On Thu, Oct 05, 2006 at 08:15:28PM +0200, Petr Salinger <[EMAIL PROTECTED]> 
wrote:
> >I forgot to send some feedback here, sorry.
> >The 18_kbsd_nsprpub.dpatch part seems fine,
> 
> Thanks.
> 
> >but I don't get some of the changes of the 88_kbsd.dpatch,
> 
> Which one's ?
> 
> They are mainly
> "defined(LINUX)"  -> "defined(LINUX) || defined(__GLIBC__)"
> 
> The notable exception is "security/coreconf/Linux.mk".
> Please take a look on the result, it also simplifies this file.
> 
> In all cases, it should not change behaviour on Linux platforms.

Sure, but I prefer to apply patches that change only what is necessary
to change for their purpose, especially on this file, which is patched
by other patches.

> >nor why they have to be debian specific.
> 
> In long term, they should not.

Then it should have been named 68_kbsd ;)

> But, in fact, 18_kbsd_nsprpub + 88_kbsd are also sufficient to build 
> firefox, but the firefox hangs for me in the first window.
> It might be due to really outdated boehm-gc, but I don't know (yet) the 
> real reason.

Does xulrunner hang the same way ? IIRC, boehm-gc is not used. It is
more likely to be an xptstubs/invoke problem. Which implementation gets
built in xpcom/reflect/xptcall/src/md/unix/ ? If it is
*_unsupported.cpp, then the resulting binary is pretty useless.

> So, I do not consider the patch finished, but in short term we 
> (kfreebsd) need libxul-dev due to reverse build-depends (such as gcj-4.1, 
> classpath, ...)
> 
> Please could you add it in mean time as debian specific ?

Not until I'm sure it give actually something that works. You can test
the mybrowser example [1] if you need an actual full test.

Mike

1. http://benjamin.smedbergs.us/xulrunner/mybrowser-0.2.2.xulapp
  unzip the given file (it's just a zip file), and run:
  xulrunner application.ini


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to