Looks like we want this on branch for Thunderbird, so requesting blocking.

Thanks Shwan. semantics, but this is sev=major (or critical) because it blocks thunderbird topcrashes (same could be said for bug 598306 Will SQLite3.7.3 be landing soon? And, will there need to be a baking period before 3.7.4 will land? Or will it be a quick hop with no wait?

(In reply to comment #2) > Will SQLite3.7.3 be landing soon? It will land when the places branch merges, which should be soon. > And, will there need to be a baking period before 3.7.4 will land? Or will it > be a quick hop with no wait? We'll want to let 3.7.4 bake on trunk before we land it on branch.

(In reply to comment #3) > (In reply to comment #2) > > Will SQLite3.7.3 be landing soon? > It will land when the places branch merges, which should be soon. > > > And, will there need to be a baking period before 3.7.4 will land? Or will it > > be a quick hop with no wait? > We'll want to let 3.7.4 bake on trunk before we land it on branch. ah, so 3.7.3 is in the places branch. I didn't ask that very well - so is the following what you envision? land places branch on trunk (with 3.7.3) bake needed?? land 3.7.4 on trunk bake land 3.7.4 on branch shake

(In reply to comment #4) > land places branch on trunk (with 3.7.3) > bake needed?? > land 3.7.4 on trunk > bake > land 3.7.4 on branch > shake Something like this, although bake time from 3.7.3 landing to 3.7.4 doesn't need to be big at all.

We need a timed plan to bring 3.7.4 to branch, bug 595530 has still to be fixed there.

(In reply to comment #5) > (In reply to comment #4) > > land places branch on trunk (with 3.7.3) > > bake needed?? > > land 3.7.4 on trunk > > bake > > land 3.7.4 on branch > > shake > Something like this, although bake time from 3.7.3 landing to 3.7.4 doesn't > need to be big at all. done - 3.7.3 landed 598306 comment 9 2010-12-16. hopefully 3.7.4 can land the week before thunderbird 3.1.8 builds start? (jan 21)

Comment on attachment 502038 [details] [diff] [review] Mozilla Changes >+++ b/db/sqlite3/README.MOZILLA >@@ -1,24 +1,25 @@ >-This is sqlite 3.7.3 >+This is sqlite 3.7.4 > >--- Shawn Wilsher <me@shawnwilsher.com>, 10/2010 >+-- Paul OâShannessy <paul@oshannessy.com>, 01/2011 > > See http://www.sqlite.org/ for more info. > > We have a mozilla-specific Makefile.in in src/ (normally no > Makefile.in there) that we use to build. > > To move to a new version: > > Simply copy the sqlite3.h and sqlite3.c files from the amalgamation of sqlite. > > Also copy test_quota.c from the full source package. > >-Be sure to update SQLITE_VERSION accordingly in $(topsrcdir)/configure.in. >+Be sure to update SQLITE_VERSION accordingly in $(topsrcdir)/configure.in as >+well as the version number at the top of this file. > > -- Shawn Wilsher <me@shawnwilsher.com> 01/2008 This is you, not me now ;) r=sdwilsh

Ben, the changes to test_quota.c cause the build to fail. Without the changes we build fine. The diff: http://hg.mozilla.org/try/diff/11130bade42e/db/sqlite3/src/test_quota.c The errors: /home/pao/mozilla/mozilla-central/storage/src/mozStorageStatementWrapper.cpp: In member function ‘virtual nsresult mozilla::storage::StatementWrapper::QueryInterface(const nsIID&, void**)’: /home/pao/mozilla/mozilla-central/storage/src/mozStorageStatementWrapper.cpp:66: warning: ‘mozIStorageStatementWrapper’ is deprecated (declared at ../../dist/include/mozIStorageStatementWrapper.h:157) /home/pao/mozilla/mozilla-central/storage/src/mozStorageStatementWrapper.cpp:66: warning: ‘mozIStorageStatementWrapper’ is deprecated (declared at ../../dist/include/mozIStorageStatementWrapper.h:157) In file included from /home/pao/mozilla/mozilla-central/storage/src/mozStorageService.cpp:58: /home/pao/mozilla/mozilla-central/db/sqlite3/src/test_quota.c: In function ‘int quotaOpen(sqlite3_vfs*, const char*, sqlite3_file*, int, int*)’: /home/pao/mozilla/mozilla-central/db/sqlite3/src/test_quota.c:312: error: invalid conversion from ‘void*’ to ‘quotaFile*’ /home/pao/mozilla/mozilla-central/db/sqlite3/src/test_quota.c: In function ‘int sqlite3_quota_set(const char*, sqlite3_int64, void (*)(const char*, sqlite3_int64*, sqlite3_int64, void*), void*, void (*)(void*))’: /home/pao/mozilla/mozilla-central/db/sqlite3/src/test_quota.c:672: error: invalid conversion from ‘void*’ to ‘quotaGroup*’ Is this something we should report upstream, something we should change in mozIStorageWrapper, or should we adjust the message in README.MOZILLA to say "maybe you should update test_quota.c"?

(In reply to comment #11) > Is this something we should report upstream, something we should change in > mozIStorageWrapper, or should we adjust the message in README.MOZILLA to say > "maybe you should update test_quota.c"? upstream!

FWIW, Ben said he hit this before, so it's not particular to 3.7.4

(In reply to comment #13) > FWIW, Ben said he hit this before, so it's not particular to 3.7.4 So how did it solve it earlier ? Does this blocks 3.7.4 from landing ?

(In reply to comment #14) > So how did it solve it earlier ? Does this blocks 3.7.4 from landing ? You just have to add two static_cast<> to the malloc lines that fail to build.

FWIW, only failures on Try were known random oranges and it doesn't look like there were any huge variations in Talos: http://perf.snarkfest.net/compare-talos/index.html?oldRevs=4a3866321a14&newRev=e1d9b2de1cbc&submit=true

Requesting blocking for 1.9.2.next on this bug. Thunderbird really needs this - we seem to get a peak of these crashes on every release and we'd really like to get this in for the next security release after the one that is just about to be cut. This will also need bug 598306 to land as well. The expectations in other bugs had been that these would land much sooner which is why I hadn't been pushing the issue until now.

Actually, based on the description of the update, I would like this version in Firefox 4 as well, Places can get a performance win dropping the wal shared memory file and going back to exclusive mode.

I have the patch ready to go for m-c. I haven't done any testing at all on 1.9.2. The patch would be different (1.9.2 is still on sqlite 3.7.1).

Pushed http://hg.mozilla.org/mozilla-central/rev/c249c39e6f21 and http://hg.mozilla.org/mozilla-central/rev/73c1b27a22e9.

Comment on attachment 502036 [details] [diff] [review] SQLite Changes Requesting approval for 1.9.2 branch - 3.7.4 has been on trunk since beta 10 and Thunderbird need to upgrade to fix some crashers.

Comment on attachment 502036 [details] [diff] [review] SQLite Changes a=LegNeato for 1.9.2.16

Comment on attachment 502038 [details] [diff] [review] Mozilla Changes a=LegNeato for 1.9.2.16

1.9.2 is still at SQLite 3.7.1 so it's a slightly different upgrade. We'll need an updated patch.

http://hg.mozilla.org/releases/mozilla-1.9.2/rev/403a2284aa27 http://hg.mozilla.org/releases/mozilla-1.9.2/rev/4888247d34c5