Proposal to reduce anti-bundling requirements

Richard Shaw hobbes1069 at gmail.com
Fri Oct 2 14:37:41 UTC 2015


On Fri, Oct 2, 2015 at 9:22 AM, Matthew Miller <mattdm at fedoraproject.org>
wrote:

> (To be clear, I personally am in favor of also allowing more relaxed
> bundling for smaller projects which are on the fringes of the system
> integration you discuss. In other words, vastly expanding the "too
> small to care" exception.)


This would help tremendously in my case...

I would not my experience asking for a bundling exception was "pleasant". I
asked for an exception for some Ham Radio software where upstream is
largely one person and not only was upstream dead, but he had modified it
where it wouldn't work with the official package even if it were in Fedora
(which is wasn't)

Since the final determination was no, ultimately upstream (of one) decided
he would break out the bundled library into it's own library so his suite
of programs could all use the same  (instead of bundling it in each one
statically). While I think this is a good outcome, he was pretty pissed
about it and only did it because he didn't want to loose Fedora users.

You could say that you don't care, the right thing happened in the end, but
then again the FPC isn't the one trying to maintain a good working
relationship with upstream. I am.

Again, trying to package one of his programs a bundled "mongoose" was
found. I tried the easy thing first, linking against the Fedora version,
but although they do official releases, it's still designed to be a copylib
as well and again, the version in his software it too modified to work with
Fedora's version. In this case it's not shared across his suite of
programs, it's the only one that uses it.

My plan again is to request a bundling exception under the "too small to
care" option, but I'm not particularly looking forward to the process.

Thanks,
Richard
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.fedoraproject.org/pipermail/devel/attachments/20151002/c19404c8/attachment.html>


More information about the devel mailing list