There's no reason nsThreadPool needs to use a reentrant monitor for
locking its event queue. Having it use a non-reentrant one should be
slightly more efficient, both in the general operation of the monitor,
and that we're not performing redundant locking in methods like
nsThreadPool::Run. This change also eliminates the only usage of
nsEventQueue::GetReentrantMonitor.
This is straightforward mapping of PR_LOG levels to their LogLevel
counterparts:
PR_LOG_ERROR -> LogLevel::Error
PR_LOG_WARNING -> LogLevel::Warning
PR_LOG_WARN -> LogLevel::Warning
PR_LOG_INFO -> LogLevel::Info
PR_LOG_DEBUG -> LogLevel::Debug
PR_LOG_NOTICE -> LogLevel::Debug
PR_LOG_VERBOSE -> LogLevel::Verbose
Instances of PRLogModuleLevel were mapped to a fully qualified
mozilla::LogLevel, instances of PR_LOG levels in #defines were mapped to a
fully qualified mozilla::LogLevel::* level, and all other instances were
mapped to us a shorter format of LogLevel::*.
Bustage for usage of the non-fully qualified LogLevel were fixed by adding
|using mozilla::LogLevel;| where appropriate.
This is straightforward mapping of PR_LOG levels to their LogLevel
counterparts:
PR_LOG_ERROR -> LogLevel::Error
PR_LOG_WARNING -> LogLevel::Warning
PR_LOG_WARN -> LogLevel::Warning
PR_LOG_INFO -> LogLevel::Info
PR_LOG_DEBUG -> LogLevel::Debug
PR_LOG_NOTICE -> LogLevel::Debug
PR_LOG_VERBOSE -> LogLevel::Verbose
Instances of PRLogModuleLevel were mapped to a fully qualified
mozilla::LogLevel, instances of PR_LOG levels in #defines were mapped to a
fully qualified mozilla::LogLevel::* level, and all other instances were
mapped to us a shorter format of LogLevel::*.
Bustage for usage of the non-fully qualified LogLevel were fixed by adding
|using mozilla::LogLevel;| where appropriate.
This is straightforward mapping of PR_LOG levels to their LogLevel
counterparts:
PR_LOG_ERROR -> LogLevel::Error
PR_LOG_WARNING -> LogLevel::Warning
PR_LOG_WARN -> LogLevel::Warning
PR_LOG_INFO -> LogLevel::Info
PR_LOG_DEBUG -> LogLevel::Debug
PR_LOG_NOTICE -> LogLevel::Debug
PR_LOG_VERBOSE -> LogLevel::Verbose
Instances of PRLogModuleLevel were mapped to a fully qualified
mozilla::LogLevel, instances of PR_LOG levels in #defines were mapped to a
fully qualified mozilla::LogLevel::* level, and all other instances were
mapped to us a shorter format of LogLevel::*.
Bustage for usage of the non-fully qualified LogLevel were fixed by adding
|using mozilla::LogLevel;| where appropriate.
This patch was generated by a script. Here's the source of the script for
future reference:
find . \( -iname "*.cpp" -o -iname "*.h" \) | \
xargs -n 1 sed -i "s/nsRefPtr<nsIRunnable>/nsCOMPtr<nsIRunnable>/g"
The way idle nsThreadPool threads wait with a timeout doesn't work well with
shutdown from nsThreadManager.
nsThreadPool doesn't need to use nsIThread for its threads. The nsIThread
interface is not useful for threads running in an nsThreadPool.
The nsIEventTarget on the nsIThreadPool should be used for dispatching events,
not an interface on the individual threads, and the threads don't need an
nsEventQueue because they use the nsEventQueue on the nsThreadPool.
Shutdown of single event threads is easier than running nested event loops for
nsIThreads, avoiding the multilevel nested event loop situations when several
threads finish and are shutdown. While the ThreadFunc is running, a nested
event loop is still used in Shutdown() in case some consumers might need it
and because that is the documented API.
This also simplifies thread creation, avoiding races that could mean there was
temporarily an extra thread or more.
--HG--
extra : transplant_source : %F7%14%16%12%EF%E9%84%19D%26%3C%FE%1F%EC%FF%A3%BAG%C4%F3
While threads are processing their initial runnable, which is the nsThreadPool, it will not be
destroyed. The threads remove themselves from mThreads before returning from Run().
This includes removal of changes from bug 900711.
--HG--
extra : transplant_source : %EF%E9%94L%E9d%F5%99b%09r%CB%024%86%88%BA%EFi%29
Bug 645263, part 0: Count sync primitive ctor/dtors. r=dbaron
Bug 645263, part 1: Migrate content/media to mozilla:: sync primitives. r=doublec
Bug 645263, part 2: Migrate modules/plugin to mozilla:: sync primitives. sr=bsmedberg
Bug 645263, part 3: Migrate nsComponentManagerImpl to mozilla:: sync primitives. sr=bsmedberg
Bug 645263, part 4: Migrate everything else to mozilla:: sync primitives. r=dbaron
Bug 645263, part 5: Remove nsAutoLock.*. sr=bsmedberg
Bug 645263, part 6: Make editor test be nicer to deadlock detector. r=ehsan
Bug 645263, part 7: Disable tracemalloc backtraces for xpcshell tests. r=dbaron
Bug 646259: Fix nsCacheService to use a CondVar for notifying. r=cjones
Bug 645263, part 0: Count sync primitive ctor/dtors. r=dbaron
Bug 645263, part 1: Migrate content/media to mozilla:: sync primitives. r=doublec
Bug 645263, part 2: Migrate modules/plugin to mozilla:: sync primitives. sr=bsmedberg
Bug 645263, part 3: Migrate nsComponentManagerImpl to mozilla:: sync primitives. sr=bsmedberg
Bug 645263, part 4: Migrate everything else to mozilla:: sync primitives. r=dbaron
Bug 645263, part 5: Remove nsAutoLock.*. sr=bsmedberg
Bug 645263, part 6: Make editor test be nicer to deadlock detector. r=ehsan
Bug 645263, part 7: Disable tracemalloc backtraces for xpcshell tests. r=dbaron
Bug 646259: Fix nsCacheService to use a CondVar for notifying. r=cjones
Nathan Froyd
Randell Jesup
Eric Rahm
Carsten "Tomcat" Book
Wes Kocher
Ryan VanderMeulen
Cervantes Yu
JW Wang
Daniel Holbert
Chih-Kai (Patrick) Wang
Patrick Wang (Chih-Kai Wang)
Birunthan Mohanathas
Shihua Zheng
Chris Peterson
Chris Pearce
Benjamin Smedberg
Ehsan Akhgari
Karl Tomlinson
Ben Turner
Joshua Cranmer
Michal Novotny
Honza Bambas
Gervase Markham
Michael Wu
Chris Pearce
Chris Jones