Re: [Evolution] About performance

[Adam Tauno Williams <awilliam whitemice org>]

On Mon, 2013-05-13 at 10:05 +0100, Pete Biggs wrote:
> > But man this is kind of the same problem of the known kernel IO problem 
> > that also I reported and nobody not much people trust just because they 
> > didn't see any problem.
> Because one person reporting problems is not usually credible - it's
> indicative of a misconfiguration or an issue elsewhere.  Yes, it
> sometimes is a problem, and given an infinite amount of resources all
> such things would be investigated - but with limited (very limited)
> resources, the developers have to concentrate on the things that have
> the most impact for the most people.

AND, it is important to note, that LINUX [and related applications] is
used very successfully in a myriad of situations by a whole lot of
people.

Most low-level bugs at this point in time are extremely narrow, it is
unreasonable to expect the world to jump on them when they harm 0.0001%
of the universe.

> > The same occurs in evolution. Just because it works well for you it does 
> > not mean it's  working well.
> > I can tell you that performance and usability degraded over time.
> but not for most people.  My experience is that Evolution has become
> more stable and more usable over the last few releases.  And yes, it has
> become faster and more responsive.

Exactly, it has gotten faster and MUCH more stable.   Some component may
have broken, or something degraded, but "Evolution" has done neither.

> To be honest most of the grouching about the stability and speed of Evo
> seems to be coming from Ubuntu users

Or people with *unbelievably* ancient versions.  Every time I see
someone complaining about 2.32 or even 2.28... my jaw drops.  I just
don't get it - why do that to yourself?

>  - that may be because there are
> just more users of Ubuntu than other distros; or it could be something
> that the Ubuntu packagers have done to Evo; or it could be some
> interaction of Evo with other libraries that Ubuntu have modified.
> That's not to say that there aren't reported problems with other
> distros, but they don't seem to make Evo unusable like it reportedly
> does on Ubuntu.

They [Ubuntu] have had some very lemony releases; so everyone gets to
experience the long-tail of those versions.

> File bugs about it.  That's the only way the developers can get a view
> on systemic problems and can spot patterns.

And bugs do get fixed.  My latest bug report was resolved in 48 hours.

> It's also very helpful if when you do submit a bug following a posting
> to this list, that you tell us the bug ID - at least then if some one
> searches the list archives (because we ALL do that before posting, don't
> we) they can at least see if the problem has been fixed, or can add a
> comment to the bug.

And perhaps people on the list can follow the bug or add something to
it.  I've done that several times.

> Finally, the developers make advances and improvements and bug fixes in
> the current version and only bug fixes in the previous version - so it
> is always worthwhile running the most up to date version before
> criticizing things too much.

Yes.  3.6.x is a major release back.  3.8.x *does* fix some performance
issues, especially related to flaky connections.  Or it certainly seems
that way to me.

-- 
Adam Tauno Williams <mailto:awilliam whitemice org> GPG D95ED383
Systems Administrator, Python Developer, LPI / NCLA