Re: RFS: pysolfc (replacement for removed package: pysol)
Am Freitag, den 02.12.2011, 02:08 +0100 schrieb Jakub Wilk:
> (I don't intend to sponsor this package.)
Thank you anyway for taking the time to review!
> Lintian says:
>
> W: pysol: transitional-package-should-be-oldlibs-extra oldlibs/optional
> I: pysolfc: conflicts-with-version pysol (<< 1:2.0-1)
> I: pysolfc: conflicts-with-version pysol-cardsets (<< 1:2.0+dfsg-1)
> X: pysolfc: package-contains-broken-symlink usr/games/pysolfc ../share/games/pysolfc/pysol.py
> (plus some pedantic tags)
>
> The last one is particularly bad.
Should be fixed now; the first one also in pysolfc-cardsets.
> Why the versions in Conflicts need "-1" or "+dfsg-1" suffix? Wouldn't
> "<< 1.2.0" be enough?
Changed to "<< 1:2.0"; also in pysolfc-cardsets.
> Anyway, what are all these conflicts for?
That's stuff that was required (or recommended) by pysol, but has either
been integrated into pysolfc or superseded by other (more standard)
requirements (like python-pygame). Is it wrong to have this kind of
stuff in Conflicts?
> The "montana" patch is not listed in debian/patches/series. Is that
> intentional?
It wasn't up-to-date. Updated, added to debian/patches/series.
> I ran pyflakes over the source code, and it detected a few undefined
> name errors:
> ./pysollib/pysolgtk/tkcanvas.py:332: undefined name 'v'
> ./pysollib/pysolgtk/toolbar.py:144: undefined name 'v'
> ./pysollib/pysolgtk/progressbar.py:146: undefined name 'timeout_add'
> ./pysollib/pysolgtk/progressbar.py:151: undefined name 'mainquit'
> ./pysollib/pysolgtk/progressbar.py:158: undefined name 'mainquit'
> ./pysollib/pysolgtk/progressbar.py:159: undefined name 'mainquit'
> ./pysollib/pysolgtk/progressbar.py:166: undefined name 'main'
> ./pysollib/pysolgtk/tkwrap.py:70: undefined name 'v'
> ./pysollib/pysolgtk/tkwrap.py:115: undefined name 'cursor_new'
> ./pysollib/pysolgtk/tkwrap.py:115: undefined name 'v'
> ./pysollib/pysolgtk/tkwrap.py:170: undefined name 'create_pixmap_from_xpm'
I think the gtk frontend is rather unmaintained (tk being upstream's
preferred choice). Okay to leave unfixed?
Am Freitag, den 02.12.2011, 18:30 +0100 schrieb Jakub Wilk:
> One more thing: python is needed in the clean target, so it belongs in
> Build-Depends, not Build-Depends-Indep.
> And it should be probably s/python/python-all/, otherwise your package
> will FTBFS in some non-minimal enironments.
Fixed both issues.
Kind regards
Bernhard Reiter
Reply to: