Ros-dev

ros-dev@reactos.org

1 participants
5793 discussions

Testing Coordinator, position open November 22.
by James Tabor 17 Oct '06

17 Oct '06

Testing Coordinator, position open November 22. Disapproved! You are going to work 90 more days due to that GOD! DAMNED! AUDIT! Thank you for your service, Management

3 2

Bug 1819
by Peter Quiring 14 Oct '06

14 Oct '06

I've fixed bug 1819 and added a patch to bugzilla, can someone verify and apply the patch. It works for me. Thanks. Peter Quiring

1 0

svn.reactos.org/iso no longer works...
by Richard 10 Oct '06

10 Oct '06

http://svn.reactos.org/iso no longer has the latest isos. Last revision was 24295.

2 1

Re: [ros-dev] [ros-diffs] [gedmurphy] 24383: add LPCDLGTEMPLATE definition
by Reuel ben Yisrael 03 Oct '06

03 Oct '06

gedmurphy(a)svn.reactos.org wrote: >============================================================================== >--- trunk/reactos/include/psdk/winuser.h (original) >+++ trunk/reactos/include/psdk/winuser.h Wed Oct 4 02:43:46 2006 >@@ -2631,6 +2631,13 @@ > short cy; > } DLGTEMPLATE,*LPDLGTEMPLATE,*LPDLGTEMPLATEA,*LPDLGTEMPLATEW; > typedef const DLGTEMPLATE *LPCDLGTEMPLATE; // <---- first >+typedef const DLGTEMPLATE *LPCDLGTEMPLATEA; >+typedef const DLGTEMPLATE *LPCDLGTEMPLATEW; >+#ifdef UNICODE >+typedef LPCDLGTEMPLATEW LPCDLGTEMPLATE; // <---- second (unicode) >+#else >+typedef LPCDLGTEMPLATEA LPCDLGTEMPLATE; // <---- second (ansi) >+#endif // UNICODE > #pragma pack(pop) > typedef struct tagDRAWITEMSTRUCT { > UINT CtlType; > > does this really compile? LPCDLGTEMPLATE appears to be duplicately typedef'd. ( see added comments above )

1 0

Re: [ros-dev] [ros-diffs] [hpoussin] 24379: Copy i8042prt driver from 0.3.1 branch to trunk. Try #2
by Alex Ionescu 03 Oct '06

03 Oct '06

hpoussin(a)svn.reactos.org wrote: > Author: hpoussin > Date: Tue Oct 3 22:04:03 2006 > New Revision: 24379 > > URL: http://svn.reactos.org/svn/reactos?rev=24379&view=rev > Log: > Copy i8042prt driver from 0.3.1 branch to trunk. Try #2 > > Added: > trunk/reactos/drivers/input/i8042prt/ > - copied from r24378, branches/ros-branch-0_3_1/reactos/drivers/input/i8042prt/ > > Why are we copying it from branch to trunk? Shouldn't it be copied from trunk to the branch? -- Best regards, Alex Ionescu Project Lead, TinyKRNL Kernel-Mode Software Design Engineer, ReactOS

3 3

Build Hive Error
by James Tabor 02 Oct '06

02 Oct '06

Rev 24339 with gcc 3.4.2, bin utils 2.15.94 [COPY] reactos/system32/win32k.sys [CC] tools/mkhive/binhive.c include/psdk/intrin.h: In function `_InterlockedCompareExchange8': include/psdk/intrin.h:193: error: output operand constraint lacks `=' include/psdk/intrin.h: In function `_InterlockedCompareExchange16': include/psdk/intrin.h:200: error: output operand constraint lacks `=' include/psdk/intrin.h: In function `_InterlockedCompareExchange': include/psdk/intrin.h:207: error: output operand constraint lacks `=' include/psdk/intrin.h: In function `_InterlockedCompareExchange64': include/psdk/intrin.h:236: error: output operand constraint lacks `=' include/psdk/intrin.h: In function `_InterlockedCompareExchangePointer': include/psdk/intrin.h:250: error: output operand constraint lacks `=' include/psdk/intrin.h: In function `_InterlockedExchange': include/psdk/intrin.h:257: error: output operand constraint lacks `=' include/psdk/intrin.h: In function `_InterlockedExchangePointer': include/psdk/intrin.h:264: error: output operand constraint lacks `=' include/psdk/intrin.h: In function `_InterlockedExchangeAdd': include/psdk/intrin.h:271: error: undefined named operand 'retval' include/psdk/intrin.h:271: error: undefined named operand 'Addend' include/psdk/intrin.h:271: error: output operand constraint lacks `=' include/psdk/intrin.h: In function `__writecr0': include/psdk/intrin.h:1009: error: undefined named operand 'value' include/psdk/intrin.h: In function `__writecr3': include/psdk/intrin.h:1014: error: undefined named operand 'value' include/psdk/intrin.h: In function `__writecr4': include/psdk/intrin.h:1019: error: undefined named operand 'value' make: *** [obj-i386/tools/mkhive/binhive.o] Error 1

2 2

Actual NTLDR boot screenshot
by Alex Ionescu 02 Oct '06

02 Oct '06

Sometimes pictures speak better then commit messages, so here's the current boot status: http://www.tinykrnl.org/ros_by_ntldr.png It fails because our Memory Manager early-boot initialization depends on some freeldr information. If I can/have time, I'll try to make it not depend on them, and/or use NTLDR structures instead. Anyways, serial debug printing works and the GDT, IDT, TSS, KPCR are all setup, so this can now be used as a good tool for blackboxing ntldr. -- Best regards, Alex Ionescu Project Lead, TinyKRNL Kernel-Mode Software Design Engineer, ReactOS

2 1

Working trunk?
by Alex Ionescu 30 Sep '06

30 Sep '06

Hi, I took a break from the project hoping any regressions I might've caused could be found/fixed... isntead, I now find myself with an OS that won't recognize my mouse/keyboard. I've tried rebooting it to re-start 2nd stage (which used to work before), but all this does is gradually damage the registry more and more, either leading to aSSERTs or freeldr codepage error. This is with a 100% clean trunk on QEMU 0.8.2. -- Best regards, Alex Ionescu Project Lead, TinyKRNL Kernel-Mode Software Design Engineer, ReactOS

1 0

Time date control panel applet bug and DrawStatusText issues
by Felipe Villarroel 28 Sep '06

28 Sep '06

Hello list!, i have two questions that i hope that someone of you could answer. The first one: - Should RegOpenKeyExW() do a SetLastError() or it just return an error code? I ask this because in the timedate control panel applet, GetSyncSetting() reads a key from the registry using that function and if it fails it calls the function GetError() that calls FormatMessage() using GetLastError() as an argument. For some reason in my computer this call fails and i receive an window that says "ERROR_INVALID_WINDOW_HANDLE" that obviously is not the right error code. This bug could be reproduced by double clicking in the Hour located in the right corner of the start menu bar. Next you click the Internet Time tab and you should receive the error. To fix this i can think of two possibilities: - Modify RegOpenKeyExW() to do a SetLastError() if this fails to set a right code for GetLastError(), or - Modify the GetError() function in the timedate control panel applet I don't know which is the right way to fix this. Finally in my opinion, if an error occurs in this situation i think that no message should appear because i think that is not too bad if this key isn't present. And the second question follows... - I have written a program that displays a Status bar using the DrawStatusText() function. In Windows it is working pretty well but in ReactOS it doesnt. DrawStatusText() in ReactOS doesnt draw a background when calling this function thus my status window draws over itself each time i receive a WM_PAINT message. So its clear that in windows DrawStatusText() draws a background before painting the text, but in ReactOS it doesnt. Currently i have hacked the DrawStatusText() function to draw a background like this: void WINAPI DrawStatusTextW (HDC hdc, LPRECT lprc, LPCWSTR text, UINT style) { RECT r = *lprc; UINT border = BDR_SUNKENOUTER; HBRUSH hBkBrush, hOldBrush; if (style & SBT_POPOUT) border = BDR_RAISEDOUTER; else if (style & SBT_NOBORDERS) border = 0; DrawEdge (hdc, lprc, border, BF_RECT|BF_ADJUST); /* now draw text */ if (text) { int oldbkmode = SetBkMode (hdc, TRANSPARENT); UINT align = DT_LEFT; if (*text == L'\t') { text++; align = DT_CENTER; if (*text == L'\t') { text++; align = DT_RIGHT; } } r.left += 3; if (style & SBT_RTLREADING) FIXME("Unsupported RTL style!\n"); // yep this fixed the problem, but i don't know what the // right color it is hBkBrush = GetSysColorBrush(GetSysColor(COLOR_BTNFACE)); hOldBrush = SelectObject(hdc, hBkBrush); FillRect(hdc, lprc, hBkBrush); DrawTextW (hdc, text, -1, lprc, align|DT_VCENTER|DT_SINGLELINE|DT_NOPREFIX); SelectObject(hdc, hOldBrush); SetBkMode(hdc, oldbkmode); } } Attached i send the code that shows this "bug" and a precompiled binary. Also i have tested this program under WINE with the same results (0.9.x i think). Thats all, Best regards, Felipe Villarroel -- _______________________________________________ Get your free email from http://www.linuxmail.org Powered by Outblaze

2 1

Re: [ros-dev] [ros-diffs] [hpoussin] 24286: Support mount manager in disk.sys
by Alex Ionescu 28 Sep '06

28 Sep '06

hpoussin(a)svn.reactos.org wrote: > Author: hpoussin > Date: Thu Sep 28 20:33:12 2006 > New Revision: 24286 > > URL: http://svn.reactos.org/svn/reactos?rev=24286&view=rev > Log: > Support mount manager in disk.sys > > Modified: If you're going to add this kind of support, I'd appreciate it if you took a look at TinyKRNL/WDK to see how this should be done. FTDISK is responsible for all the code you've added, not disk.sys, and if we're going to implement NT5+ Storage stack features on top of an NT4 driver, this will mean either one of these two things: 1) A crappy "ROS" stack that's incompatible with anything NT offers (and the storage stack is heavily documented in NT, so we're at a big loss if we do this). 2) Someone will have to eventually re-write even more code to make it compatible. I suggest you work on implementing an XP/2K3 style classpnp driver first. The WDK and TinyKRNL have sample source code (non-reversed), and there's 50 pages of documentation on it. Once that works, a new disk.sys can be written (same information as above is available). Then ftdisk should follow. It handles all Fault-Tolerance stuff, but about 6000 lines of it are responsible for the glue layer between the storage stack. TinyKRNL has just finished implementing it, so the information and documentation located there should be helpful (using TinyKRNL documentation and code for learning purposes is fully in-line with our IP policy and the whole point of the project). Then partmgr, mountmgr can be written. These are also implemented in TinyKRNL completely. (Except GPT support). Finally, some parts of the kernel's IoMgr needs to be changed in the method in which it sets up disk letters, and kernel32 needs some new code (which Filip has on his disk). -- Best regards, Alex Ionescu Project Lead, TinyKRNL Kernel-Mode Software Design Engineer, ReactOS

1 0

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Ros-dev