bmalloc's fastMallocGoodSize(size) just returns 'size', so it's of no utility anymore.
Created attachment 245595 [details] Patch
Fast malloc can't give us the bucket size efficiently?
s/Fast Malloc/bmalloc
Comment on attachment 245595 [details] Patch k
Comment on attachment 245595 [details] Patch Rejecting attachment 245595 [details] from commit-queue. Failed to run "['/Volumes/Data/EWS/WebKit/Tools/Scripts/webkit-patch', '--status-host=webkit-queues.appspot.com', '--bot-id=webkit-cq-01', 'apply-attachment', '--no-update', '--non-interactive', 245595, '--port=mac']" exit_code: 2 cwd: /Volumes/Data/EWS/WebKit Last 500 characters of output: cpp patching file Source/WTF/wtf/FastMalloc.cpp Hunk #1 FAILED at 219. Hunk #2 FAILED at 334. Hunk #3 succeeded at 2753 (offset -33 lines). 2 out of 3 hunks FAILED -- saving rejects to file Source/WTF/wtf/FastMalloc.cpp.rej patching file Source/WTF/wtf/FastMalloc.h Hunk #1 succeeded at 35 with fuzz 2. patching file Source/WTF/wtf/Vector.h Failed to run "[u'/Volumes/Data/EWS/WebKit/Tools/Scripts/svn-apply', '--force', '--reviewer', u'Anders Carlsson']" exit_code: 1 cwd: /Volumes/Data/EWS/WebKit Full output: http://webkit-queues.appspot.com/results/5577581122289664
Created attachment 246147 [details] Patch for landing
Comment on attachment 246147 [details] Patch for landing Clearing flags on attachment: 246147 Committed r179743: <http://trac.webkit.org/changeset/179743>
All reviewed patches have been landed. Closing bug.
For reasons I don't yet understand, this has absolutely destroyed our ability to run DumpRenderTree on Windows. I'll see if this method was being used manually in the Windows code somewhere.
Re-opened since this is blocked by bug 141335
I think the problem is that our Open Source test bots are running with WebKitQuartzCoreAdditions.dll that was linked with the WTF::fastMallocGoodSize() symbol. We need to roll out updates to WKQCA before we land this.
(In reply to comment #11) > I think the problem is that our Open Source test bots are running with > WebKitQuartzCoreAdditions.dll that was linked with the > WTF::fastMallocGoodSize() symbol. > > We need to roll out updates to WKQCA before we land this. I think the easier way is to just leave the symbol in there as a functioning stub, and just remove the use from Vector.h right now.
This also broke WebKit nightlies, which may be more challenging to fix. See bug 141344.
I guess it's no more challenging to fix if it just returns size though.
Comment on attachment 245595 [details] Patch Cleared Anders Carlsson's review+ from obsolete attachment 245595 [details] so that this bug does not appear in http://webkit.org/pending-commit.
We only have this call-site using it: https://searchfox.org/wubkat/source/Source/WebCore/platform/sql/SQLiteDatabase.cpp#96 void SQLiteDatabase::useFastMalloc() { #if ENABLE(SQLITE_FAST_MALLOC) int returnCode = sqlite3_config(SQLITE_CONFIG_LOOKASIDE, 0, 0); RELEASE_LOG_ERROR_IF(returnCode != SQLITE_OK, SQLDatabase, "Unable to reduce lookaside buffer size: %d", returnCode); static sqlite3_mem_methods fastMallocMethods = { [](int n) { return fastMalloc(n); }, fastFree, [](void *p, int n) { return fastRealloc(p, n); }, [](void *p) { return static_cast<int>(fastMallocSize(p)); }, [](int n) { return static_cast<int>(fastMallocGoodSize(n)); }, [](void*) { return SQLITE_OK; }, [](void*) { }, nullptr }; returnCode = sqlite3_config(SQLITE_CONFIG_MALLOC, &fastMallocMethods); RELEASE_LOG_ERROR_IF(returnCode != SQLITE_OK, SQLDatabase, "Unable to replace SQLite malloc: %d", returnCode); #endif }