Lucas Nussbaum <[EMAIL PROTECTED]> wrote:

> sbuild checks for output to stdout. It means that during 90min, there
> was no output to stdout.
>
>>      # Create Unicode encoded subfonts `$(uninamestem)00' .. 
>> `$(uninamestem)ff'.
>>      # This will take more than an hour.  Make yourself a cup of tea.
>>      @echo fontforge -script subfonts.pe $(typeface).ttf $(uninamestem) 
>> $(sfddir)/Unicode.sfd
>>      @( cd $(builddir) && fontforge -script subfonts.pe $(typeface).ttf 
>> $(uninamestem) $(sfddir)/Unicode.sfd > log 2>&1 || ($(ERR)) )
>>      @echo
>> 
>> and similar, all write their stdout to log.  And I'd say there's one
>> message every couple of minutes usually.  So if this didn't happen,
>> maybe this is more than a normal or wishlist bug...
>
> Maybe, then, they shouldn't output to the logfile. This would allow to
> see progress and please sbuild, too.

It would give very long boring logfiles.  Well, around 100 lines and 100
blank lines for every "make yourself a cup of tea, and I don't know how
many for the following calls.  Anyway, maybe that's the way to go.

Personally, I was never sure that we should discard this.  Danai, what
do you think?  

I wouldn't make an upload just for this, but once we make a new upload,
anyway, we can just drop the redirections and the $ERR handling.

>> >> touch build-stamp.gkai
>> >> dh_testdir
>> >> touch build-stamp
>> >> Build killed with signal TERM after 90 minutes of inactivity
>> >> Build killed with signal KILL after 5 minutes of inactivity
>> >> Build killed with signal KILL after 5 minutes of inactivity
>> >> Build killed with signal KILL after 5 minutes of inactivity
>> 
>> This looks as if a couple of processes timed out, where killed but the
>> build proceeded?  Can you provide a full log?
>
> Yes, see
> http://http://ox.blop.info/bazaar/buildlogs/20061114/latex-cjk-chinese-arphic_1.15_etch32.buildlog

I'm still unsure whether this means that means the process was killed
right after touching build-stamp, which would indicate a bug - or
whether it just means that it was killed four times, much before it
failed? 

Regards, Frank
-- 
Dr. Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX/TeXLive)

Reply via email to