May 2005 - Wine-Devel - list.winehq.org

Re: Borland C++ 4.5

by Andrew Neil Ramage

The screen was corrupted - the install was a small window in the corner (ie, it took over the shell window), but when it tried to create a desktop icon (I think) the whole screen became a patchwork of colours. Andrew You can be the captain I will draw the chart Sailing into destiny Closer to the heart Closer to the Heart by Rush (A Farewell to Kings, 1977) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ wino(a)piments.com wrote: > I had a similar situation installing Dragon Nat. Speaking with > installshield. I thought the thing was hanging and infact it was > popping up a modal dlg with a yes/no or whatever and wine was failing > to bring it to the front. > > If the BC installer is hogging the whole screen and you are getting a > similar disfunction in wine you may think it is hung when it is not. > > There is a line in .wine/config that allows you to catch badly behaved > installers in 600x800 window so that you can still access the rest of > the system. > > It's fairly unlikely that an installer will really cripple your kernel > to the point where you have to power-off. > > HTH > > > > On Tue, 31 May 2005 21:53:19 +0200, Kuba Ober <kuba(a)mareimbrium.org> wrote: > >> On Tuesday 31 May 2005 11:29, Andrew Neil Ramage wrote: >> >>> This was a BC++ 4.5 CD I bought years ago. The system I use is SusE 9.2 >>> Desktop (hoping to upgrade to Pro soon) and the latest wine cvs. There >>> was no debugging information available as the system hung, so I could >>> not access the shell window I started the install from. >> >> >> Did you try ssh'ing into the system and executing 'killall X' as >> root? The >> mere fact that X doesn't react to keystrokes doesn't necessarily mean >> that >> the system is hanging. >> >> Cheers, Kuba >> >> > > > > >

May 31, 2005

Re: Borland C++ 4.5

by Uwe Bonnes

>>>>> "Mike" == Mike Hearn <mh(a)codeweavers.com> writes: Mike> On Tue, 31 May 2005 15:53:19 -0400, Kuba Ober wrote: >> Did you try ssh'ing into the system and executing 'killall X' as >> root? The mere fact that X doesn't react to keystrokes doesn't >> necessarily mean that the system is hanging. Mike> If the system no longer responds to keystrokes then it *has* hung Mike> and we should treat fixing as it a a+ priority, regardless of what Mike> is going on under the hood. There's no difference for users Mike> between "X is hung" and "Linux is hung" ... There are some ways that wine can make Linux unusable, like: - A modal box hogging the Display in the foreground - Some error triggering the rebuilding of the Font database . As X is single threaded, this will hog X ... We need more input... -- Uwe Bonnes bon(a)elektron.ikp.physik.tu-darmstadt.de Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt --------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

May 31, 2005

Re: Borland C++ 4.5

by Mike Hearn

On Tue, 31 May 2005 15:53:19 -0400, Kuba Ober wrote: > Did you try ssh'ing into the system and executing 'killall X' as root? The > mere fact that X doesn't react to keystrokes doesn't necessarily mean that > the system is hanging. If the system no longer responds to keystrokes then it *has* hung and we should treat fixing as it a a+ priority, regardless of what is going on under the hood. There's no difference for users between "X is hung" and "Linux is hung" ...

May 31, 2005

Re: Borland C++ 4.5

by Andreas Mohr

Hi, On Tue, May 31, 2005 at 10:31:33PM +0200, wino(a)piments.com wrote: > It's fairly unlikely that an installer will really cripple your kernel to > the point where you have to power-off. Well, it's fairly likely if you have a program allocating the whole world and even a bit more than that (e.g. by stack trashing which causes a pointer value to be used as the size to allocate: ugh) with a user that forgot to set a memory limit... (ulimit -v?). Then just have the Linux OOM mechanism fail to kill the misbehaving program properly, and that's it. (this perhaps happened here) Andreas Mohr

May 31, 2005

Re: Borland C++ 4.5

by wino＠piments.com

I had a similar situation installing Dragon Nat. Speaking with installshield. I thought the thing was hanging and infact it was popping up a modal dlg with a yes/no or whatever and wine was failing to bring it to the front. If the BC installer is hogging the whole screen and you are getting a similar disfunction in wine you may think it is hung when it is not. There is a line in .wine/config that allows you to catch badly behaved installers in 600x800 window so that you can still access the rest of the system. It's fairly unlikely that an installer will really cripple your kernel to the point where you have to power-off. HTH On Tue, 31 May 2005 21:53:19 +0200, Kuba Ober <kuba(a)mareimbrium.org> wrote: > On Tuesday 31 May 2005 11:29, Andrew Neil Ramage wrote: >> This was a BC++ 4.5 CD I bought years ago. The system I use is SusE 9.2 >> Desktop (hoping to upgrade to Pro soon) and the latest wine cvs. There >> was no debugging information available as the system hung, so I could >> not access the shell window I started the install from. > > Did you try ssh'ing into the system and executing 'killall X' as root? > The > mere fact that X doesn't react to keystrokes doesn't necessarily mean > that > the system is hanging. > > Cheers, Kuba > >

May 31, 2005

Re: wine/dlls/setupapi parser.c

by Alexandre Julliard

Robert Shearman <rob(a)codeweavers.com> writes: > Alexandre Julliard wrote: > > >Index: wine/dlls/setupapi/parser.c > >diff -u -p wine/dlls/setupapi/parser.c:1.15 wine/dlls/setupapi/parser.c:1.16 > >--- wine/dlls/setupapi/parser.c:1.15 Tue May 31 17:10:06 2005 > >+++ wine/dlls/setupapi/parser.c Tue May 31 17:10:06 2005 > >@@ -852,7 +852,9 @@ static DWORD parse_buffer( struct inf_fi > > struct parser parser; > > const WCHAR *pos = buffer; > > - parser.start = buffer; > >+ if (*pos == 0xfeff) pos++; /* skip Unicode BOM signature */ > >+ > >+ parser.start = pos; > > parser.end = end; > > parser.file = file; > > parser.line = NULL; > > > > This shouldn't be necessary. The Unicode BOM (or Zero Width > Non-Breaking Space) has CT_CTYPE1 of C1_SPACE | C1_BLANK so it should > be skipped just like any other whitespace character. Hmmm I guess you are right. It seemed to help at the time but that may have been because of the various problems with RtlIsTextUnicode. -- Alexandre Julliard julliard(a)winehq.org

May 31, 2005

Re: ntdll.dll: mark dirlink as REPARSE_POINT (v2)

by Detlef Riekenberg

Am Dienstag, den 31.05.2005, 00:35 -0400 schrieb Dimi Paun: > On Tue, 2005-05-31 at 06:21 +0200, Detlef Riekenberg wrote: > > - if (S_ISDIR( st.st_mode ) && !show_dir_symlinks) return NULL; > > + if (S_ISDIR( st.st_mode )) > > + { > > + if (!show_dir_symlinks) return NULL; > > + name_is_link = FILE_ATTRIBUTE_REPARSE_POINT; > > + } > > } > > Hmm, thinking more about it, if windows has support for dir symlinks > (I'm sure they made sure there are no problems with that), should we Some Informations about a REPARSE_POINT: 1. A REPARSE_POINT is a Symlink or a Mountpoint 2. A REPARSE_POINT in win2000 is used for the "Remote Storrage Service" ans as a Volume Mountpoint (kill the Driveletter-Limit). 3. The Windows-Explorer does not copy the REPARSE_POINT but the complete data, represented by the REPARSE_POINT. 4. The only tool in win2000, which is able convert a normal Directory into a REPARSE_POINT (mountpoint) is the Disk-Management Utility. a: According to "b", there is the "linkd"-app in the w2k-Reskit for creating junctions. b: With "Junction.exe" you can view REPARSE_POINTs and create/delete dir-symlinks ("http://www.sysinternals.com/files/jnctnsrc.zip") (Freeware with source) c: You can create Hardlinks for Files on an NTFS-Volume with ln.exe from "www.losoft.de/download/ls-tools.zip" (Freeware with source) d: Another Link-Tool for NTFS: "http://ashedel.chat.ru/source/" (GPL) e: A clone of ntifs.h with struct FILE_LINK_INFORMATION, struct _REPARSE_DATA_BUFFER and also "*CreateSymbolicLink*" "http://www.acc.umu.se/~bosse/ntifs.h" (GPL) (Also in the mingw-cvs) > even bother having this show_dir_symlinks option? Maybe we should just > nuke it, and just have: I think, we should not do that, before the Rest of a "REPARSE_POINT" is implemented in wine (Create/Info). > On a different note, are the brackets above indented properly? In v2, I used MC. The Tabsize is 8 instead of 4. -- By By ... ... Detlef

May 31, 2005

Search, and screenshots [Was Re: Half Life 2 error -- err:mmio:MMIO_ParseExtA]

by J. Grant

>> Also if there is already a document on setting up Half Life2 could >> someone direct me to it please, I could not find one already. > > I'm not sure about your specific error, but I was able to get HL2 > to start with Oliver Stieber's d3d patches. However, it was far from > playable for me. Models were missing, and there were other glitches. > I made a web site about it, but nobody has contacted me since then > to help debug. For info, please see appdb > (http://appdb.winehq.org/appview.php?versionId=2890) Great! the appdb.winehq.org search seems to have a problem when I enter "half life 2", it does not match that against "half-life-2" I had previously looked but not found your work. Could that search functionality be revised so it supports non 100% perfect matches? The non-perfect-match search could simply be a case-insensitive count of the letters in a name if it was too much work to do checks on fragments of words. Also there is the frustrating way that someone has hidden the screenshots behind javascript pop-up code. Is there any reason these are not simply image links? When clicking on the javascript popup window of the small screenshot the window is closed! If something is clickable I would have expected it to go to the original full size image at least.. The image is also scaled so small that the text in the screenshots is unreadable. Please include my email address in any replies. Kind regards JG -- Homepage: http://jguk.org/ Blog: http://jguk.org/blog.rss Radio: http://jguk.org/#radio

May 31, 2005

Re: Borland C++ 4.5

by Kuba Ober

On Tuesday 31 May 2005 11:29, Andrew Neil Ramage wrote: > This was a BC++ 4.5 CD I bought years ago. The system I use is SusE 9.2 > Desktop (hoping to upgrade to Pro soon) and the latest wine cvs. There > was no debugging information available as the system hung, so I could > not access the shell window I started the install from. Did you try ssh'ing into the system and executing 'killall X' as root? The mere fact that X doesn't react to keystrokes doesn't necessarily mean that the system is hanging. Cheers, Kuba

May 31, 2005

Re: wine/dlls/setupapi parser.c

by Robert Shearman

Alexandre Julliard wrote: >Index: wine/dlls/setupapi/parser.c >diff -u -p wine/dlls/setupapi/parser.c:1.15 wine/dlls/setupapi/parser.c:1.16 >--- wine/dlls/setupapi/parser.c:1.15 Tue May 31 17:10:06 2005 >+++ wine/dlls/setupapi/parser.c Tue May 31 17:10:06 2005 >@@ -852,7 +852,9 @@ static DWORD parse_buffer( struct inf_fi > struct parser parser; > const WCHAR *pos = buffer; > >- parser.start = buffer; >+ if (*pos == 0xfeff) pos++; /* skip Unicode BOM signature */ >+ >+ parser.start = pos; > parser.end = end; > parser.file = file; > parser.line = NULL; > This shouldn't be necessary. The Unicode BOM (or Zero Width Non-Breaking Space) has CT_CTYPE1 of C1_SPACE | C1_BLANK so it should be skipped just like any other whitespace character. -- Rob Shearman

May 31, 2005