Re: [be] Re: Possible Debian/Ubuntu packaging issue

[Jonathan Marsden]

Teus Benschop wrote:

> Thinking about BibleTime, and any relationship that the debian package
> says might be there between the two, I don't see any relationship strong
> enough that this needs to be set at the package level. Bibledit is an
> independent Bible program, and so is BibleTime. Yes, if BibleTime runs,
> Bibledit notices that and sends references to is. But that is about all
> what the two have in common. If you would drop any relationship
> information from the package, that would be fine with me.

OK, understood.  If it causes issues, I'll drop it.  I'd like to
understand what the issue(s) really are, first, if I can!  After all,
that relationship has apparently been specified for years, across
multiple Debian releases, already.

At this point, there is no clear evidence that the Recommends: line is
causing replicable issues.  We don't even know for certain that Kim was
updating to my 3.6-1 package from my PPA, nor do we know why the attempt
caused the error it did.  Can you (or anyone else) duplicate this issue,
on Debian or Ubuntu?  Can we get a more complete bug report from Kim,
including steps to reproduce it, and the info I suggested earlier?  It
may be that Kim was seeing some sort of Ubuntu variant of Debian bug
#524194 -- we just don't know.

> I remember having seen a "bug" filed on the debian system that said that
> bibledit could not be compiled for one or more processor architectures.
> Can't remember which ones. The reason was that BibleTime didn't compile
> on that processor. Bibledit itself would have been fine on that
> processor, it was only that Recommends: line that caused this "bug". If
> this line were completely removed, than bibledit would be more "clean",
> and probably compile on any processor that debian supports.

OK.  It seems odd that a recommends relationship would cause a FTBFS.  I
don't think it is "supposed" to work that way!

As of right now, there are only two open bugs in the Debian BTS against
bibledit, as you can see at http://bugs.debian.org/bibledit -- and one
of them is #524194, a new packaging issue in 3.6-1 that I already fixed
in 3.6-2 and 3.7-1.  The remaining bug is an older wishlist bug which I
plan to investigate at some point (if it is indeed still applicable to
the new codebase -- there are no patches at all in 3.7-1, so I strongly
suspect it is also now fixed, but I need to check on that!).

So it looks like the (old?) bug you are referring to has already been
closed.  However, I don't see it in the list of all bugs known to the
Debian BTS for the bibledit package, which is only six bugs, visible at
http://bugs.debian.org/cgi-bin/pkgreport.cgi?archive=both;package=bibledit
, so I am a bit puzzled as to where this bug report really is!

Looking at building on various CPU architectures, see
http://ftp.debian.org/debian/pool/main/b/bibledit/ .  The current
bibledit official 2.2-2.1 packages in Debian stable (Lenny), which
include the Recommends: bibledit line, built on 13 platforms.

My first upload of bibledit to Debian, bibledit 3.6-1, has been built on
7 architectures already.  The others may have an issue with a fix I made
regarding depending on a virtual package (ping)... or it may be other
packages bibledit needs that are broken in sid... I'm looking at that in
more detail tonight.  Other than that, I see nothing suggesting that
bibledit can't build on all of the others too.  Softening the Recommends
into a Suggests can't hurt -- but to be honest, I doubt it will make any
difference to this at all.

The current overall bibledit package status page in Debian PTS at
http://packages.qa.debian.org/b/bibledit.html shows this information,
too.  The two Lintian warnings and bug #524194 are already fixed in
packages that I hope to get uploaded to Debian unstable this week.

Jonathan