We can hit the limit very easily when writing javascript bytecode as alternative data to the cache entry because all data is written at once but CacheFileOutputStream splits it into chunks which are then written on a backgound thread. 40MB was chosen because bytecode is usually 4x-10x larger than the original data, so it can occupy most of the cache entry which is limited to 50MB.
The patch changes default cache size on desktop and mobile. The smart cache size calculation is changed to grow faster. It also introduces a cache size limit for users who have enabled clearing cache on shutdown, which should reduce number of shutdown crashes (bug 1356853).
When writing to alt-data output stream fails for whatever reason, we now try to
truncate alternative data and keep the original data instead of dooming the
whole entry. The patch also changes how is the predicted size passed to the
cache. Instead of a dedicated method it's now an argument of openOutputStream
and openAlternativeOutputStream methods which fail in case the entry would
exceed the allowed limit.
There's no need to allocate a completely new nsCString when all we want
to do is parse a character string into an integer. We can allocate a
dependent string instead, which will avoid some memory churn.
The change to RootAccessible.cpp fixes an obvious bug introduced in bug 741707.
The visibility changes in gfx/thebes are because NS_DECL_ISUPPORTS has a
trailing "public:" that those classes were relying on to have public
constructors.
MozReview-Commit-ID: IeB8KIJCGhU
This macro is identical to NS_INTERFACE_MAP_END and encourages the
reader to think that there's something extra-special threadsafe about QI
implementations that use the macro, when in reality there's nothing of
the sort.
We must close cache iterators before shutting down CacheIndex because they hold reference to CacheIndex which causes that CacheIndex is not destroyed when it's dereferenced in CacheIndex::Shutdown.
--HG--
extra : amend_source : d74243f4eb4f4073ecc72ac9b331091456a91504
* A memory leak occurs when this happens in the content process
* I added an assertion that we only create it in the parent process
MozReview-Commit-ID: 1UTyusRg0qx
--HG--
extra : rebase_source : 400cfb21ee094ad33e15681f213af6dbe20ebbf5
* A memory leak occurs when this happens in the content process
* I added an assertion that we only create it in the parent process
MozReview-Commit-ID: 1UTyusRg0qx
--HG--
extra : rebase_source : 646c1c82d141cadc6c1b19843b4cc750e1a1ce59
For netwerk/cache2/CacheFileInputStream.cpp:148 and netwerk/protocol/http/nsHttpHeaderArray.cpp:358,
missing "()" in the if statement.
For netwerk/base/rust-url-capi/test/test.cpp:29, netwerk/streamconv/converters/nsHTTPCompressConv.cpp:297,
and netwerk/streamconv/converters/nsHTTPCompressConv.cpp:300, null pointer will be returned but the
original memory buffer will not be freed if |realloc| fails. We should remember the original memory
buffer and free it if error is detected.
MozReview-Commit-ID: 2ggXsL73jYV
--HG--
extra : rebase_source : e47e41f2b37f717207bd13990efead22a14db1c0
We should not be declaring forward declarations for nsString classes directly,
instead we should use nsStringFwd.h. This will make changing the underlying
types easier.
--HG--
extra : rebase_source : b2c7554e8632f078167ff2f609392e63a136c299
These are all easy cases where an nsXPIDLCString local variable is set via
getter_Copies() and then is null checked. The patch uses IsVoid() to replace
the null checks (and get() and EqualsLiteral() calls to replace any implicit
conversions).
--HG--
extra : rebase_source : 484ad42a7816b34b86afbe072e04ba131c1619c6
These are all easy cases where an nsXPIDLCString local variable is set via
getter_Copies() and then is only used in ways that nsCStrings can also be used
(i.e. no null checks or implicit conversions to |char*|).
In every case the patch trivially replaces the nsXPIDLCString with an
nsCString. (Also, there are a couple of unused nsXPIDLCString variables that
the patch simply removes.)
This parameter isn't used by any implementation of onDispatchedEvent,
and keeping the parameter makes later refactorings in this bug more difficult.
MozReview-Commit-ID: 90VY2vYtwCW
We have a minimum requirement of VS 2015 for Windows builds, which supports
the z length modifier for format specifiers. So we don't need SizePrintfMacros.h
any more, and can just use %zu and friends directly everywhere.
MozReview-Commit-ID: 6s78RvPFMzv
--HG--
extra : rebase_source : 009ea39eb4dac1c927aa03e4f97d8ab673de8a0e
Michal Novotny
Kris Maglione
Chris Peterson
Andi-Bogdan Postelnicu
Sebastian Hengst
Tristan Bourvon
Nika Layzell
Honza Bambas
Miko Mynttinen
Tom Ritter
Nathan Froyd
Cosmin Sabou
Boris Zbarsky
Masatoshi Kimura
Michal Novotny
Sylvestre Ledru
Nicholas Nethercote
Tom Schuster
Valentin Gosu
Ho-Pang Hsu
Shih-Chiang Chien
Eric Rahm
Bill McCloskey
Carsten "Tomcat" Book
Bevis Tseng
Kartikaya Gupta