Tobias Netzel
ef4327744e
Bug 812932 - Have EndianMacros.h recognize __BIG_ENDIAN__ preprocessor macro - r=bbondy
2012-11-19 09:09:17 -08:00
Jeff Muizelaar
7406d0ccb6
Bug 812602. Don't decode jpegs progressively when we have all the data. r=joe
...
Decoding progressively is slower because we duplicate work like color
conversion and color correction for each outputted scan. We already suppress
paints when have all of the data, so we shouldn't do this extra work.
--HG--
extra : rebase_source : e6dd733546e1bb9e6e3279d9ceea0b4c0a88e049
2012-11-18 20:18:52 -05:00
Josh Matthews
41e92624f2
Bug 803125 - Notify status tracker when image request is destroyed. r=joe
2012-11-17 13:33:20 +00:00
Ryan VanderMeulen
d64a665956
Backed out changeset fb3e2fb82000 (bug 802390) for mochitest bc orange.
2012-11-16 16:58:11 -05:00
Joe Drew
9bbca08240
Bug 802390 - Gracefully handle shutting down a decoder that hasn't had a chance to do any work. r=jrmuizel
2012-11-16 11:43:24 -08:00
Jeff Muizelaar
5737de8f11
Bug 795648. Add a commment describing why we need to call RequestDecode. r=mwoodrow
...
--HG--
extra : rebase_source : bd0b47c361eee7c0ed6f00f92eda8aabc5ef08e9
2012-11-16 11:00:06 -08:00
Jeff Muizelaar
0827ec3aaa
Bug 794957. Use imgLoader instead of imgILoader. r=joe
...
This makes use of imgLoader instead of imgILoader in more places.
--HG--
extra : rebase_source : 6b391e5ecc2fd9558516c7493116e7e7852b71b3
2012-11-16 10:24:58 -08:00
Matt Woodrow
26e7462dbf
Bug 795648 - Don't synchronously decode images during painting. r=joe
2012-11-16 10:04:29 -08:00
Honza Bambas
4b1361b1c5
Bug 734062 - Followup: fixed unused varialble use, r=joedrew
2012-11-14 13:56:56 +01:00
Josh Matthews
9416cf2c50
Bug 803121 - Avoid using virtual functions in imgRequestProxy's destructor. r=joe
2012-10-19 15:39:38 -04:00
Jeff Muizelaar
20d1de1575
Bug 811461. Switch IMAGE_DECODE_LATENCY to microseconds. r=taras
...
Currently this is recording in ms when we expect all of the values to be <5ms
I believe this was just a typo.
--HG--
extra : rebase_source : ed71dc9bd7dc9ef0903e2e84152b44ba13d97adf
2012-11-13 15:38:54 -08:00
Seth Fowler
bd282adc07
Bug 809181 - SVGDocumentWrapper::UpdateViewportBounds() should check if anything has changed instead of relying on callers. r=dholbert
2012-11-08 14:54:47 -05:00
Honza Bambas
a335e4be51
Bug 734062 - stale image in the image cache when an updated version is fetched into HTML5 offline cache in background, r=joedrew
2012-11-08 19:54:00 +01:00
Josh Matthews
27b8218689
Bug 805013 - Paper over the browser-chrome crashes from certain circumstances that run ~RasterImage off the main thread. r=joe
2012-11-07 17:25:27 -05:00
Josh Matthews
f9c594dd58
Bug 805013 - Diagnostic bandaid to ensure that all RasterImages are removed from the discard tracker. r=joe
2012-11-07 14:03:10 -05:00
Jeff Walden
793acbdae0
Bug 803439 - Add removeFrom() that takes a list and asserts the element is initially present in it. r=jlebar
...
--HG--
extra : rebase_source : c2aad35774a71167f97f940fb416735862b78950
2012-10-23 17:43:23 -07:00
aceman
cc2af4f753
Bug 808379 - fix Assertion failure: isEmpty(), at ../../dist/include/mozilla/LinkedList.h:258 causing Thunderbird crash since landing of fix to bug 803688. r=jlebar
...
Clear the sDiscardableImages linked list in DiscardTracker::Shutdown() so that the list destructor finds an empty array, which is required after bug 803688.
2012-11-04 23:20:52 +00:00
Seth Fowler
4388c1b02d
Bug 807556 - Remove imgIRequest::loadImage's aRequest argument. r=joe
2012-11-03 23:04:07 -04:00
Kyle Huey
1482122303
Bug 798491: Add an option to stick all chrome JSMs/JS components in the same compartment. r=mrbkap,philikon
...
--HG--
extra : rebase_source : 98337b6a8c07d05e8c961a452dd05a7d75c3c60b
2012-10-31 09:13:28 -07:00
Vicamo Yang
8297cdabd6
Backout 67cb43bb8865: Breaks B2G Marionette
2012-10-31 16:35:11 +08:00
Kyle Huey
bdf47895bb
Bug 798491: Add an option to stick all chrome JSMs/JS components in the same compartment. r=mrbkap,philikon
2012-10-30 12:28:11 -07:00
Ed Morley
777eed8444
Backout a145ded68994, e0cf397089ec & 1545e91c658e (bug 798491) for bustage on a CLOSED TREE
2012-10-30 17:02:31 +00:00
Kyle Huey
aa1997a51a
Bug 798491: Add an option to stick all chrome JSMs/JS components in the same compartment. r=mrbkap,philikon
2012-10-30 08:26:12 -07:00
Nathan Froyd
bd55a07f0e
Bug 806618 - rewrite PR_NewLogModule calls to not generate static initializers; r=ehsan
2012-10-29 19:32:10 -04:00
Jacek Szpot
0b35b0e3d8
Bug 792180 - Replace NS_{UN,}LIKELY with MOZ_{UN,}LIKELY; r=ehsan
2012-10-26 15:32:10 +02:00
Jeff Muizelaar
d3f69a6169
Bug 805119. Add a note about the deprecation of findEntryProperties. r=joe
...
DONTBUILD
2012-10-24 14:40:41 -04:00
Josh Matthews
ce46a7b003
Bug 802485 - Quiet intermittent assertion for further investigation. r=orange
2012-10-22 15:40:05 -04:00
Nick Hurley
a115c7b5bc
Bug 536324 part 2 - Change implementers of nsIChannel. r=sworkman
2012-10-22 10:51:07 -07:00
Abhishek Potnis
2509c81393
Bug 795511: Removed usages of PR_BIT and PR_BITMASK from the tree; r=ehsan
2012-10-20 07:31:43 +05:30
Josh Matthews
05a584c750
Bug 802485 - Remove special handling for canceled image request proxies that are changing owners. r=joe
2012-10-22 12:44:53 -04:00
Joe Drew
eb26e24d1b
Bug 802168 - Don't shut down the decoder when we're destroying RasterImages; instead, just remove ourselves from the decoding list. r=jrmuizel,jdm
2012-10-22 13:16:57 -04:00
Phil Ringnalda
397a0510c4
Bug 761049 - Back out 6ba8fbe719cc to reenable browser_bug666317.js on Linux
2012-10-19 21:59:18 -07:00
Seth Fowler
775198e7ab
Bug 803349 - Restore NS_DECL_IMGICONTAINER to VectorImage / RasterImage. r=joe
2012-10-19 13:27:11 -07:00
Isaac Aggrey
cbb17fe7b2
Bug 794510: Part 2: Use more stdint types instead of protypes.h types; r=ehsan
2012-10-11 18:38:04 -05:00
Nicholas Nethercote
6669152d43
Bug 744686 - Comment out unused |alpha_bits| variable to fix a compile warning. r=jmuizelaar.
...
--HG--
extra : rebase_source : 48be31fe3eab889e6f7930f44b96140ba076c389
2012-10-17 15:54:25 -07:00
Joe Drew
c6ec6d5484
Bug 801358 - Part 2 - Account for downscaled frames in about:memory. r=jrmuizel
...
--HG--
extra : rebase_source : c5a507f0f5d4fa5b700f19cdb359fd8281baa80d
2012-10-13 22:17:40 -04:00
Joe Drew
67ad3c2619
Bug 801358 - Part 1 - Discard downscaled frames. r=jrmuizel
...
--HG--
extra : rebase_source : 1b699918323e7c116521cbde579aafb07846b2f6
2012-10-13 22:17:30 -04:00
Joe Drew
d8f7ba2fd2
Bug 795940 - Part 0.5 - Move ScaleRequest and the runners into the CPP file so consumers don't have to care about them. BY THE DEMAND OF AND r=jrmuizel
...
--HG--
extra : rebase_source : eeb96eb9680017d2be9ee99ac0c760249a74bcbe
2012-10-12 18:24:47 -04:00
Joe Drew
38c28c41f2
Bug 795940 - Part 0.4 - Only try to downscale images that have a single instance on a web page, since that's all we support anyhow. r=jrmuizel
...
--HG--
extra : rebase_source : 1bda15c8edb7e2d1d3ca26354ce8429640b03b43
2012-10-12 18:20:25 -04:00
Joe Drew
573a5425f8
Bug 795940 - Part 0.3 - Support stopping requests (on a best-effort basis). r=jrmuizel
...
--HG--
extra : rebase_source : 793d04fa7954bbf060f01436bc0d257edd05a28a
2012-10-12 18:20:22 -04:00
Joe Drew
99d0a521a3
Bug 795940 - Part 0.2 - Change from singleton ScaleWorker and DrawWorker (and on-demand allocated ScaleRequest) to nsIRunnables, allocated on demand, that know how to run themselves. r=jrmuizel
...
--HG--
extra : rebase_source : 9b98ed62eaf5fdddf6ee117ec5f26d049d752eaf
2012-10-12 18:20:19 -04:00
Joe Drew
78dd322d79
Bug 795940 - Part 0.1 - Allocate ScaleRequests dynamically and only hold on to a ScaleResult in RasterImage. Let ScaleRequests only hold a weak pointer to RasterImage so their lifetimes aren't bound. Finally, make ScaleRequests hold on to references to their surfaces instead of imgFrames. r=jlebar,jrmuizel
...
--HG--
extra : rebase_source : 335837e469e8318f0a70907060470b0a0e14ef21
2012-10-10 11:35:23 -04:00
Daniel Holbert
155b4147de
Bug 802409: Mark a bunch of /image subdirectories as FAIL_ON_WARNINGS. r=joe
2012-10-17 08:32:12 -07:00
Daniel Holbert
461f07928c
Bug 802458: Remove useless "NS_SUCCEEDED(rv)" in icon/gtk/nsIconChannel.cpp, and declare rv down where it's used. r=karlt
2012-10-16 22:18:10 -07:00
Josh Matthews
2c140b0eaa
Bug 802144 - Don't delete memory used for image frames while the decoder could still try to use it. r=joe
2012-10-16 22:45:14 -04:00
Josh Matthews
23b999d286
Bug 801405 - Avoid attempting to optimize frames that no longer exist. r=joe
2012-10-15 18:06:43 -04:00
Josh Matthews
f1cf8a6ed3
Bug 801701 - Fix references to non-existent requests in status trackers for extracted frames. r=joe
2012-10-15 16:05:01 -04:00
Ehsan Akhgari
4770edb259
Silence a compiler warning about initialization order of RasterImage members, no bug
...
--HG--
extra : rebase_source : a109b2d20f362add2dbb38a6b2037adfab69941e
2012-10-13 17:28:39 -04:00
Patrick McManus
f77e276cbe
bug 801013 - imgRequest::AdjustPriority wrongly adjusts loadgroup instead of channel r=joedrew
2012-10-13 12:47:36 -04:00
Jeff Muizelaar
491b02a2e0
Bug 799335. Fix unintenional image flash on tab switch. r=joe,dholbert
...
This was caused by bug 792199.
There are two parts to this fix:
1. Always decode during StartDecoding even if we already have a Decoder
(hopefully this is safe) This was a bug in the original patch.
2. After calling StartDecoding() we recheck if we're complete and abort if
we're not doing a SYNC_DECODE. Before this regression we would usually be done
decoding the image because of the decoding we did when we Locked all of the
images on tab switch.
--HG--
extra : rebase_source : 9db93075aad2f45c6cc70d27ef67468b4507642c
2012-10-12 16:04:32 -04:00