> Mike Nordell and I have started a page on the wiki listing issues with
> the 0.2.5 release that downloaders/user should be aware of (if they're
> not deemed to be show-stoppers). If we don't warn our users of the
> issues we know about that they are definitely going to experience, we
> may end up spending a lot more energy covering the ground later.

[Mike Nordell]

Jason Filby wrote:
> Aren't we duplicating reactos.com/bugzilla here?

in respose to
> Royce Mitchell III wrote:
> > Mike Nordell and I have started a page on the wiki listing issues with
> > the 0.2.5 release ...

To which I respond:

Yes. But since I seemingly couldn't create a bugzilla account no matter what
I did, I thought it was better to follow Royce's suggestion of storing my
findings somewhere (wiki) rather than nowhere.

/Mike

[Gge]

Mike Nordell wrote:
> Jason Filby wrote:
> > Aren't we duplicating reactos.com/bugzilla here?
>
> in respose to
> > Royce Mitchell III wrote:
> > > Mike Nordell and I have started a page on the wiki listing issues with
> > > the 0.2.5 release ...
>
> To which I respond:
>
> Yes. But since I seemingly couldn't create a bugzilla account no matter what
> I did, I thought it was better to follow Royce's suggestion of storing my
> findings somewhere (wiki) rather than nowhere.
>
> /Mike

The bugs should be reported to Bugzilla  (normally) in my opinion.
I still do not know when bug may be created for users ? During "RC" 
phase only ?

Regards

[Ge van Geldorp]

> From: Gge
>
> The bugs should be reported to Bugzilla  (normally) in my 
> opinion. I still do not know when bug may be created for 
> users ? During "RC" phase only ?

No, please feel free to report bugs at any time.

Ge van Geldorp.

[Mike Nordell]

Gge wrote:
> The bugs should be reported to Bugzilla  (normally) in my opinion.

Your opinion is mine. If we now had a workin bz account creation it would
not be an issue...

To spell it out:
ROS BZ is fscked up, and I can't even create an accout to file bugs.

/Mike

[Steven Edwards]

--- Mike Nordell [e-mail] wrote:
> To spell it out:
> ROS BZ is fscked up, and I can't even create an accout to file bugs.

Don't create a account on the bugzilla page. Create one on www.reactos.org and it will work
with Bugzilla. I just did it for user ac.

http://reactos.com/bugzilla/show_bug.cgi?id=459

There are some bugs and I have noted them here.

Thanks
Steven

__________________________________ 
Do you Yahoo!? 
Send a seasonal email greeting and help others. Do good. 
http://celebrity.mail.yahoo.com

[Mike Nordell]

Steven Edwards wrote:
> Don't create a account on the bugzilla page.

Please don't go there again. T has been notified IIRC.
> Create one on www.reactos.org and it will work
> with Bugzilla. I just did it for user ac.

Good for you. It failed for me.

Whatever. I'm in too much (physical) pain to even bother. Bloody
christianity weeekend with broken sholders in the slippery ice&snow. Merry
Sunday. and wiki..

/Mike

[Royce Mitchell III]

Jason Filby wrote:
> Hi
>
> Aren't we duplicating reactos.com/bugzilla here?
>
> Cheers
> Jason

Yes, sorta, but..

1) account creation for bugzilla is broken.

2) this is supposed to be a list of specific issues with 0.2.5 that 
we're likely to get a lot of support issues on.

3) bugzilla seems quite out of date. Everybody needs to get in there and 
clean it up.

Perhaps once bugzilla gets cleaned up we could use it to generate "known 
issues" lists for each release.
> On Fri, 24 Dec 2004 16:51:55 -0600, Royce Mitchell III [e-mail] wrote:
> > Mike Nordell and I have started a page on the wiki listing issues with
> > the 0.2.5 release that downloaders/user should be aware of (if they're
> > not deemed to be show-stoppers). If we don't warn our users of the
> > issues we know about that they are definitely going to experience, we
> > may end up spending a lot more energy covering the ground later.
>
> Ros-dev mailing list
> [e-mail]
> http://reactos.com:8080/mailman/listinfo/ros-dev
>
> .