bug-texinfo
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: texinfo-6.0.91 pretest


From: Eli Zaretskii
Subject: Re: texinfo-6.0.91 pretest
Date: Sat, 02 Jan 2016 13:26:23 +0200

> Date: Sat, 2 Jan 2016 10:45:39 +0000
> From: Gavin Smith <address@hidden>
> Cc: Texinfo <address@hidden>
> 
> >      diff --strip-trailing-cr -a -u -r 
> > .//diffs/staging_res/htmlxref_nodes/test_refs.2 
> > .//out_parser/htmlxref_nodes/test_refs.2
> >      --- .//diffs/staging_res/htmlxref_nodes/test_refs.2     2014-12-19 
> > 22:07:28 +0200
> >      +++ .//out_parser/htmlxref_nodes/test_refs.2    2016-01-02 11:36:15 
> > +0200
> >      @@ -1,2 +1,2 @@
> >      -./htmlxref.cnf:15: warning: unrecognized type: foo
> >      -./htmlxref.cnf:16: warning: missing type
> >      +htmlxref.cnf:15: warning: unrecognized type: foo
> >      +htmlxref.cnf:16: warning: missing type
> >      FAIL test_scripts/htmlxref_htmlxref_nodes.sh (exit status: 1)
> >
> > So it's a "./foo" vs "foo" issue.  I don't think this is a real
> > problem, is it?  If so, can this be fixed somehow so as not to cause
> > bogus failure indications?  (Btw, why two slashes in "//diffs"?)
> 
> It's due to the Windows version of File::Spec->catfile canonicalizing
> the path to remove a leading "./". It's called in "locate_init_file"
> in texi2any.pl.
> 
> It would seem to be a loss to avoid this canonicalization for the sake
> of the test suite.

Agreed.

> Maybe this test failure will just have to be accepted under Windows.

That's undesirable, since one then has to use "make -k check".  But
certainly not a catastrophe, if no better solution is found.

> There are ways out, of course, e.g. multiple reference output files,
> only one of which would have to match, or conditional tests somehow,
> but I can't immediately think of anything easy.

One way is to run the output through Sed, before comparing, to replace
"^htmlxref" with "./htmlxref".




reply via email to

[Prev in Thread] Current Thread [Next in Thread]