SeaMonkey 2.0 Release Candidate 1

Discussion about Seamonkey builds
User avatar
Philip Chee
Posts: 6475
Joined: March 1st, 2005, 3:03 pm
Contact:

SeaMonkey 2.0 Release Candidate 1

Post by Philip Chee »

Date: Sat, 10 Oct 2009 18:41:02 +0200
From: Robert Kaiser <kairo@kairo.at>
Subject: SeaMonkey 2.0 Release Candidate 1

SeaMonkey 2.0 Release Candidate 1 is now available for free download on the SeaMonkey website. We encourage testers to get involved in discussing and reporting problems as well as further improving the product.

Full news article:
http://www.seamonkey-project.org/news#2009-10-10

Downloads for all available platforms and languages:
http://www.seamonkey-project.org/releases/2.0rc1

Release notes (preliminary ones for 2.0, RCs don't get separate relnotes):
http://www.seamonkey-project.org/releases/seamonkey2.0

System Requirements:
http://www.seamonkey-project.org/doc/2.0/system-requirements

Newsgroups: http://www.seamonkey-project.org/community#groups
File a bug: http://bugzilla.mozilla.org/enter_bug.cgi?format=guided
Get Involved: http://www.seamonkey-project.org/dev/get-involved

Robert Kaiser
SeaMonkey project coordinator
Rickkins
Posts: 522
Joined: January 12th, 2004, 3:25 pm

Re: SeaMonkey 2.0 Release Candidate 1

Post by Rickkins »

Outstanding..!!
007simple
Posts: 8
Joined: July 29th, 2009, 8:02 am
Contact:

Re: SeaMonkey 2.0 Release Candidate 1

Post by 007simple »

waiting for final version to relase
User avatar
Uncle Spellbinder
Posts: 3519
Joined: May 28th, 2004, 4:52 pm
Location: Highland, IN - U.S.A.
Contact:

Re: SeaMonkey 2.0 Release Candidate 1

Post by Uncle Spellbinder »

There's still an RC2 to come before final release, I believe.
My Firefox Add-Ons Collection: Firefox Essentials
User avatar
Philip Chee
Posts: 6475
Joined: March 1st, 2005, 3:03 pm
Contact:

Re: SeaMonkey 2.0 Release Candidate 1

Post by Philip Chee »

Uncle Spellbinder wrote:There's still an RC2 to come before final release, I believe.

If there are no problems with RC2, that will be the final release.

Phil
User avatar
Uncle Spellbinder
Posts: 3519
Joined: May 28th, 2004, 4:52 pm
Location: Highland, IN - U.S.A.
Contact:

Re: SeaMonkey 2.0 Release Candidate 1

Post by Uncle Spellbinder »

Cool. Thanks, Phil.
My Firefox Add-Ons Collection: Firefox Essentials
rsx11m
Moderator
Posts: 14404
Joined: May 3rd, 2007, 7:40 am
Location: US

Re: SeaMonkey 2.0 Release Candidate 1

Post by rsx11m »

The RC2 builds are currently in progress, they needed yet another respin after colliding with the Firefox 3.5.4 respin (I always thought that the tagging process is instantaneous on Mercurial as you are tagging changesets rather than individual file revisions, but apparently it is not).
User avatar
Philip Chee
Posts: 6475
Joined: March 1st, 2005, 3:03 pm
Contact:

Re: SeaMonkey 2.0 Release Candidate 1

Post by Philip Chee »

rsx11m wrote:The RC2 builds are currently in progress, they needed yet another respin after colliding with the Firefox 3.5.4 respin (I always thought that the tagging process is instantaneous on Mercurial as you are tagging changesets rather than individual file revisions, but apparently it is not).

In Mercurial to tag a changeset you add a line to the .hgtags file saying <changeset> <tag> e.g.

Code: Select all

49342a1d9d9390addb6f803cb55ea77c03c7dd15 FIREFOX_3_5_4_BUILD3
e178dbd2a38939075df72c8d3468408f7314a9c0 FIREFOX_3_5_4_RELEASE
49342a1d9d9390addb6f803cb55ea77c03c7dd15 FIREFOX_3_5_4_RELEASE
49342a1d9d9390addb6f803cb55ea77c03c7dd15 SEAMONKEY_2_0rc2_BUILD3
e178dbd2a38939075df72c8d3468408f7314a9c0 SEAMONKEY_2_0rc2_RELEASE
49342a1d9d9390addb6f803cb55ea77c03c7dd15 SEAMONKEY_2_0rc2_RELEASE

Phil
rsx11m
Moderator
Posts: 14404
Joined: May 3rd, 2007, 7:40 am
Location: US

Re: SeaMonkey 2.0 Release Candidate 1

Post by rsx11m »

So, why does it take 11 minutes to do that, thus allowing race conditions like this to happen? :-k
User avatar
Philip Chee
Posts: 6475
Joined: March 1st, 2005, 3:03 pm
Contact:

Re: SeaMonkey 2.0 Release Candidate 1

Post by Philip Chee »

rsx11m wrote:So, why does it take 11 minutes to do that, thus allowing race conditions like this to happen? :-k

That time interval has nothing to do with
http://hg.mozilla.org/releases/mozilla-1.9.1/rev/2fe392234f8c

See: https://wiki.mozilla.org/SeaMonkey:Release_Process:2.0rc2#Build_3
# Looks like the Firefox build process removed the SeaMonkey build2 tags from mozilla-1.9.1, so I updated release_config.py for a build3 and commented out all tagging except the Mozilla repo in release_master.py

Phil
rsx11m
Moderator
Posts: 14404
Joined: May 3rd, 2007, 7:40 am
Location: US

Re: SeaMonkey 2.0 Release Candidate 1

Post by rsx11m »

My thinking was that the order of events was something like this:
  1. hg pull (SM)
  2. hg pull (FF)
  3. hg tag (SM)
  4. hg tag (FF)
  5. hg push (SM)
  6. hg push (FF)
where step #6 would have overridden the tags set in step #5 as observed. I thought Mercurial would take care of those mid-air collisions, thus apparently the manual intervention resulted in the loss of the tags.

Anyway, RC2 build 3 seems to be fine now.
User avatar
Philip Chee
Posts: 6475
Joined: March 1st, 2005, 3:03 pm
Contact:

Re: SeaMonkey 2.0 Release Candidate 1

Post by Philip Chee »

rsx11m wrote:My thinking was that the order of events was something like this:
  1. hg pull (SM)
  2. hg pull (FF)
  3. hg tag (SM)
  4. hg tag (FF)
  5. hg push (SM)
  6. hg push (FF)
where step #6 would have overridden the tags set in step #5 as observed. I thought Mercurial would take care of those mid-air collisions, thus apparently the manual intervention resulted in the loss of the tags.

Anyway, RC2 build 3 seems to be fine now.

This is known as a "bad merge". Since ffxbld and seabld are buildbots, their tiny little pythonesque minds probably didn't notice the problems.

Phil
rsx11m
Moderator
Posts: 14404
Joined: May 3rd, 2007, 7:40 am
Location: US

Re: SeaMonkey 2.0 Release Candidate 1

Post by rsx11m »

Yeah, that's a common problem of version control systems. You'd have to lock the pull in step #1 and release it in #5, then the conflicting buildbot in step #2 should notice that it can't get a lock on that branch. I couldn't find anything in the hg description though if locking is supported, thus they'll have to coordinate running their bots not too much in parallel in the future...
User avatar
Philip Chee
Posts: 6475
Joined: March 1st, 2005, 3:03 pm
Contact:

Re: SeaMonkey 2.0 Release Candidate 1

Post by Philip Chee »

rsx11m wrote:Yeah, that's a common problem of version control systems. You'd have to lock the pull in step #1 and release it in #5, then the conflicting buildbot in step #2 should notice that it can't get a lock on that branch. I couldn't find anything in the hg description though if locking is supported, thus they'll have to coordinate running their bots not too much in parallel in the future...

Like most if not all distributed version control systems, Mercurial doesn't have the concept of "locking". The assumption is that the person doing the merge would resolve the merge conflicts (using say rebase, or even manually). Unfortunately the "persons" here are just a couple of cranky buildbots.

Phil
rsx11m
Moderator
Posts: 14404
Joined: May 3rd, 2007, 7:40 am
Location: US

Re: SeaMonkey 2.0 Release Candidate 1

Post by rsx11m »

Phil, thanks for the explanation (and I've given up by now trying to imagine how a distributed lock might work ;-) ).
Post Reply