Bug#576457: config.h regenerated at wrong place for amd64

2010-04-10 Thread Giuseppe Iuculano
> The failure to start seems to have been fixed (or at least, "fixed") by the
> 0.9.21-1.2+b1 binNMU on amd64.

Yes, the problem was that the "touch config.h.in" didn't have any effect
probably because it was built in a very fast buildd and ext3 only has a
1 second resolution.
It's a timing/resolution related issue, calling explicitly autoconf will
fix the issue.


Cheers,
Giuseppe



signature.asc
Description: OpenPGP digital signature


Bug#576457: config.h regenerated at wrong place for amd64

2010-04-08 Thread Simon McVittie
On Wed, 07 Apr 2010 at 11:54:23 +0200, Torsten Marek wrote:
> I've checked some other architecture's buildd logs, they look exactly
> like the local one.

The failure to start seems to have been fixed (or at least, "fixed") by the
0.9.21-1.2+b1 binNMU on amd64.

Simon



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org



Bug#576457: config.h regenerated at wrong place for amd64

2010-04-07 Thread Torsten Marek
Hi,

something is strange in there. If you check the build log, you'll see
that configure is run twice, then config.h.in is regenerated using
autoheader, but config.h is build via config.status only after "make
all-recursive" is executed.

In a local build, config.h is created immediately after config.h.in, and
judging from the rules, there is no way around that:


all: $(BUILT_SOURCES) config.h
$(MAKE) $(AM_MAKEFLAGS) all-recursive


...

config.h: stamp-h1
@if test ! -f $@; then \
  rm -f stamp-h1; \
  $(MAKE) $(AM_MAKEFLAGS) stamp-h1; \
else :; fi

stamp-h1: $(srcdir)/config.h.in $(top_builddir)/config.status
@rm -f stamp-h1
cd $(top_builddir) && $(SHELL) ./config.status config.h
$(srcdir)/config.h.in:  $(am__configure_deps) 
($(am__cd) $(top_srcdir) && $(AUTOHEADER))
rm -f stamp-h1
touch $@

(copy from a local Makefile)

Excerpt from the buildd log:

make[1]: Leaving directory 
`/build/buildd-pulseaudio_0.9.21-1.2-amd64-2bGIeC/pulseaudio-0.9.21'
make[1]: Entering directory 
`/build/buildd-pulseaudio_0.9.21-1.2-amd64-2bGIeC/pulseaudio-0.9.21'
echo 0.9.21 > .version-t && mv .version-t .version
(CDPATH="${ZSH_VERSION+.}:" && cd . && /bin/bash 
/build/buildd-pulseaudio_0.9.21-1.2-amd64-2bGIeC/pulseaudio-0.9.21/missing 
--run autoheader)
rm -f stamp-h1
touch config.h.in
/usr/bin/make  all-recursive
... many many lines ...
cd . && /bin/bash ./config.status config.h
config.status: creating config.h


Excerpt from local log:

make[1]: Leaving directory `/tmp/buildd/pulseaudio-0.9.21'
make[1]: Entering directory `/tmp/buildd/pulseaudio-0.9.21'
echo 0.9.21 > .version-t && mv .version-t .version
(CDPATH="${ZSH_VERSION+.}:" && cd . && /bin/bash 
/tmp/buildd/pulseaudio-0.9.21/missing --run autoheader)
rm -f stamp-h1
touch config.h.in
cd . && /bin/bash ./config.status config.h
config.status: creating config.h
/usr/bin/make  all-recursive

I've checked some other architecture's buildd logs, they look exactly like the 
local one.

best,

Torsten
-- 
.: Torsten Marek
.: University of Zurich
.: Institute of Computational Linguistics
.: http://www.cl.uzh.ch/en/tmarek.html






-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org