Re: [Evolution] Evo-mail - too many files issue

[Nick Jenkins <nickpj gmail com>]

> FWIW, it was meaningful enough to me. The word 'Ubuntu' means that
> he's using a version of Evolution that is *so* old that I cannot be
> persuaded to care about it.

Given that we all seem to agree that Ubuntu are doofuses for not
shipping the latest stable Evo release with their distro releases from
10.10/Oct 2010 and onwards, I have to ask: Would there be any merit in
an upstream repository of pre-built Evo packages? E.g. every night,
there could be automated builds of Evo, for each of the current stable/
previous stable/current dev or head or unstable trees, to produce a .rpm
& .deb, for amd64 and x86-32 bit architectures (so 3 trees x 2 package
manager formats x 2 architectures = 12 automated builds a night).

That way:
* Laggard distros could be bypassed.
* Upgrading wouldn't be such a big deal. I'd estimate that maybe a third
of email threads in response to end user problems on this list at some
point will contain a "that's an old version, you need to upgrade" type
of response, irrelevant of distro. But one big reason that people are
reluctant to upgrade is that it's a hassle and typically involves
upgrading everything on the system, which in turn will probably break
other things - but you won't know exact what until you upgrade and then
find the regressions. That all seems a big ask when that person is only
bothered by one specific Evo-related problem.
* Users could have a much closer relationship to upstream (something I
would be entirely happy with, since it's upstream who fix the bugs & add
features).
* More users could track the latest development version.
* Upstream gets feedback much quicker, collapsing the currently quite
slow loop between things being added to git, then later rolled into a
stable Evo release, then being packaged by a distro, then being shipped
to users, that they then install as part of a complete system upgrade,
which only then can they provide useful feedback on.

But I don't know if such a thing is even feasible. For example:
* It seems a massive ask of the already very busy developers to set
something like this up. It looks like there is already some continuous
building going on (see http://live.gnome.org/BuildBrigade/ and
http://build.gnome.org/evolution ), but the aim doesn't seem to be to
produce end-user installable .deb and .rpm packages, which would have to
be an extra workload, so it may not be practical.
* Is it even technically possible to just upgrade Evo, or does the whole
gnome desktop need to be upgraded? I.e. does Evo rely on having the very
latest version of gnome, or will it run on the previous stable version
of gnome or the stable version before that? I suspect that it relies on
the latest gnome, but if the whole desktop has to be upgraded, then it's
not too far a leap from there to having to upgrade the whole distro.

-- All the best,
Nick.