Re: debian-installer/sarge
On Fri, Aug 18, 2006 at 12:47:53AM +0200, Frans Pop wrote:
> On Friday 18 August 2006 00:24, Christian T. Steigies wrote:
> > Wouters mail was the first official request for help with this package.
>
> That's not completely fair. I know that Dann Frazier has done several
> requests for this last week, though maybe not on the 68k list.
Right, he asked me on IRC, Wouter asked me on IRC, but that was on friday
night when my IRC client ran at work... seems we live in different
timezones. I started some builds on monday, but as I said I grabbed the
wrong source. When that was finished another build on q650 had already been
started, but the timeout wasn't adjusted properly. It does not help blaming
other people now, but it would help for all packages if they would produce
some output if they are still working on something, especially if that part
takes a long time.
> Anyway, don't tell me that the package will be ready tomorrow, tell
> Martin. He's the guy waiting for it. And it'll still need BYHAND
> processing.
Which Martin? Anyhow, it seems I spoke too soon, the problem that I saw
already in my useless build with the old sarge version has become worse.
There were several segfaults, and now the build fails due to a segfault:
Info: creating .pdf file...
Info: creating temporary .html file...
Info: creating .txt file...
./buildone.sh: line 292: 2643 Segmentation fault /usr/bin/w3m -dump
$tempdir/install.${language}.corr.html -o display_charset=$CHARSET >$destdir/install.${language}.txt
Error: build of txt failed with error code 139
Warning: The following formats failed to build: txt
make: *** [build-stamp] Error 2
buildd@aahz:~>w3m
Segmentation fault
This looks a bit like
#318806: w3m: Segfaults on ia64 every time
Package: w3m (0.5.1-3); Severity: important; Reported by: Roland Rosenfeld; 1 year and 30 days old
Starting program: /usr/bin/w3m
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
(no debugging symbols found)
[Thread debugging using libthread_db enabled]
[New Thread 16384 (LWP 5572)]
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 16384 (LWP 5572)]
0xc00619ca in GC_mark_from () from /usr/lib/libgc.so.1
I wonder why the build completed earlier this week when I built the sarge
version. I had segfaults in the same places, but the build continued. Has
something been changed in the Makefiles?
Christian
Reply to: