Bug#410463: [Pkg-octave-devel] Bug#410463: octave2.9-forge: FTBFS: error: invalid matrix index = 1

[Rafael Laboissiere]

* Thomas Weber <thomas.weber.mail at gmail.com> [2007-02-21 15:23]:

> Package: octave2.9
> reassign 410463 octave2.9-forge
> thanks
>
> Am Samstag, den 10.02.2007, 21:38 +0100 schrieb Kurt Roeckx:
> > error: invalid matrix index = 1
> > error: evaluating assignment expression near line 158, column 12
> > error: evaluating if command near line 156, column 3
> > error: called from `test' in file `/usr/share/octave/2.9.9/m/testfun/test.m'
> > error: near line 7 of file `fntests.m'
> 
> This is a bug in the way octave2.9-forge creates its test files (they
> are simply not valid). However, fixing this turned up a bug in Octave
> 2.9. This bug is nasty, as it eats all memory during octave2.9-forge's
> unit tests. I guess buildd admins won't like that too much. Fortunately,
> it's already fixed upstream and a small patch suffices.

Although this bug does not concern the versions of the packages now in
testing (octave2.9_2.9.9-7 and octave2.9-forge_2006.07.09+dfsg1-7), I
understand that it would be good to have the version in testing also fixed.

> So either we fix the tests and ask debian-release for a new version of
> both octave2.9 and octave2.9-forge  or we disable them and try to get
> octave2.9-forge in its current state into Etch. 

Alternatively, we could let things as they are now.  After all, the packages
in testing are not affected.

At any rate, if we really want this bug fixed in testing, we would maximize
the odds of having octave2.9 accepted by applyin both patches
50_correct_test and 50_parse-inline-brackets to 2.9.9-7 and asking the
release managers to consider unblocking the resulting 2.2.9-7etch1 version.

-- 
Rafael