----- Original Message -----
On 03/27/2012 03:14 AM, Bohuslav Kabrda wrote:
> One more email from me:
> I don't know if you have been following the discussion about the
> new Ruby Guidelines at the fpc ticket [1] or at the packaging list
> (mainly, see last 3 comments, which summarize current state), but
> this week (wednesday, 17:00 UTC), the fpc is going to finish the
> draft, vote on it and close it. Please, if anyone of you has
> something to add, write it to the ticket or come to the meeting, I
> myself am going to be there to discuss the remaining things.
> This is really the last chance to alter something, so I would
> highly appreciate if more of us could come to the meeting or at
> least support our opinions at the ticket.
> We've been trying hard to carry through as much of our draft as we
> could, together with Vit (who is now enjoying a well deserved
> vacation). I hope you will find our opinions on the fpc changes
> reasonable and will support us. (And if you don't find our
> opinions reasonable, there is still time to say so, at least.)
>
> Thanks a lot!
>
Hey guys really appreciate the hard effort. Will try to make the FPC
meeting (this is on #fedora-meeting-1 right).
Just glancing over the new guidelines real quick, overall they look
good, but the extended bits to build gems does look like its going to
be
a PITA. I understand the reasoning behind it, but seems to add alot
of
overhead (and reading your last comment on the trac issue, will there
be
situations when it doesn't work?). Also would like to discuss some of
the other bits such as all rubygems providing ruby(libraryname) and
the
bits about interpreter independence.
Well, I agree with you on the building to be a great PITA, but fpc guys are pretty touchy
about that one after last Vit's email on this on packaging list [1], so I guess we
will have to accept that (from a certain point of view, they are right, but I still
believe that our solution would be better, as I have stated in my last comment in the fpc
ticket).
As for the provides, I am thinking the same and I've been arguing about that with
Toshio for a while. The problem he doesn't see is, that according to his proposal, we
can have a non-gem library, which provides ruby(foo) and a rubygem-foo, which also
provides ruby(foo), which will then result into unexpected behaviour when requiring
ruby(foo). I also wrote that in my last comment on the fpc ticket and I will bring it up
on the meeting.
Finally, please expect some maybe-not-so-friendly atmosphere, because there were some
heated discussions on the packaging list. The fpc members complain that we don't
listen to them and want to do things our way, but they basically do the same - they
changed the draft without understanding ruby (or consulting with us first) and in some
cases they were very wrong, but still arguing about it in the way "Look, I don't
understand Ruby, but...", which in turn made me and Vit very non-happy (which was
expressed in our reactions ;)). I would however very much like to try to throw all that
away and start all over with Ruby-SIG<->FPC relations (if someone form fpc reads
this, yes, I know we have our differencies, but we have to work together, so let's
calm down everybody and start again).
The trac issue / email threads have gotten pretty long, am hoping
the
irc convo tomorrow will be a little easier to follow. In any case,
take
care.
-Mo
--
Regards,
Bohuslav "Slavek" Kabrda.
[1]
http://lists.fedoraproject.org/pipermail/packaging/2012-March/008221.html