On 06.08.2012 07:51, Justin Stottlemyer wrote:
Hi,
I recently built/recompiled 2.6Stable (currently in use at my company
in
other places) 2.7 Stable, 3.1.20, and 3.2.latest build. I've
deployed
squid dozens of times at different places and as usual ran into some
problems on the newer / later builds. I'm interested in seeing
where you
are heading, possibly mentioning some of the problems I ran into (
silentish errors) and seeing where discussions are currently leading.
That would be great Justin, thank you. I'm actively trying to resolve
all these minor glitches in 3.2+ and make the upgrade process more
seamless.
The version forks were:
2.5->2.6->2.7->3.2
2.5->3.0->3.1->3.2
2.6/2.7->3.0/3.1 cross-fork upgrade issues are not that interesting,
since the version features were so different various sized issues
including complete blockers are to be expected.
But the X.Y -> 3.2 transition is very important to get seamless or at
least documented clearly now that we are officially re-merging the 2.x
fork. Any input you can help provide towards writing a release notes
section, or pointing out remaining "bungled config" issues, would be a
great help.
NP: The key tool being promoted to ease those upgrades is "squid -k
parse" and its output.
I have figured out my configuration to get things working and its not
the
prettiest but it is definitely very functional. Squid 3.2 looks to
have
everything I would want it to have, I just want to try and keep
abreast of
where things are for now.
I once again considered some alternatives like Varnish (missing
features),
and lighttpd with mod_cache ( Not sure I want to bother) and as usual
came
back to Squid.
I saw mention of no bugs in 14 days in squid 3.2 so I thought I would
ask
if I should help out and re-compile the latest automated 3.2 build
and
attempt to re-use my configuration and mention any of the problems I
run
into, if they are still there.
Please do. Our testing infrastructure still has a light coverage on the
tests it can perform, so we do rely on real installations a lot for
reporting how the releases are going.
In addition, is this team still interested in 3.1 bugs?
We are. If they are resolved already in 3.2 they get closed against
that release and we (usually) won't donate time re-fixing. But it is
still good to have a record that they existed, since the OS
distributions are largely still shipping 3.1 packages for a few years to
come. You just have to be careful to check the closed bugs as well
before reporting to avoid "already-fixed" duplicates.
Amos
