Re: Patchy email

[James]

On 28/09/16 10:24, David Kastrup wrote:
> address@hidden writes:
>
> > 07:29:54 (UTC) Begin LilyPond compile, previous commit at       
> > 8a493b7bc9aa4d34cd2970fc2223ec6920f0ed02
> > 07:29:58 Merged staging, now at:        8a493b7bc9aa4d34cd2970fc2223ec6920f0ed02
> > 07:29:59        Success:                ./autogen.sh --noconfigure
> > 07:30:16        Success:                /tmp/lilypond-autobuild/configure 
> > --enable-checking
> > 07:30:18        Success:                nice make clean
> > 07:40:20        Success:                nice make -j3 CPU_COUNT=3
> > 07:50:03        Success:                nice make test -j3 CPU_COUNT=3
> > 08:54:44        Success:                nice make doc -j3 CPU_COUNT=3
> > 08:56:12 *** FAILED STEP ***
> >         merge from staging
> >         Command '['git', 'fetch']' returned non-zero exit status 128
> > ssh: Could not resolve hostname git.sv.gnu.org: Name or service not known
> > fatal: Could not read from remote repository.
> >
> > Please make sure you have the correct access rights
> > and the repository exists.
> > 08:56:12 Traceback (most recent call last):
> >    File 
> > "/usr/local/tmp/lilypond-extra/patches/compile_lilypond_test/__init__.py", line 
> > 530, in handle_staging
> >      self.merge_push ()
> >    File 
> > "/usr/local/tmp/lilypond-extra/patches/compile_lilypond_test/__init__.py", line 
> > 469, in merge_push
> >      run ("git fetch")
> >    File 
> > "/usr/local/tmp/lilypond-extra/patches/compile_lilypond_test/__init__.py", line 
> > 105, in run
> >      raise FailedCommand ("Command '%s' returned non-zero exit status %d\n%s" % 
> > (cmd, returncode, stderr.strip ()))
> > FailedCommand: Command '['git', 'fetch']' returned non-zero exit status 128
> > ssh: Could not resolve hostname git.sv.gnu.org: Name or service not known
> > fatal: Could not read from remote repository.
> >
> > Please make sure you have the correct access rights
> > and the repository exists.
> Intermittent net failure.  I took the liberty of still counting this as
> success and manually pushing this staging to master.
>
> Surprising how "dangerous" the manual completion of this Patchy run
> feels after a few years of our automated procedure.
Well it is pretty robust, you'd be surprised how many times I get these 
kinds of errors over a month (due to repo/network issues).

I've never seen a partial push/merge it fails or it doesn't and if it 
fails, then patchy clears up any stale 'lock' branches and starts over.

I'm impressed just how resilient these scripts have been :)

Now if only Jean-Charles, Graham et al, could script checking reg test 
diffs and finding obscure make errors in all the .log files we generate ;)

James