Am 22.01.2016 um 22:32 schrieb Georg Baum <georg.b...@post.rwth-aachen.de>:
>
> Stephan Witt wrote:
>
>> No. It’s not wrong. You’re right with the use of the _own_ userdir
>> for tests. But the real life matters too. It’s the wrong userdir
>> for standard orpeation. And this is Georg message, IMO.
>
> Yes. Also, depending on the reaason for the wrong default user dir, also the
> special one passed on the command line of the tests might have been used
> incorrectly by tex2lyx.
>
> I had a closer look, and I am now pretty sure that this is not the case: The
> user dir for the tests seems to work well, and the default user dir is not a
> regression (see my other mail). Since we do not know whether the failing
> tests are a regression, I am not sure anymore whether they should be
> considered a beta blocker or not.
>
> In the log of the failing tests one can see some strange things. For
> example, the CJK tests contain the line
>
> sepackage{CJKutf8}
>
> in the preamble. somehow the first two letters of \usepackage have been
> swallowed. Also, \setlength{\parindent}{3mm} in test-insets.tex is
> completely swallowed and not parsed.
>
> Stephan, can you reproduce one of the failing tests when running tex2lyx
> manually?You are asking for running tex2lyx in the debugger? Yes, I’m able to do so. What should I try exactly? I started tex2lyx in debugger with these options: -f -roundtrip /Users/stephan/git/lyx/src/tex2lyx/test/algo2e.tex /Users/stephan/git/lyx-build/algo2e.lyx I got the attached algo2e.lyx and algo2e.tex. Stephan > If yes, the quickest way to find the cause of the failing tests > would be to run it in the debugger and to step through the preamble parsing > code. Something is really fishy there. > > > Georg > >
algo2e.tex
Description: TeX document
algo2e.lyx
Description: Binary data
