September 2002 - Wine-Devel

Re: Using winbind with Wine

by Tim Potter

On Thu, Sep 26, 2002 at 01:47:38PM +0200, Martin Wilck wrote: > To initiate this process we'd "only" need a standardized protocol for > the socket communication. Andrew said that doesn't exist and won't with > regard to winbind. I'd like to focus the discussion in this direction. > > - is the winbind team willing to standardize the protocol, or at least > ensure backward compatibility in future versions? There is a LGPL client library (well library is probably doing it more justice than it deserves) called wb_client.c which is used in the NSS modules that talk to winbindd. I would think that would be a preferable way of talking to winbindd rather then rewriting code to talk the winbindd protocol. > - is the winbind team willing to add more RPC calls to the interface? Not unless they are related to authentication or user/group enumeration. > "winebind" would be linked against Samba libraries, and therefore be GPL > from the start. Heh - cute name. (-: Tim.

Sept. 26, 2002

Re: winedbg is not finding debug symbols, anybody know what I'm doing wrong?

by Greg Turner

Eric POUECH wrote... > but I assume you double checked that more likely, aren't you using gcc 3.2 > by any chance ? if so, you need to force ELF symbol emission, not > Dwarf II which became default in 3.2 Bingo. Guilty as charged. Thank you SO MUCH for this, I should have figured gcc3.2 was the culprit, that thing is feisty! I'll take a look at whether I can find a generic solution to this and post a patch. Now that you mention it, I think gcc3.2 has some other interesting warnings, something about some alignment thingys not supported anymore (probably I just change "-m$X" to "-f$X" to fix), I'll see about those too, could well be important... -gmt -- "Waiting periods are only a step. Registration is only a step. The prohibition of private firearms is the goal." -U.S. Attorney General Janet Reno, December 1993

Sept. 26, 2002

Re: [Controversial] Developper only CVS sandbox

by Eric POUECH

> >Some systems may be better than CVS at this, admitted. > >But as long as there are only a few branches and people keep their stuff > >more or less in sync with main, I think it'd be fine - especially for > >people working on relatively separate subtrees like D3D8. just to react on Alexandre flame war proposal. what about bitkeeper ? A+

Sept. 26, 2002

Re: winedbg is not finding debug symbols, anybody know what I'm doing wrong=?iso-8859-1?Q?=3F ?=

by Eric POUECH

> Whenever I run winedbg (or when wine runs it in response to an exception), > it fails to find debug symbols. Looking at disassembled code is making my > brain numb. I "export STRIP=false" before configure to ensure that the > Makefile doesn't strip anything, and gdb finds everything just fine. So what > am I doing wrong? so you compiled and install wine yourself. did you install it in /usr/local or /usr/ ? (I just wonder if you don't have two sets of wine, one in /usr/local the other one in /usr). but I assume you double checked that more likely, aren't you using gcc 3.2 by any chance ? if so, you need to force ELF symbol emission, not Dwarf II which became default in 3.2 A+

Sept. 26, 2002

How to tell winemaker and winebuild what I want to tell it

by Bill Medland

Anyone able to tell me what I am doing wrong here before I spend days trying to figure it out? My winelib wrapper wants to call MultiByteToWideChar and DOSFS_GetFullName. As far as I am aware they are available from ntdll.dll (ntdll.dll.so) So I want to to get the build to "link" to ntdll.dll I thought it should simply require "-intdll" in the argument to winemaker, but during the link I still get complaints about not finding those two functions. What am I doing wrong? Bill p.s. Thanks to Francois for the fixes to Winemaker

Sept. 26, 2002

dlls/xxx/xxx_main.c

by Dimitrie O. Paun

Can we please rename all these dlls/xxx/xxx_main.c to dlls/xxx/xxx.c It should be possible now, right? Should I send a patch in? -- Dimi.

Sept. 26, 2002

Re: [patch] Listview error handling

by Dimitrie O. Paun

On September 26, 2002 12:56 pm, Paul Rupe wrote: > This fixes a problem in my previous listview patch reported by Rein Klazes. > > Changelog: > Paul Rupe <prupe(a)myrealbox.com> > Remove freed sublist on failure in LISTVIEW_InsertItemT > + HDPA hdpaSubItems = NULL; I don't understand what you want to do. Why is this needed? Why did you change the code in the first place? -- Dimi.

Sept. 26, 2002

Re: wine/dlls/comctl32 listview.c

by Paul Rupe

On 26 Sep 2002, dpaun(a)rogers.com wrote: > Of course it does. Damn, I was tracking this bug myself, but I > thought my changes introduced them. :) I'll look into it. But > why the change in the first place??? It was my fault, sorry. I changed it because I had a crash where one of those steps failed, but it left the freed pointer in the array to be used later. I was hoping simply reordering things would prevent that, but I guess I missed a dependency. I submitted another patch that puts the order back the way it was and does the extra work in the cleanup instead. -- Paul Rupe "She smiled, in the end." p r u p e @ m y r e a l b o x . c o m | Oppose government police-ware on your PC! | Stop the Consumer Broadband and Digital Television Promotion Act! | <http://www.eff.org/alerts/20020322_eff_cbdtpa_alert.html>

Sept. 26, 2002

Re: Using winbind with Wine

by Luke Howard

>Because DCE/RPC is *horrible* ? :-) :-). If you need a new >RPC protocol please use ONC/RPC/NDR not DCE :-). I agree, I think I've grown a few gray hairs from DCE. I guess the argument in favour of it depends on how similar the kinds of lookups WINE needs to make are to existing, implemented, RPCs. If there isn't much overlap then yes, choosing DCE RPC doesn't buy you much in terms of reduced amounts of code on the server. -- Luke -- Luke Howard | PADL Software Pty Ltd | www.padl.com

Sept. 26, 2002

Re: Using winbind with Wine

by jra＠dp.samba.org

On Fri, Sep 27, 2002 at 12:50:57AM +1000, Luke Howard wrote: > > > - is the winbind team willing to standardize the protocol, or at least > > ensure backward compatibility in future versions? > > Rather than inventing new protocols, why not just use DCE RPC over domain > sockets or TCP/IP? > > The only catch is that you need a DCE RPC client library. We're using the > OSF DCE runtime (actually, FreeDCE), which is BSD-licensed. Because DCE/RPC is *horrible* ? :-) :-). If you need a new RPC protocol please use ONC/RPC/NDR not DCE :-). Jeremy.

Sept. 26, 2002