Re: Various problem with xconq

[Martin Quinson <Martin dot Quinson at ens-lyon dot fr>]

On Thu, Jan 31, 2002 at 01:53:32AM +0100, Hans Ronne wrote:
> >Hello,
> >I just downloaded the 7.4.1 tarball, and I'm trying to compile it. Sadly, I
> >run in a lot of problems. And I also found problem reported in the Debian
> >bug tracking system I want to tell you.
> 
> Nice to see that somebody still cares about xconq. This list has been
> almost dead for the last few months.

In fact, I come from the Debian world, and the problem is the same their.
The last packaged version is 7.2.2 (belong the package I did last night ;)
 
> >I can't manage to build the documentation in ps format.
> >Ohhh, well. What am I supposed to do now ?
> 
> You can always use the html version of the manual which is enclosed with
> the Mac binaries. This texinfo stuff can really give you gray hairs.

Mmm. If so, could you also include it in the source tarball ? I'm not
willing to fetch it for the Debian package, for example.
 
> >--------------------------------------------------------------------------
> >Beside of that, I noticed a whole bench of gcc warning like
> >oplayer.c:1129: warning: chance1' might be used uninitialized in this
> >function
> >don't you get afraid that the code may be wrong in these places ?
> 
> I checked the code and the warning is correct, i.e. chance1 is not
> initialized in some cases. Not sure if it ever happens, but the possibility
> exists. Never saw this warning in CodeWarrior, so it must be more tolerant
> than gcc. If you mail me the other warnings you got, I can look into it.

I'll send you the whole output of the build, but privatelly, it's pretty big.

> >--------------------------------------------------------------------------
> >I also noticed that the config.guess is of version "2000-05-30" and
> >config.sub of version "2000-06-10". That's kinda problematic for new
> >architectures (see http://bugs.debian.org/78672 for more details)
> 
> The configuration files for various unix architectures is something I never
> messed with. I have just been happy if xconq compiles under LinuxPPC, which
> it does. Stan used to handle this stuff, but he has not been active for
> almost a year. It would be great if somebody who is familiar with different
> unix architectures could take a look at this, if it is a problem.

The problem is really easy to solve: just include the last version of this
two files. If you want to, I can send them to you..

> >--------------------------------------------------------------------------
> >I also found this patch in the Debian bug tracking system:
> >http://bugs.debian.org/107313
> >The author claims that without that, the patch is necessary to compile on
> >the GNU/Hurd architecture, but I've no way to figure out if that's true.
> 
> Same answer here.

Well, just check it out. It's pretty straighforward:
| xconq still uses sys_errlist.  strerror has superceded this and
| thus GNU HURD only has strerror().  This patch updates xconq to use
| strerror() to compile and run xconq on GNU HURD

And the patch seems raisonable, and is less than 100 lines.

> >--------------------------------------------------------------------------
> >Please also note that the page
> >http://sources.redhat.com/xconq/lists.html
> >is wrong. It points to xconq7@sourceware.cygnus.com where it should be
> >xconq7@sources.redhat.com
> 
> This is something that only Stan can fix. It's his web site.

Ok.

> >--------------------------------------------------------------------------
> >Last point, there is a major security issue in the parameter handeling. That
> >was reported on bugtraq more than one year ago, and a patch were contributed
> >a short while ago. See
> >http://bugs.debian.org/80576
> >for more details.
> 
> This patch seems straightforward. I will see to it that it goes into the
> CVS sources.

Good. Thank you. Please let me know.

> >I really like this program, and I hope you will concider all the issues I
> >raised in this mail.
> 
> Thanks for your comments. I will fix those problems that I can handle.

I'm really happy from this rapid answer. I'm subscribed to the list for a
few month, and I was fearing that the project is dead. I'm happy to see the
contrary !


Thanks, Mt.