Re: Side effect of webwml alioth transition

[Matt Kraai <kraai@ftbfs.org>]

On Tue, Jul 07, 2009 at 12:24:52AM +0900, Osamu Aoki wrote:
> On Sun, Jul 05, 2009 at 10:10:38PM +0200, Nicolas François wrote:
> > On Sun, Jul 05, 2009 at 09:58:30PM +0200, Nicolas François wrote:
> > > 
> > > If this does not fix the issue, could anybody point me to the
> > > beyond-pkging.dbk and po4a/ja/beyond-pkging.po files?
> > 
> > I could find the files.
> > 
> > I tried on an Etch and I could reproduce the issue on an Etch chroot, and
> > pre-pending the po4a command with COLUMNS=1000 fixed it.
> > (Fixing the PO file should also fix the problem, setting COLUMN would just
> > be a hack to avoid such problem in the future)
> 
> Is it good to change parent Makefile to:
> 
> $(SUBDIRS-publish):
>         -/usr/bin/time -f "%C took %E real, %U user, %S sys time" \
>           COLUMNS=1000 $(MAKE) -C $(subst -publish,,$@) publish $(publish_args)
> 
> I did it but does this export COLUMNS ?

I think you need to put the COLUMNS=1000 before /usr/bin/time, as the
log files currently contain lines like the following:

 /usr/bin/time -f "%C took %E real, %U user, %S sys time" \
   COLUMNS=1000 make -C apt-howto publish PUBLISHDIR=/org/www.debian.org/www/doc/manuals install_file="install -p -m 664" install_dir="install -d -m 2775"
 /usr/bin/time: cannot run COLUMNS=1000: No such file or directory
 Command exited with non-zero status 127

-- 
Matt Kraai                                           http://ftbfs.org/