598306 - (SQLite3.7.3) Upgrade to SQLite 3.7.3

Status: RESOLVED FIXED

(Core :: SQLite and Embedded Database Bindings, defect)

Description

[Shawn Wilsher :sdwilsh]

This will fix at least one crasher for us.

Comment 1

[Marco Bonardo [:mak]]

also fixes an assertion and a performance regression (bug 595530, moving a query from 100ms to 12s, other effects are unknown so far).

Comment 2

[Marco Bonardo [:mak]]

3.7.3 is officially out

Comment 3

[Shawn Wilsher :sdwilsh]

Attachment: SQLite Changes

Comment 4

[Shawn Wilsher :sdwilsh]

Attachment: Mozilla Changes

Comment 5

[Shawn Wilsher :sdwilsh]

Attachment: SQLite Changes

non-empty this time...

Comment 6

[Shawn Wilsher :sdwilsh]

http://hg.mozilla.org/projects/places/rev/40e57eeb5fb0
http://hg.mozilla.org/projects/places/rev/588eb6345711

Comment 7

[Shawn Wilsher :sdwilsh]

This is doing great on the places branch plus stuff in comment 1, so blocking final!

Comment 8

[Dietrich Ayala (:dietrich)]

Blocks a beta 8 blocker, approval granted to land.

Comment 9

[Marco Bonardo [:mak]]

http://hg.mozilla.org/mozilla-central/rev/40e57eeb5fb0
http://hg.mozilla.org/mozilla-central/rev/588eb6345711
http://hg.mozilla.org/mozilla-central/rev/70d77529b523

Comment 10

[Shawn Wilsher :sdwilsh]

In order to fix bug 595530 on branch, we'll need to take this on 1.9.2.

Comment 11

[Mark Banner (:standard8)]

(In reply to comment #10)
> In order to fix bug 595530 on branch, we'll need to take this on 1.9.2.

Thunderbird would also like to see us get this so that we can then upgrade to 3.7.4 which will fix some significant upgrade crashers.

Comment 12

[Daniel Veditz [:dveditz]]

(In reply to comment #10)
> In order to fix bug 595530 on branch, we'll need to take this on 1.9.2.

That bug is several months old. If it were a true blocker shouldn't it have been nominated before now? Landing a library upgrade a few days before code-freeze doesn't give anyone a great feeling. Can this wait for the next .x release? How big is the set of changes? How long has this upgraded version been tested? About a month of nightlies, but a lot of that was holidays. Did it make it into a beta?

Comment 13

[Shawn Wilsher :sdwilsh]

(In reply to comment #12)
> That bug is several months old. If it were a true blocker shouldn't it have
> been nominated before now? Landing a library upgrade a few days before
> code-freeze doesn't give anyone a great feeling. Can this wait for the next .x
> release? How big is the set of changes? How long has this upgraded version been
> tested? About a month of nightlies, but a lot of that was holidays. Did it make
> it into a beta?
It's in beta 9, which is why I hadn't requested it sooner.  I'm fine with waiting to make sure things look OK on beta 9 with it (so it can slip to the next .x release).

Comment 14

[Wayne Mery (:wsmwk)]

I thought these sql landings were considered very low risk given the sql test regime.

comment 11 refers to this blocking landing 3.7.4 bug 618315, which is #3 and #10 thunderbird topcrash for many months. IMO it doesn't need to wait.

Comment 15

[Wayne Mery (:wsmwk)]

(In reply to comment #14)
> I thought these sql landings were considered very low risk given the sql test
> regime.
> 
> comment 11 refers to this blocking landing 3.7.4 bug 618315, which is #3 and
> #10 thunderbird topcrash for many months. IMO it doesn't need to wait.

nevermind. There's zero possibility (I assume) bug 618315 will get into 1.9.2 before firefox 3.6.14 freeze - heck it's not even on trunk.  And from Thunderbird POV getting only 3.7.3 without 3.7.4 into 1.9.2 is of no value. Ludo or standard8 can correct, but I'm not aware of 3.7.3 blocking thunderbird issues other than bug 552050 crasher. 

In light of this, for thunderbird, these bugs need to block 1.9.2.next

Comment 16

[Mark Banner (:standard8)]

Comment on attachment 482998 [details] [diff] [review]
Mozilla Changes

Requesting approval for 1.9.2 branch - 3.7.3 has been on trunk since beta 9 and Thunderbird need to upgrade to sqlite 3.7.4 (via bug 618315) to fix some crashers.

Comment 17

[Marco Bonardo [:mak]]

Also Places has a longstanding performance problem from 1.9.2.10 that we would really like to fix (slowing down locationbar and all history queries)

Comment 18

[LegNeato]

We're taking bug 618315 for branch instead.