Commit Graph

349 Commits

Author SHA1 Message Date
Dan Witte
0533b2cfca Bug 572223 - too much cookies.sqlite io. Part 3: sync read per domain. r=sdwilsh 2010-07-29 12:17:42 -07:00
Steve Fink
aed0c5459e Bug 580019: Fix functiontimer crash when opening db with NULL filename 2010-08-02 14:29:59 -07:00
Neil Rashbrook
0b700a41bb Backout previous changes to try to fix bustage 2010-08-02 00:30:12 +01:00
Neil Rashbrook
f1b32deb08 Bug 522770 followup to fix shared build bustage r=khuey 2010-08-02 00:00:56 +01:00
Dan Witte
dfcfc976fb Revert bug 572223 to fix creation of duplicate cookies. 2010-08-01 12:24:38 -07:00
Kyle Huey
932ea79b49 Bug 522770: Link directly to object files instead of intermediate static libs. Patch originally written by Ted Mielczarek. Checking into a CLOSED TREE as a bustage fix. reviews pending.
--HG--
extra : rebase_source : 54a506a5815dcdedaf3f61ad7d6bdd0529d7f5f7
2010-08-01 01:48:21 -07:00
Robert Sayre
073bde8d2b Merge mozilla-central to tracemonkey. 2010-07-30 21:57:41 -07:00
Dan Witte
e25f157ff9 Bug 572223 - too much cookies.sqlite io. Part 3: sync read per domain. r=sdwilsh 2010-07-29 12:17:42 -07:00
Robert Sayre
55a8586519 Merge mozilla-central to tracemonkey. 2010-07-27 13:09:04 -07:00
Shawn Wilsher
6d9ba47ee5 Bug 578939 - Storage service cannot be initialized off of the main thread.
r=asuth
a=blocking2.0
2010-07-24 11:21:17 -07:00
Luke Wagner
99411cc93d Bug 549143 - fatvals 2010-07-14 23:19:36 -07:00
Robert Sayre
70fc03b113 Merge mozilla-central to tracemonkey. 2010-07-02 17:25:52 -07:00
Jeff Walden
d52d02e814 Remove a couple checks in test code that had asserted that Object.getOwnPropertyNames wasn't implemented. r=orange 2010-07-02 00:26:09 -07:00
Shawn Wilsher
e59d57ca10 Bug 575667 - split about:memory reporting into more detailed sections for SQLite
r=vlad
a=vlad
2010-07-22 10:08:59 -07:00
Benjamin Smedberg
a249678c04 Final merge and landing of bug 568691 and dependent bugs to mozilla-central on a CLOSED TREE.
* bug 568691, register XPCOM components statically using manifests. r=mossop, with some test and build stuff r=vlad,sdwilsh
* bug 573557, rename categories with spaces to use hypens. r=jst
* bug 573739, Don't get the private browsing service during layout module initialization, r=ehsan

This initial merge does not include some work planned to land imminently:
* The extension manager still restarts. I need to solve a problem re-reading default preferences from extensions.
* MOZ_OMNIJAR is broken: there is a patch which I need to update in bug 568691.
* I will concurrently land a mobile-browser fix for component registration, but it may require some additional packaging changes.
2010-07-01 14:26:25 -04:00
Taras Glek
bb78f82fbe Bug 416330 - Suboptimal SQLite page size r=sdwilsh 2010-07-01 10:57:13 -07:00
Benjamin Smedberg
762def30bb Merge mozilla-central into the static-xpcom-registration branch. 2010-06-23 17:26:25 -04:00
Benjamin Smedberg
45af085fff Merge mozilla-central into the "branch" for but 568691. The omnijar XPT/component/chrome registration was removed during this merge, and will be re-added to the component manager shortly.
--HG--
rename : content/base/public/nsIChromeRegistry.idl => chrome/public/nsIChromeRegistry.idl
2010-06-21 11:41:42 -04:00
Shawn Wilsher
688f6b6cd2 Bug 571599 - Use sqlite3_unlock_notify
sr=vlad
r=bent
r=asuth
2010-06-21 14:23:00 -07:00
Marco Bonardo
fa6cc21c35 backout bug 547190 due to test failure 2010-06-22 03:49:47 +02:00
Marco Bonardo
0393cc87f3 Bug 547190 - AsInt64 (and other AsXXX cpp helpers) ignores GetInt64 failures. r=asuth 2010-06-22 02:01:33 +02:00
Benjamin Smedberg
f13550a209 Bug 568691 part B - mechanical changes to in-tree binary modules needed to get them building and registering correctly. After this patch, xpcshell appears to work. 2010-06-10 14:11:40 -04:00
Phil Ringnalda
9256e9cab3 Remove test from backed out bug 416330 so the CLOSED TREE can actually pass tests 2010-06-28 18:19:39 -07:00
Shawn Wilsher
5aafb87384 Backout of bug 416330 so we can freeze for the beta on a CLOSED TREE. 2010-06-28 16:31:47 -07:00
Taras Glek
ce040e2786 Bug 416330 - Suboptimal SQLite page size r=sdwilsh
--HG--
extra : rebase_source : 053c7b1d975304ed7d402bb619874c889c71425c
2010-06-28 12:13:26 -07:00
Shawn Wilsher
4e43c7eae6 Bug 571581 - Propagate SQLITE_CONSTRAINT as an nsresult
r=asuth
sr=vlad
2010-06-16 09:59:24 -07:00
Dave Townsend
07c355da2b Bug 570529: mozIStorageStatement.execute() should reset itself even if an error occurs. r=sdwilsh 2010-06-07 15:06:56 -07:00
Saint Wesonga
3ed57717d1 Bug 492625 - "Stop using underscores to start #defines" [r=vlad]
--HG--
extra : rebase_source : d685fb8bc375ba4fcf97742298e9361ebfe9b851
2010-06-01 08:38:00 -04:00
Jeff Walden
52312bac03 Bug 518663 - ES5: Object.getOwnPropertyNames. r=jorendorff 2009-09-24 14:33:14 -07:00
Chris Jones
8095e43052 Backed out changeset c60e3e48ea38 2010-06-24 19:56:14 -05:00
Taras Glek
065ac23871 Bug 416330 - Suboptimal SQLite page size r=sdwilsh 2010-06-24 16:39:24 -07:00
Ehsan Akhgari
d6b32f425f Bug 567254 - Crash in Connection::getFilename for memory-based databases; r=sdwilsh
--HG--
extra : rebase_source : 9fb1300af8c6b35a4e1aa229af032cd08a76e721
2010-05-20 20:21:09 -04:00
Ehsan Akhgari
a87b134220 Bug 560647 - add startup timeline instrumentation; r=vlad 2010-05-19 19:22:19 -04:00
Vladimir Vukicevic
d562978a21 b=566442; add memory reporters for sqlite; r=sdwilsh 2010-05-19 13:46:08 -07:00
Shawn Wilsher
38ecfb2f5c Test for bug 562866. r=jorendorff r=asuth
--HG--
rename : storage/test/unit/test_js_helpers_enumerate.js => storage/test/unit/test_js_helpers.js
2010-05-11 14:14:54 -07:00
timeless@mozdev.org, Jason Orendorff
4bea23532d Bug 562866 - StatementParams::NewResolve mishandles JSVAL_IS_STRING(aId). r=sdwilsh.
--HG--
extra : rebase_source : f9e1ea925fa38a26c806ba6dbddc52ef83646d55
2010-05-06 13:56:39 -05:00
Daniel Holbert
8d40d741e4 Bug 560383, part 2: (storage directory) Use "!!" for converting int to bool, to avoid MSVC compile warning. r=sdwilsh 2010-05-06 11:53:12 -07:00
Marco Bonardo
21198dad1f Bug 562787 - Fix getResultByName idl description to specify it needs an AS clause. r=asuth 2010-05-05 15:28:01 +02:00
timeless@mozdev.org
785392580d Bug 562855 - hook_sqlite_mutex in test_true_async.cpp should check the return value from sqlite3_initialize() and sqlite3_shutdown(), r=asuth
--HG--
extra : rebase_source : 648695de6bf0b68965cea43b27c7f9c089d8b945
2010-05-02 01:57:00 -07:00
Mitchell Field
9818f00bf2 Bug 560095 - Use mozilla::services::GetObserverService(). r=biesi,dveditz,gavin,josh,jst,mrbkap,roc,sdwilsh,shaver,sicking,smontagu,surkov 2010-04-29 18:59:13 +02:00
Marco Bonardo
01e6a8ee80 Bug 559678 - followup, make the warning less scarying. rs=sdwilsh 2010-04-22 14:53:56 +02:00
Marco Bonardo
7a4702f9fa Bug 559678 - Better errors for Storage transactions misuse. r=asuth 2010-04-16 14:31:18 +02:00
Andrew Sutherland
b4419b8d6f Bug 555124 - Intermittent "ASSERTION: AsyncStatementSpinner not thread-safe: '_mOwningThread.GetThread() == PR_GetCurrentThread()', file .../mozilla/storage/test/test_true_async.cpp, line 162". null out references on the main thread to avoid destruction on the async thread. r=sdwilsh. 2010-03-29 10:54:41 -07:00
Chris Jones
014cb688de Bug 550425: Bump up the timeout for expensive deadlock tests and print more info on failures. r=bsmedberg,sdwilsh 2010-03-24 19:12:32 -05:00
Andrew Sutherland
ddd3c4327f Bug 507414 - Add a createAsyncStatement API on mozIStorageConnection. r=sdwilsh, sr=vlad.
--HG--
rename : storage/src/mozStorageStatementJSHelper.cpp => storage/src/mozStorageAsyncStatementJSHelper.cpp
rename : storage/src/mozStorageStatementJSHelper.h => storage/src/mozStorageAsyncStatementJSHelper.h
rename : storage/src/mozStorageStatementParams.cpp => storage/src/mozStorageAsyncStatementParams.cpp
rename : storage/src/mozStorageStatementParams.h => storage/src/mozStorageAsyncStatementParams.h
2010-03-24 00:32:40 -07:00
Shawn Wilsher
5bdff598be Bug 536978 - Cookies should write to the database asynchronously.
This makes cookie insertion, updates, and deletions happen asynchronously off of
the main thread.  The only I/O done on the main thread is the initial loading of
the database at each startup.
r=dwitte
2010-03-12 07:18:38 -08:00
Shawn Wilsher
6371523ca7 Bug 552092 - Expose a length property on mozIStorageBindingParamsArray
Expose the length so that consumers don't have to track if they added anything
to the array themselves.
r=asuth
sr=rob_strong
2010-03-15 10:23:39 -07:00
Shawn Wilsher
b5e02f4c11 Bug 552003 - Passing an empty mozIStorageBindingParamsArray does not error
Throw instead of just asserting, which makes debugging this so much easier.
r=asuth
2010-03-15 10:23:39 -07:00
Marco Bonardo
28774cea88 Backing out bug 536978, due to a Storage assertion causing orange. CLOSED TREE 2010-03-12 17:41:28 +01:00
Shawn Wilsher
6fb0361405 Bug 536978 - Cookies should write to the database asynchronously.
This makes cookie insertion, updates, and deletions happen asynchronously off of
the main thread.  The only I/O done on the main thread is the initial loading of
the database at each startup.
r=dwitte
2010-03-12 07:18:38 -08:00
Shawn Wilsher
e7bf5fc1f2 Bug 496019 - mozilla::storage::Connection::Close can spin a nested event loop
Creates a new close method that must be used when using asynchronous statements,
and disallow Close from being called in that case.
r=asuth
sr=vlad
2009-11-09 09:58:34 -08:00
Dan Witte
ddfc2078d1 Fix regression in mozIStorageStatement.bindUTF8StringParameter such that it accepts arbitrary octets by
not converting string types. b=544496, r=sdwilsh
2010-02-22 13:57:45 -08:00
Justin Lebar
e730a1afc6 Bug 500328 - Implement History.pushState(), History.replaceState() methods. r=smaug, r=zpao, sr=sicking 2009-09-01 09:45:05 -07:00
Daniel Holbert
46ba50b0cb Backed out changeset dc7a04be6904 on suspicion of causing bug 543034. 2010-01-29 18:50:46 -08:00
Justin Lebar
a827ba17fc Bug 500328 - Implement History.pushState(), History.replaceState() methods. r=smaug, r=zpao, sr=sicking 2009-09-01 09:45:05 -07:00
Reed Loden
cdabfc45ae Backout patch from bug 489897 as the likely cause of pwmgr test failures. 2009-12-31 01:15:11 -06:00
David Dahl ext:(%2C%20Shawn%20Wilsher%20%3Csdwilsh%40shawnwilsher.com%3E)
4a25f2e3a8 Bug 489897 - implement StatementRow::NewEnumerate
for (let colName in stmt.row) now works!
r=sdwilsh
r=ddahl
sr=mrbkap
2009-10-12 17:49:00 -05:00
Shawn Wilsher
02575db216 Bug 494828 - Stop using our own mutexes and use SQLite's where possible.
Part 2: Use the helper object in a few places where we had our own mutexes.
r=asuth
2009-07-22 15:18:33 -07:00
Shawn Wilsher
9fc75352b9 Bug 525356 - windows debug unit tests: test_connection_executeAsync.js and test_statement_executeAsync.js failing frequently (fatal assertion)
Make sure to release the database connection on the thread it was opened on, not
the background thread!
r=asuth
a=beltzner
2009-12-02 10:53:46 -08:00
Shawn Wilsher
7dac22ddd2 Backed out changeset f91a016416d1 (bug 496019) 2009-12-01 15:05:52 -08:00
Shawn Wilsher
09b6e528f9 Bug 496019 - mozilla::storage::Connection::Close can spin a nested event loop
Creates a new close method that must be used when using asynchronous statements,
and disallow Close from being called in that case.
r=asuth
sr=vlad
a=dbaron

--HG--
extra : rebase_source : 8470e30ef4ca1e9fb516284cafb77b070e46fde3
2009-11-09 09:58:34 -08:00
Shawn Wilsher
9f62654fa8 Check to make sure that sqlite has been compiled with SQLITE_SECURE_DELETE (Bug 445164); r=asuth 2009-11-23 18:46:39 -05:00
L. David Baron
ab0caddb75 Backed out changeset 4a0bf3ffd767 due to test crashes. 2009-11-19 13:51:31 -08:00
Shawn Wilsher
0d37582e30 Bug 525356 - windows debug unit tests: test_connection_executeAsync.js and test_statement_executeAsync.js failing frequently (fatal assertion)
Release on the calling thread so when the connections destructor is called, it
is not on the background thread.
r=asuth
2009-11-19 11:49:20 -08:00
Blair McBride
b954b9cae2 Bug 528076 - Unsafe LIKE warning wrongly references mozIStorageConnection::escapeStringForLIKE. r=sdwilsh 2009-11-14 10:15:12 +01:00
Shawn Wilsher
116173a4b7 Backout of bug 496019 and bug 526601 due to window build bustage. 2009-11-09 12:16:39 -08:00
Shawn Wilsher
5281dcc164 Bustage fix for windows. 2009-11-09 10:35:26 -08:00
Shawn Wilsher
e6c35ff3f0 Bug 496019 - mozilla::storage::Connection::Close can spin a nested event loop
Creates a new close method that must be used when using asynchronous statements,
and disallow Close from being called in that case.
r=asuth
sr=vlad
2009-11-09 09:58:34 -08:00
Saint Wesonga
fe1b96674d Bug 479730 - mozStorageConnection::Close logs closing improperly. r=vlad 2009-11-05 14:52:55 +01:00
David Dahl
e55bd8baf9 Bug 524787 - crash [@mozStorageResultSet::GetNextRow(mozIStorageRow**) ]. r=sdwilsh 2009-10-31 11:38:31 +01:00
Ginn Chen
6a61d86670 Bug 522157 Make -D__FUNCTION__=__func__ global for Sun Studio on Solaris r=benjamin 2009-10-26 14:25:40 +08:00
L. David Baron
0e462f7e2d Disable test_deadlock_detector for failing on OS X. (Bug 523378) 2009-10-20 20:26:04 -04:00
Marco Bonardo
e68c36ec27 Bug 519326 - Variant.h(152) : warning C4244: '=' : conversion from 'PRInt64' to 'PRInt32', possible loss of data. r=sdwilsh 2009-10-09 22:15:19 +02:00
Dietrich Ayala
b682a0bba3 Bug 419893 - sort out PGO-triggered bugs (r=ted)
--HG--
extra : rebase_source : 028cd891e11df11abf768107f64b1d6bca541dca
2009-10-08 11:01:13 -07:00
Robert O'Callahan
6b41eca841 Backed out changeset 59a6c227df86 2009-10-02 15:23:36 +10:00
Shawn Wilsher
168578cfbc Bug 494828 - Stop using our own mutexes and use SQLite's where possible.
Part 2: Use the helper object in a few places where we had our own mutexes.
r=asuth

--HG--
extra : rebase_source : 3f2f703274df0405bc395191dec8f522e5aa4e40
2009-07-22 15:18:33 -07:00
Shawn Wilsher
e0356a19b4 Remove this test. It's not terribly useful (testing basic behavior of a
database which is covered by tests the SQLite group runs), and relies on a "bug"
that newer versions of SQLite fix.  rs=mconnor
2009-09-30 14:38:43 -07:00
Shawn Wilsher
3d91db712f Test only change to make tests clean up better by using the free profile directory given to us. No bug. r="duh, why didn't we do this sooner" 2009-09-30 14:17:06 -07:00
Paul O’Shannessy
5b3c2717f0 Bug 518440 - Deprecate mozIStorageStatementWrapper & mozIStorageStatement.step() in IDL. r=sdwilsh
--HG--
extra : rebase_source : 0ede87a6eb9a34475d9efafa433809e571a104d3
2009-09-25 15:32:21 -07:00
Shawn Wilsher
3889b768e7 Bug 507674 - Safer protection of AsyncExecuteStatements::mState and AsyncExecuteStatements::mCancelRequested
This is essentially a backout of bug 506805, followed by changing the locking
protection rules to only protect mCancelRequested writing on the calling thread,
and reading on the background thread.

--HG--
extra : rebase_source : 249a61e40c2fd140527f034e36e8c1a53518baac
2009-09-02 13:30:24 -07:00
Shawn Wilsher
7e79575c8f Bug 507199 - Fix race condition in finalizing asynchronous storage statements
(was "test_browserGlue_corrupt_nobackup_default.js, test_browserGlue_corrupt_nobackup.js, test_browserGlue_shutdown.js randomly crash")
Fixes a race condition with asynchronous statements that causes a crash.  We do
not always have access to the background thread, so we should not assume we
always can get it.
2009-08-25 23:54:56 -07:00
Benjamin Smedberg
2e9438fd2f Followup to bug 398573 - remove REQUIRES from the tree since it is no longer used... automatically generated patch, rs=ted 2009-08-25 08:59:31 -07:00
Ginn Chen
40156e2b51 Bug 507978 test_deadlock_detector.cpp failed to compile on Solaris r=bsmedberg 2009-08-11 14:56:58 +08:00
Jacek Caban
1f66701ce7 Bug 505734 - mozStorageStatementParams.cpp compilation fails on MinGW. r=sdwilsh 2009-08-09 22:30:38 +02:00
Akshay Gupta
fea6ebf827 Bug 433456 - mozIStorageFunction should be a [function]; r=sdwilsh 2009-08-06 20:00:39 +02:00
Shawn Wilsher
11b33f0ab8 Bug 506805 - Remove locking in AsyncExecuteStatements
This removes the use of the shared mutex in AsyncExecuteStatements.  We now rely
on PR_AtomicSet and the volatile keyword.  This results in zero lock contention
between the calling thread and the background thread if cancel is ever called.
r=asuth
r=bent
sr=vlad
2009-07-29 17:24:50 -07:00
Shawn Wilsher
9981a52b9b Bug 506959 - AsyncExecuteStatements assumes ms, but calls functions that assume microseconds
This fixes the code to use functions that expect milliseconds instead of
microseconds, and updates to use the new TimeStamp and TimeDuration class.
r=asuth
2009-07-28 13:18:10 -07:00
Shawn Wilsher
463b989609 Bug 506022 - Avoid obtaining the database mutex at all costs in Connection::ExecuteAsync
Stop calling any SQLite functions that would acquire the database mutex inside
Connection::ExecuteAsync.  Additionally, do not actually bind parameters when
the binding functions are called, but rather when we execute (so for async
statements, that will happen on the background thread creating no mutex
contention with the main thread).
r=asuth
2009-07-28 10:21:03 -07:00
Shawn Wilsher
3e9850b9ff Merge places to mozilla-central. 2009-07-27 08:50:26 -07:00
Curtis Bartley
1192731c18 Bug 498938 - Add Levenshtein Edit Distance function to Sqlite so we can use it in queries. r=sdwilsh sr=vlad 2009-07-24 11:39:34 -04:00
Shawn Wilsher
1515fe6be7 Merge places to mozilla-central. 2009-07-23 13:10:26 -07:00
Shawn Wilsher
7a0b8db179 Merge places to mozilla-central. 2009-07-23 07:58:59 -07:00
timeless@mozdev.org
ecebe2a338 Bug 491196 don't implement static functions in headers
r=sdwilsh
Moving JSValStorageStatementBinder to mozStoragePrivateHelpers as bindJSValue

storage/src/mozStorageStatementParams.h: At global scope:
storage/src/mozStorageStatementParams.h:78: warning: ‘PRBool mozilla::storage::JSValStorageStatementBinder(JSContext*, mozIStorageStatement*, int, jsval)’ defined but not used
storage/src/mozStorageStatementParams.h:78: warning: ‘PRBool mozilla::storage::JSValStorageStatementBinder(JSContext*, mozIStorageStatement*, int, jsval)’ defined but not used

You get this once for each file that includes a header that defines a static function which is not used in that file.
2009-07-22 17:31:03 -05:00
Shawn Wilsher
2681223754 Bug 505550 - Acquire the db's mutex when cloning statements in Connection::ExecuteAsync
Hold the lock around all our API calls so we don't lose so much to the
background thread.
r=asuth
2009-07-22 15:18:33 -07:00
Shawn Wilsher
c61e6f67fa Bug 494828 - Stop using our own mutexes and use SQLite's where possible.
Part 2: Use the helper object in a few places where we had our own mutexes.
r=asuth
2009-07-22 15:18:33 -07:00
Shawn Wilsher
dfa9ebd065 Bug 494828 - Stop using our own mutexes and use SQLite's where possible.
Part 1: Create helper objects to make using sqlite3_mutex safer and easier.
r=cjones

--HG--
rename : xpcom/tests/TestDeadlockDetector.cpp => storage/test/test_deadlock_detector.cpp
2009-07-22 15:18:33 -07:00
Shawn Wilsher
5fcdac946b Backout of bug 494828 and bug 505550 due to orange. 2009-07-23 10:34:54 -07:00
Benjamin Smedberg
266f5b1dec Fix bustage again from bug 494828 by linking against libxpcomglue_s like we were, but continue to link against libxpcom_core also. 2009-07-23 12:18:12 -04:00
Benjamin Smedberg
fa3ab235cc Fix bustage from bug 494828 by linking against libxpcom_core.so for tracemalloc symbols. 2009-07-23 11:42:32 -04:00
Shawn Wilsher
6631ab6c3b Bug 505550 - Acquire the db's mutex when cloning statements in Connection::ExecuteAsync
Hold the lock around all our API calls so we don't lose so much to the
background thread.
r=asuth
2009-07-22 15:18:33 -07:00