On Mon, May 12, 2014 at 12:02:26PM -0700, Jonathan Nieder wrote:
> Mark Brown wrote:

> > This doesn't address the issue in the original report: this isn't
> > something we should be doing off our own back in Debian, it's something
> > that upstream should be doing.  The issue isn't that it's hard to build
> > shared libraries, the issue is that it's hard to maintain them and so it
> > needs to be something that upstream are engaged with.

> I'm a little puzzled.  Upstream has rules for installing libminizip at
> contrib/minizip/Makefile.am.  Are you saying that they are buggy or
> unmaintained?

No, I'm saying that I wasn't aware that this had now been done - I
didn't check given that there didn't seem to be anything addressing the
content of the discussion in the original report.

I do note that this has not been added to the standard Makefile
though, that is still static only, so it's not clear to me that the
developer (who is mostly a Windows guy) is really aware of this or
understands shared libraries on Unix.  I'm also not seeing symbol
versioning which would be nice but I guess shouldn't be a blocker.

> > Please discuss this with upstream.

> If you have bug reports in upstream's libminizip support, I'm happy to
> work on them upstream.  Last time I checked, it Just Worked (tm).

> I don't think it would make sense to go upstream and say "Mark Brown
> is not happy with how contrib/minizip deals with the shared library.
> Discuss" without further details. ;-)

Well, figuring out if they realise it is building a shared library and
understand the stable ABI requirements that Unix like systems have would
be the main thing I guess.  The comments in MiniZip64_info.txt about the
changes for 64 bit are a bit worrying here but seem to mostly refer to
the on disk format rather than the interfaces for users of the library.

Attachment: signature.asc
Description: Digital signature

Reply via email to