Re: etc/refcards/ and new-style backquotes

[Miles Bader <miles.bader@necel.com>]

Reiner Steib <reinersteib+gmane@imap.cc> writes:
>> [I _tentatively_ favor the latter, as it's always seemed to me that
>> the changes to the 5.10 branch are sometimes a wee bit aggressive
>> considering it's a release branch.]
>
> Do you have specific changes in mind?  I don't recall that we
> introduced any regressions or other problems.

No.  But some changes made on the 5.10 branch look a bit involved to me
(and often differ in details from the same change on the trunk), and my
impression is that the threshold for including a change on the 5.10
branch is kind of low (I think the same is true of the Emacs rel-22
branch, BTW).

> I don't understand your argument here.  As the Emacs trunk is open for
> any changes now, it seems quite dangerous to me to sync changes from
> Emacs trunk to gnus-5.10, isn't it?

Well, in general you're right, though in practice it's maybe not a big
deal -- almost every change made to Gnus on the Emacs trunk is basically
pretty trivial (they're almost all spelling fixes and random stuff like
that).

> - gnus-5.10 <-> emacs-rel-22
> - gnus-trunk <-> emacs-trunk
> - gnus-5.10 -> gnus-trunk and/or emacs-rel-22 -> emacs-trunk
>
>   To ensure bug fixes also go into the development series.
>   "and/or", because both mostly do the same, if we have the two other
>   channels in place.
>
> Could you explain what kind of (new) problems you'd expect in this
> scenario?

Only that many changes will be showing up via multiple paths, and the
simplistic mechanisms used to resolve such "conflicts" are fairly
brittle.

If the change rate on the branch(es) is low, it's not a big deal, but
it's kind of annoying slogging through merging
similar-but-not-quite-identical versions of the same change.

Merging the Gnus trunk into the Emacs trunk of course needs someone (not
me) to deal with the PGG mess.

-Miles

-- 
Suburbia: where they tear out the trees and then name streets after them.