adac41015b
Bug 462432 - Leak mAsyncExecutionMutex
...
r=sdwilsh
2008-11-27 12:34:40 -08:00
287ddd5ebc
Addressing review nit for bug 464202 that was improperly addressed in changeset 59d84ac24da2
...
original patch r=mrbkap, a=beltzner
2008-11-20 14:11:03 -05:00
26fa978b71
Address followup nit for bug 464202.
2008-11-19 11:54:29 -08:00
6017aa4aa8
Bug 457743 - Automatic wrapper creates cycles, and thus leaks. r=dcamp, r=asuth, a=beltzner for 1.9.1b2
2008-11-19 11:33:00 -08:00
14720845a3
Bug 464202 - nsGlobalWindow leak running browser-chrome Mochitests. r=mrbkap a=beltzner
2008-11-19 01:11:30 -05:00
deb7217ab3
Bug 461344 - abort if the version of sqlite we are using is not new enough
...
This adds a runtime check to ensure that the version of SQLite that is running
is the same as the one that we were compiled with. If the check fails, we will
display a prompt to the user, and abort.
r=bsmedberg
2008-10-31 20:38:58 -04:00
f653780142
Bug 451815 - Need a more reliable way to indicate SQLITE_BUSY
...
This changeset adds a new error result for storage (NS_ERROR_STORAGE_BUSY) that
maps directly to when SQLite would return SQLITE_BUSY.
r=asuth
--HG--
rename : storage/src/mozStorage.h => storage/src/mozStoragePrivateHelpers.h
2008-10-30 18:50:00 -04:00
801290a84c
Bug 462434 - Transaction semantics are bad
...
This propagates errors in starting and ending transactions.
r=asuth
2008-10-30 18:46:25 -04:00
da47747760
Bug 454740 - Asynchronous storage should batch/chunk results
...
This changeset batches results obtained by the async storage API so we are not
flooding the calling thread with so many events.
r=asuth
2008-10-29 13:13:32 -04:00
0430a4d084
Bustage fix (= missed additional change) for
...
Bug 460635 - There should be one async execution thread per mozStorageConnection connection; v1.2; r=(bugmail + sdwilsh)
2008-10-29 05:09:23 +01:00
2bb839d2b7
Bug 460635 - There should be one async execution thread per mozStorageConnection connection; v1.2; r=(bugmail + sdwilsh)
2008-10-29 04:53:19 +01:00
dd6c7ee933
Bug 458998 - AsyncExecute helper classes can complete() before they are added to mPendingEvents
...
This changeset also fixes a few issues where we were holding a lock while and
kills off a few other race conditions that were noticed in the course of fixing
this bug.
r=bent
2008-10-24 18:30:07 -04:00
18700696dc
Bug 461263 - Race condition exists with mozStorageTransaction
...
While the race was technically "safe", we could have possibly wrongly tried to
commit or rollback. This changeset makes us operate sanely in all conditions.
r=bent
2008-10-24 16:50:01 -04:00
d28587dd1d
Bug 458811 - Allow for multiple statements to be executed at in a transaction asynchronously
...
This changeset allows consumers to execute a series of statements, in order,
asynchronously in a transaction.
r=dcamp
sr=sicking
2008-10-13 18:45:40 -04:00
af33d57a63
Bug 458756 - Use raw sqlite3_stmt pointers instead of the mozStorageStatement object for async statements
...
This ends up saving us a bunch of work when we execute a statement
asynchronously, and makes bug 458811 easier to do.
r=dcamp
2008-10-10 17:41:31 -04:00
17a0817cd9
comment change to kick off new builds
...
rs=me
2008-10-09 20:29:14 -04:00
d29d8419b3
Bug 453722 - Do not need to safeguard more than one call to Init for storage service
...
Bug 449822 made it so services can not be initialized more than once when being
called from multiple threads. Storage used to protect against this, but no
longer needs to do this (which this changeset makes us do).
r=bent
2008-09-25 14:28:29 -04:00
1234432adf
Bug 456489 - WinCE Compiler keyword interface used in mozStorageRow.cpp. patch by wolfe, r=sdwilsh
2008-09-23 11:03:48 -07:00
7216b611ec
Bug 456471 - mozIStorageError IDL file needs ERROR undefined for WinCE / WinMobile. patch by wolfe, r=sdwilsh
2008-09-23 11:02:55 -07:00
65b3bb8092
Bug 455202 - executeAsync without listener/callback results in segfault when error encountered; test part; r=sdwilsh
2008-09-15 18:50:21 +02:00
8d86adfc37
Bug 455202 - executeAsync without listener/callback results in segfault when error encountered; fix part; r=sdwilsh
2008-09-15 18:49:18 +02:00
3d079626e8
Bustage fix - double semi-colon.
2008-09-12 17:40:55 -04:00
78e4ede55b
Bug 452897 - Automatically provide the wrapper for JavaScript
...
This changeset adds a language helper to mozStorageStatement so it gets the
attributes and methods that mozIStorageStatementWrapper has. We can now mark
the wrapper as depreciated.
r=vlad
r=mrbkap
--HG--
rename : storage/test/unit/test_storage_statement_wrapper.js => storage/test/unit/test_statement_wrapper_automatically.js
2008-09-12 17:30:41 -04:00
3e49f59cca
Bug 452897 - Automatically provide the wrapper for JS
...
This changeset has no code changes - it just separates three classes that were
all contained in one file into three files (with appropriate headers). This is
being done to make the actual code changes for the bug simpler.
rs=me
2008-09-12 17:30:37 -04:00
4c52e158c3
Bug 452836 - Optimize calls to sqlite3_prepare_v2
...
This changeset makes all calls pass -1 for length which indicates that the
string is null terminated.
r=Neil
2008-09-03 12:20:10 -04:00
6b5d53f655
Bug 407216 - DOM quick stubs - faster paths for top N DOM methods (r+sr=jst, security r=mrbkap, build r=bsmedberg)
...
* * *
* * *
* * *
2008-08-19 21:38:24 -05:00
6dea9a020c
Bug 452836 - Optimize calls to sqlite3_prepare_v2
...
This makes sqlite not copy the sql string when preparing it, saving us a string
copy every time.
r=bholley
2008-08-29 17:19:23 -04:00
cfba414480
Bug 452833 - Do not loop if we get SQLITE_SCHEMA
...
We used to have to try again if we got SQLITE_SCHEMA back from sqlite when we
prepare a statement. However, this isn't needed with the new v2 interface,
which we now use.
r=bholley
2008-08-29 17:17:01 -04:00
7132ee9652
Bug 451058 - Address post-review comments.
...
This addresses some comments that were brought up after the code had already
landed.
r=Neil
2008-08-29 16:46:14 -04:00
9b3a321600
Bug 452841 - Remove hack in mozStorageStatement::Initialize
...
This removes some code that was at one time needed to work around a sqlite but
that no longer exists.
r=vlad
2008-08-29 16:42:12 -04:00
97a796274f
Bug 451463 - mozStorageStatement needs threadsafe refcounting
...
This changeset makes mozStoageStatement use threadsafe refcounting. The class
is still not threadsafe, but JS consumers cannot create a statement off of the
main thread since it would be garbage collected on the main thread.
r=bent
2008-08-27 13:11:02 -04:00
d0e6e8b795
Bug 448607 - Includes and comments are not accurate in mozStorageBackground.*
...
This fixes some comments that are wrong and removes unneeded includes.
rs=me; there's really no change here.
2008-08-27 13:05:03 -04:00
abf12d8076
Backing out changeset cf6c811e1272 (bug 407216) due to debug assertions.
2008-08-19 22:28:26 -05:00
09f9ac1d0a
Bug 407216 - DOM quick stubs - faster paths for top N DOM methods (r+sr=jst, security r=mrbkap, build r=bsmedberg)
...
* * *
* * *
2008-08-19 21:38:24 -05:00
b4c01f2f50
Bug 450665: test_bug-365166.js creates file in /tmp instead of in build directory. r=sdwilsh
2008-08-16 16:17:47 +01:00
78b08dae0c
Make mozIStorageConnection threadsafe - bug 448476 r=robarnold
2008-08-12 17:50:25 -07:00
9a06360391
Avoid returning false without setting an exception. bug 429521, r=sdwilsh
2008-08-04 10:04:27 -07:00
31a2fea1d6
Test that the object has the right class, even when called from native code. bug 444233, r=shaver
2008-07-16 19:29:06 +02:00
9b3bae8da4
Bug 444821 - Re-enable test_storage_statement_executeAsync.js
...
This fixes the test to run properly on slower running systems (like the
tinderbox machines).
r=vlad
2008-07-15 11:04:52 -04:00
56be74f21c
Disable stupid test (stupidity is my own fault). Bug 444821 to fix it properly
...
and re-enable it.
2008-07-11 17:29:45 -04:00
a583d673c5
Bug 429986 - Provide an option for database to be asynchronous
...
This adds unit tests for the async storage work.
r=vlad
2008-07-11 15:47:38 -04:00
35ba2ed351
Bug 429986 - Provide an option for database access to be asynchronous.
...
This adds a method to mozIStorageStatement to allow for a statement to execute
asynchronously and report to a callback. For writes, this can move fsyncs,
which can be painful, off of the main thread.
r=vlad
sr=shaver
2008-07-11 15:47:33 -04:00
3da654e7c2
Bug 435994 - Implement ResultSet and Row for mozIStorageStatement::executeAsync
...
These creates the result set and row objects to be used with the async storage
API added in bug 429986.
r=vlad
sr=shaver
2008-07-11 15:47:28 -04:00
fb401724a9
Bug 435995 - Implement an nsIVariant object storage can use.
...
This creates a templated C++ class for storage to map SQLite data types to the
correct nsIVariant.
r=robarnold
r=vlad
2008-07-11 15:47:21 -04:00
3436da9283
Bug 444347 - Make NSPR logging a bit more useful
...
This adds additional NSPR logging to storage. Logging is now done when a sql
statement is prepared, a connection is opened, and a connection is closed.
Additionally, when sqlite3_trace is called, we also log the address of the
connection so you know what queries are running on which database. The address
is also logged at connection open so you can perform the mapping.
r=vlad
2008-07-11 15:00:58 -04:00
140f31f75e
Bug 434838 - mozStorageConnection uses nsDataHashtable when it should use nsInterfaceHashtable
...
This makes mozStorageConnection no longer manually AddRef and Release user
defined functions.
r=sdwilsh
r=bent
2008-07-08 18:38:25 -04:00
bad1189195
Bug 442762 - Annotate mozStorageStatementScoper with NS_STACK_CLASS, r=sdwilsh
2008-06-30 21:03:29 -04:00
faddef0df3
Bug 434796
...
Remove mozIStorageStatement::initialize
r=vlad
sr=shaver
2008-06-11 11:00:09 -04:00
13420bfff8
Bug 434247
...
Remove mozStorageStatement's mStatementString
r=vlad
2008-06-11 10:25:20 -04:00
6691f61f94
Backed out changeset 5ccc3423220f
2008-06-04 12:55:44 -04:00
Andrew Sutherland
Shawn Wilsher
Jeff Walden
Serge Gautherie
Doug Turner
Jason Orendorff
Lukas Blakk
Blake Kaplan
Arpad Borsos
Benjamin Smedberg