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

Bug#639144: lintian: doesn't detect any -> all (>= b:V) dependencies



Package: lintian
Version: 2.5.2
Severity: wishlist

Please see #636189 for details, According to Cyril Brulebois (KiBi on
#debian-buildd), this is actually a bug in lintian as well for not
detecting it:

<markos> guys, also any ideas how to deal with #636189? it really
blocks lots of packages from building
<KiBi> make it strip the binNMU part when generating the depends.
<markos> the +b1?
<KiBi> yes. and that's a serious bug
<markos> I see
<markos> ok, I'll send an update to the bug report and raise the security
<markos> er, severity
<KiBi> make it >= ${source:Version}
<KiBi> (I wasn't sure before looking whether it was computed in
debian/rules, or set using variables in debian/control; but that's the
former case, so easily fixed)
<markos> ok
<KiBi> I guess it wasn't noticed by lintian due to >=.
<KiBi> having a = ${binary:Version} between arch: any and arch: all
packages is plain wrong. but >= too.
<KiBi> so you might want to check that on the lintian side and report
a bug there.
<phil> Did somebody upgrade the severity?
<KiBi> phil: markos was on his way to
<KiBi> comments say:
<KiBi> # (b1) any -> any (= ${source:Version})          -> use b:V
<KiBi> # (b2) any -> all (= ${Source-Version}) [or b:V] -> use s:V
<KiBi> b2 should check for >= as well, I think.
<KiBi> in lintian's checks/version-substvars for the
not-binnmuable-any-depends-all tag.
<KiBi> markos: will you please file that lintian bug?
<markos> KiBi, will do
<KiBi> cheers

while fixing the bug, it was suggested to replace ${Source-Version}
with ${binary:Version} to remove confusion with ${source:Version} as
well.

Thanks

Konstantinos



Reply to: