Mailinglist Archive: opensuse-factory (249 mails)

< Previous Next >
Re: [opensuse-factory] Status: distribution
  • From: Marcus Meissner <meissner@xxxxxxx>
  • Date: Mon, 1 Feb 2010 10:50:40 +0100
  • Message-id: <20100201095040.GA12960@xxxxxxx>
On Wed, Jan 27, 2010 at 10:01:35AM +0100, Stephan Kulow wrote:
Am Dienstag 26 Januar 2010 schrieb Stefan Seyfried:
2010/1/26 Stephan Kulow <coolo@xxxxxxxxxx>:
The status of ppc is the same as the last months: I do not have time for
it and I invite people who care to look it up on build.opensuse.org and
fix the packages. So far I only see Marcus working on it.

I wanted to take a look at them (first the failing ones, which, if fixed,
will probably fix some of the expansion errors). However, stuff
like this:

gcc libdmraid-events.c -L../lib -ldmraid -lm -g -fPIC -Wall -shared
-Wl,-soname,libdmraid-events.so.0 -o libdmraid-events.so.0
/tmp/ccAt8rbT.s: Assembler messages:
/tmp/ccAt8rbT.s:16: Error: unknown pseudo-op: `.cfi_sections'
make: *** [libdmraid-events] Error 1

is above my ppc competence ;) This seems to be a common failure,
so maybe if someone with ppc or toolchain knowledge fixes this
".cfi_sections" errors, the situation would improve dramatically.

Good luck :-)

Where are you looking?

I see 44 failures in standard/ppc that are not also in either i586 or
x86_64. Some of them are really ppc specific code (Mesa[1] for example),
others are plain packaging bugs (kiwi from the look at it) and others
are random failures that happen on other archs too (perl). The worst of
all is the failing glibc on ppc. The yast2-core failure looks strange too.

.cfi_section is "staging", which is the gcc 4.5 preparation area.

For PowerPC meanwhile fixed:
- installation-images: submitted to snwint
- ruby: submitreq done
- kiwi: submitreq done
- libqt4 (and the whole KDE4) stack: fixed by last binutils checkin from
Micha, is rebuilding fine
- glibc: fixed by upgrade
- mozilla-xulrunner192: submitreq against wireless-tools done (that redefined
"inline" to a
non-C++ compliant "inline __attribute__((always_inline))"

Unfixed major issue:
- gtk2 stack (everything that misses "GIR" files)...

Caused by new "gobject-introspection", it cannot read in
the typelibs it wrote a second before due to a funny offset issue.
I suspected an endianess issue, but while it happens only on Big Endian
it seems to be a different issue.

Ciao, Marcus
--
To unsubscribe, e-mail: opensuse-factory+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse-factory+help@xxxxxxxxxxxx

< Previous Next >