Re: Build test using TeXInfo from experimental

On Sun, Feb 22, 2026 at 06:52:54PM +0100, Preuße, Hilmar wrote:

> > * octave
> > 
> > This one also fails while processing texinfo files:
> > 
> > Chapter 32 [915] [916] [917] [918] [919] [920] [921] [922] [923] [924] [925]
> > [926]Function-Index.html: reached EOF without finding data start pattern at ./doc/interpreter/mk-qthelp.pl line 77, <$
> > HTML> line 4580.
> >   [927]make[3]: *** [Makefile:32530: doc/interpreter/octave_interpreter.qhc] Error 255
> > 
> > so it also seems texinfo-related.
> > 
> I've seen the error message in the logs, but it seems the TeXInfo calls
> itself run fine, so I can't also see a relationship to TeXInfo upgrade.

Well, just to clarify:

The relationship is that I built it once with new texinfo and it failed once,
then I built it with the old texinfo (same version, same day, probably
same build-dependencies) and it built ok.

However, there are actually two variables which changed between those
two cases: One of them is the new texinfo, and the other variable is
that I still use the traditional schroot backend of sbuild for this
kind of rebuild (i.e. using an extra private repository).

So, to be sure, I'm going to build this package four more times, two
with new texinfo (which in my setup uses schroot), and two more in
pure unstable using also schroot, to discard that it's schroot what
triggers the difference.

In case texinfo is the variable which makes the package to fail, I'd
recommend that you put "texinfo" somewhere in the subject when
reporting the bug (i.e. "FTBFS with new texinfo"). Of course we would
leave the "why" to the maintainer.

Thanks.