User Push date [To Local] | Changeset | Patch author — Commit message |
---|---|---|
bsmedberg@mozilla.com Wed Nov 10 15:29:47 2010 +0000 | bdbef533364f1cdecbbefe16d8d8d9f5e0fd8100 | Benjamin Smedberg — Final merge of bug 596451. a=blocking2.0 - Asynchronous layer-based painting on Windows. Various pieces r=josh/romaxa/karlt/jmathies/roc. ← 35 merge changesets [Collapse] |
30b383d73524f23ce5ea32b176c580077e21c0e9 | Benjamin Smedberg — Fix test_xulbrowser_plugin_visibility to be more robust and use fewer magic timeouts, review note from bug 583109 comment 52. | |
cff6e330a3deab267508bc44d73b81681131c7d4 | Benjamin Smedberg — Additional review fix for bug 583109 comment 49 (karlt). | |
949de4d79ea9dd68d17535aea975be5d64479b23 | Benjamin Smedberg — Address review comments for the test in bug 583109 by polling for visibility changes in the plugins. | |
d12c46b9e27fe367434b588954805578f5faac5c | Benjamin Smedberg — Fix review comments from bug 583109 (changeset 6ff3fcbb7845), comment 46. | |
be57c513c2ace26a60194acdca019e5ad7684345 | Benjamin Smedberg — Bug 607958 - Opacity rounding of 0.5 is different when using a layer and painting directly. Use a value of 0.8 which divides evenly into 255 to avoid rounding differences. Bug 602200 did not fix this rounding difference. r=roc | |
40cc37bb536ed7026b0f69215504c0ed69cabdbe | Benjamin Smedberg — Merge mozilla-central into the branch for asynchonous plugin painting (bug 596451), especially to pick up and test whether the fix for 602200 also fixes bug 607958. | |
4ab9ea27785d333afa1fd8f22c8222ee015b8807 | Benjamin Smedberg — Fix nsObjectFrame.cpp back to unix-style line endings. DOS endings crept in during the last merge for reasons I don't understand. | |
0aaabea2708e4ff27340b91a5337bc788231b1ae | Benjamin Smedberg — Merge mozilla-central into the branch for asynchronous plugin painting on Windows. | |
671bd739dbff035d5b1df3ce1e53e2974efdc3a2 | Benjamin Smedberg — With async plugin painting, crashtest 522512 is no longer valid (and the condition which caused it is entirely gone). Remove it. | |
df0cf7e8e9e6611f0a723c3239625bd5b1e94339 | Benjamin Smedberg — Test for bug 583109 - Add an .isVisible() method to the testplugin. test_visibility.html checks whether a visibility: hidden plugin suspends and resumes painting properly. test_xulbrowser_plugin_visibility.xul tests whether switching tabs properly suspends and resumes plugin painting. r-pending=karlt | |
6ff3fcbb784501e3dc241e511aeea04ff759a468 | Benjamin Smedberg — Bug 606285 and Bug 583109 - Send the "visible" notification when painting with async layers, so that when a plugin goes from invisible to visible it starts sending asynchronous paints. Suspend sending surface updates for asynchronous plugins when they are invisible, and set the empty clipRect so that the plugin knows that drawing is suspended per the new API. Assume that plugins are visible again when a Paint or BuildLayer call is received. r-pending=karlt | |
24a2e0ac428f81256fc54025a8a9437dafde37b8 | Jan Arne Petersen — Bug 583109 - Add visibility notification for plugins by setting the NPWindow.clipRect to [0, 0, 0, 0] when the plugin is not visible, either in a background tab, or visibility: hidden. r=karlt Patch rebased against trunk by Benjamin Smedberg. | |
167349934e8dbc73de2dcf7eef7174574d6b8bd0 | Benjamin Smedberg — Followup to bug 596451 part E (cset 37188e297b07) - double-pass painting is still used for in-process plugins, so fix the test expectations accordingly. | |
3db23492203ff76d97ade9e85436b99a165d5040 | Benjamin Smedberg — Merge mozilla-central to the branch for bug 596451. | |
9adc8f43d987fdf4ff0bd5945fbe9753f9c08bed | Benjamin Smedberg — Fix testplugin painting issues from bug 596451 part J - when doing non-solid painting, paint to x/y, not 0/0. | |
005779cf3e419a4613459e764c90700bec6ad421 | Benjamin Smedberg — Bug 596451 part K - Make sure we set up the same plugin quirks for async plugin rendering that we do in AnswerNPP_SetWindow (fixes Flash/Silverlight context menus and mouse pointers), r=jmathies | |
ac385c4d7b3704c533e14fc1d4fb552e8959ed37 | Benjamin Smedberg — Bug 596451 part J - Fix our test plugin to paint without using GDI functions which overwrite the opacity incorrectly, r=jmathies | |
5d416c488505e30c1d1c2446e4e9cadca4b0e444 | Benjamin Smedberg — Build bustage from bug 596451 part I - always return a value from PluginInstanceChild::ReadbackDifferenceRect. | |
9f1f861a9ff742ace1c25ec97c273eeaa5e326ab | Benjamin Smedberg — Bug 607002 - Classes forward-declared as structs in IPDL-generated code cause MSVC warnings. Forward-declare them as classes. r=cjones | |
c6a351a89dde8ee22010da6be19da3f9177bdb01 | Benjamin Smedberg — Bug 596451 part H - Make our quirk of silverlight transparent by default active in the plugin process, to match the logic in nsNPAPIPluginInstance. This makes the Microsoft Surface website paint correctly. r=jimm | |
900cdb9c89529e731edabd78f853a3458ff6d13c | Benjamin Smedberg — Bug 596451 part G - Fix a condition where FixupWindow is called too early and doesn't know that we are a windowless plugin yet. Call FixupWindow from nsPluginInstanceOwner::CreateWidget, r=roc | |
eda1909e1416781c4ab9576180179d3b8c4e732c | Benjamin Smedberg — Bug 596451 part F - fix mouse/window positioning and remove the Silverlight quirk by changing the logical->device offset before painting, r=jimm | |
88f740a74afc33cc44ee08cbab6175f2f7b0e575 | Benjamin Smedberg — Bug 596451 part I - Read back from the current Windows surface, instead of repainting the entire client area every time, r-pending=jmathies | |
37188e297b073acf5d8319dddf5e8c1aa15722f3 | Benjamin Smedberg — Bug 596451 part E - Implement transparency using DIBs with builtin transparency. Flash at least does this correctly, and since nobody else really uses windowless transparency, this is going to be the Way It Is. r-pending=jimm/karlt | |
425bd6be7289f7439c3968fb6afbd407aa80c0fe | Benjamin Smedberg — More build bustage from Bug 596451 part D. Explicitly initialize the const structures for GCC. | |
866e9298ee159c57b2a434b199b39d5a811cc061 | Benjamin Smedberg — Non-Windows build bustage fix from bug 596451 part D - base::SharedMemoryHandle is not available to IPDL on non-Windows, so use typedefs. | |
94324cad0457d50276216fec9cd21bdbc191580e | Benjamin Smedberg — Bug 596451 part D - Hook up asynchronous plugin painting on Windows. This part allows opaque plugins to paint correctly. r=jmathies | |
da433c7320abe71c4d2d51bffd6a213ef61e3ac4 | Benjamin Smedberg — Bug 606976 - Don't map the memory from a SharedDIB: CreateDIBSection creates another mapping, and we don't actually use the mapping for anything, r=jimm | |
f7b85d9a0900caf0983551fb0ff0a579c212524e | Benjamin Smedberg — Review comment from bug 596451 part C - the XSync is required so that pending composition using this surface is complete before the surface is re-used by the plugin. | |
f409549bea96b3c19bd55d28f16479cfddb13431 | Benjamin Smedberg — Bug 596451 part C - simplify the behavior of the returning surface from Show, because it's only used to hand back ownership of IPDL shared memory. | |
6435ea1c2fbaca0806f06b6ee6957ce2519108c3 | Benjamin Smedberg — Merge mozilla-central with the asynchronous plugins painting branch. | |
7c2451323c882ef1fd7415aa597dbc3517da8fec | Benjamin Smedberg — Merge mozilla-central into asynchronous plugins branch. | |
cda406eada98c9727943a320f2e246623fa51164 | Benjamin Smedberg — Bug 596451 part B - UseAsyncPainting really shouldn't need to dig very far or use an outparam, r=romaxa | |
67f89755fa94da8fd15dcad9b549d56c09f36d69 | Benjamin Smedberg — Bug 596451 part A - fix PluginInstanceChild to use bools, r=jimm/josh |