summaryrefslogtreecommitdiffstats
path: root/chrome/app/framework.order
Commit message (Collapse)AuthorAgeFilesLines
* Work around a ridiculous bug in ATS, a deprecated system framework.mark@chromium.org2011-08-181-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | ATS likes to write to memory it doesn't own on Mac OS X 10.7 ("Lion"). This happens in SendDeactivateFontsInContainerMessage, called by ATSFontDeactivate, used by Chrome. SendDeactivateFontsInContainerMessage has some really sloppy memory handling that can be avoided by making sure that a certain symbol, __CTFontManagerUnregisterFontForData, is available. Note that the system's CoreText framework has a _CTFontManagerUnregisterFontForData symbol, which is probably what ATS should be looking for, but that's actually an entirely different ridiculous bug in ATS. ATS seems to have this bug on 10.6 ("Snow Leopard") too, but we haven't noticed the corruption there. Maybe we're just lucky. BUG=93191, 90884 TEST=1. Visit http://www.justgiving.nl/ 2. Click the British flag in the top-right 3. Click the blue JustGiving logo in the top-left. 4. Wait for the page to finish loading, then repeat step 3. Expect: to be able to continue performing step 4 repeatedly. No sad tabs. No renderer crashes. No messages logged to the console about memory (malloc) errors Review URL: http://codereview.chromium.org/7655040 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@97320 0039d316-1c4b-4281-b951-d872f2087c98
* OOL code in CarbonPluginWindowTracker.erg@google.com2011-02-241-0/+2
| | | | | | | | | BUG=carnitas TEST=compiles Review URL: http://codereview.chromium.org/6581038 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@75968 0039d316-1c4b-4281-b951-d872f2087c98
* Re-land earlier patch that moves the NPAPI implementation from ↵brettw@chromium.org2010-12-211-7/+7
| | | | | | webkit/glue/plugins to webkit/plugins/npapi git-svn-id: svn://svn.chromium.org/chrome/trunk/src@69808 0039d316-1c4b-4281-b951-d872f2087c98
* Revert "Move the NPAPI files from webkit/glue/plugins to webkit/plugins/npapi"chase@chromium.org2010-12-211-7/+7
| | | | | | | | | | | | Manually reverting r69755, which broke the tree. BUG=none TEST=none TBR=dmaclach@chromium.org Review URL: http://codereview.chromium.org/5998002 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@69771 0039d316-1c4b-4281-b951-d872f2087c98
* Revert "Revert 69755 - Move the NPAPI files from webkit/glue/plugins to ↵chase@chromium.org2010-12-201-7/+7
| | | | | | | | | | | | | | | webkit/plugins/npapi" Manually reverting r69766, which was itself a failed revert of r69755. BUG=none TEST=none TBR=dmaclach@chromium.org Review URL: http://codereview.chromium.org/5996003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@69768 0039d316-1c4b-4281-b951-d872f2087c98
* Revert 69755 - Move the NPAPI files from webkit/glue/plugins to ↵brettw@chromium.org2010-12-201-7/+7
| | | | | | | | | | | | | | | webkit/plugins/npapi and put them in the webkit::npapi namespace. BUG=none TEST=none Review URL: http://codereview.chromium.org/6012002 TBR=brettw@chromium.org Review URL: http://codereview.chromium.org/5961004 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@69766 0039d316-1c4b-4281-b951-d872f2087c98
* Move the NPAPI files from webkit/glue/plugins to webkit/plugins/npapi and putbrettw@chromium.org2010-12-201-7/+7
| | | | | | | | | | them in the webkit::npapi namespace. BUG=none TEST=none Review URL: http://codereview.chromium.org/6012002 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@69755 0039d316-1c4b-4281-b951-d872f2087c98
* Update Native Client revision in DEPSmseaborn@chromium.org2010-11-081-0/+1
| | | | | | | | | | | | | | | | In order for tools/build/mac/verify_order to pass, we have to add operator delete(void*, void*) to chrome/app/framework.order. (The mangled name for this function is "__ZdlPvS_".) This gets defined as a side effect of gdp_rsp's use of STL in NaCl. BUG=http://code.google.com/p/nativeclient/issues/detail?id=469 BUG=http://code.google.com/p/nativeclient/issues/detail?id=583 TEST=nacl_ui_tests Review URL: http://codereview.chromium.org/4690001 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@65382 0039d316-1c4b-4281-b951-d872f2087c98
* Use an opaque handle type to pass WebPluginDelegateImpl across dylib boundariesthakis@chromium.org2010-10-281-6/+6
| | | | | | | | | | | This works around a wicked new clang linker optimization, but also feels cleaner in general. BUG=http://llvm.org/bugs/show_bug.cgi?id=8478 TEST=clang/mac builder stays green with newer clang version Review URL: http://codereview.chromium.org/4191003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@64289 0039d316-1c4b-4281-b951-d872f2087c98
* Mac: app mode loader (shim).viettrungluu@chromium.org2010-05-181-0/+3
| | | | | | | | | | | | Define an interface using which Chrome can be loaded from a minimal loader. Produce a binary for this loader (which can be put into a bundle). BUG=13148 TEST=not yet Review URL: http://codereview.chromium.org/2066004 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@47576 0039d316-1c4b-4281-b951-d872f2087c98
* Rename FakePluginWindowTrackerstuartmorgan@chromium.org2010-03-021-7/+7
| | | | | | | | | | | Removes confusion with the new "fake window" code related to the GPU process, and makes the name better reflect the reality of how the Carbon plugin windows are actually used since we no longer expect to replace them entirely with interposing. BUG=none TEST=none (no behavioral changes) Review URL: http://codereview.chromium.org/660407 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@40459 0039d316-1c4b-4281-b951-d872f2087c98
* Plumb Carbon SetCursor through.avi@chromium.org2010-01-251-0/+1
| | | | | | | | | BUG=http://crbug.com/32703 TEST=when hiding cursor (Flash 10.0), cursor hides (but doesn't stay hidden; that gets fixed later) Review URL: http://codereview.chromium.org/552132 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@37028 0039d316-1c4b-4281-b951-d872f2087c98
* Send Cocoa NPAPI plugins an initial window activation eventstuartmorgan@chromium.org2010-01-211-0/+1
| | | | | | | | | | | | | Until we have real window activation status, just send a window focus event so that Cocoa plugins always think their window is active (as is currently the case with Carbon plugins). Also wire up the Carbon interposing on window activation checks to talk to the delegate, so once the delegate has correct information Carbon will Just Work. BUG=31847 TEST=Flash 10.1 should track the mouse when the plugin is loaded, without having to click in it first. Review URL: http://codereview.chromium.org/552067 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@36754 0039d316-1c4b-4281-b951-d872f2087c98
* (Mac) Allow the tracking of the "active" plugin delegate separate from the ↵avi@chromium.org2010-01-191-8/+8
| | | | | | | | | | | fake windows. BUG=http://crbug.com/20717 TEST=none; no visible change Review URL: http://codereview.chromium.org/549085 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@36533 0039d316-1c4b-4281-b951-d872f2087c98
* (Mac) Intercept (Carbon) cursor setting by plugins and forward it on properly.avi@chromium.org2010-01-151-0/+1
| | | | | | | | | BUG=http://crbug.com/20717 TEST=as in bug; this covers old Flash Review URL: http://codereview.chromium.org/550062 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@36416 0039d316-1c4b-4281-b951-d872f2087c98
* Fix clicking in QuickTime content on the Macstuartmorgan@chromium.org2010-01-131-0/+2
| | | | | | | | | | | QuickTime uses FindWindow in its click handling to get the WindowRef, so we need to interpose it to give back the plugin window (which isn't actually on screen). This makes clicks work except for the menu button in the control strip. BUG=24952 TEST=Click/double-click in a QuickTime video to pause/play, or use the control strip (except for the menu button). Review URL: http://codereview.chromium.org/536047 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@36182 0039d316-1c4b-4281-b951-d872f2087c98
* Add Cocoa window interposing for plugin processstuartmorgan@chromium.org2009-12-081-4/+4
| | | | | | | | | | | | | | Adds support for noticing window show/hide done via Cocoa APIs. Also adds interposing on a couple more Carbon calls that I came across while debugging. Refactors to make things a bit less messy: - Moves responsibility for re-activating the browser when the last plugin window closes into the browser, to simplify that logic. - Moves utility method for activating the plugin process into the plugin, rather than the interpose library, so the Cocoa code can use it too. - Add a data-holder and a couple of wrapper methods to reduce code duplication in the window show/hide handling. BUG=28009, 29321 TEST=Plugin-opened windows (Gmail attachments, full-screen Flash, etc.) should work as expected again. Review URL: http://codereview.chromium.org/465111 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@34074 0039d316-1c4b-4281-b951-d872f2087c98
* Add nacl .globl symbols to framework.order.jrg@chromium.org2009-12-041-0/+4
| | | | | | | | TBR=mark@chromium.org, thomasvl@chromium.org Review URL: http://codereview.chromium.org/464046 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@33871 0039d316-1c4b-4281-b951-d872f2087c98
* List the gcov symbols to get the mac code coverage bot happy.thomasvl@chromium.org2009-12-041-0/+17
| | | | | | | | TEST=code coverage bot doesn't compile fail on verify_order. BUG=none Review URL: http://codereview.chromium.org/464036 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@33820 0039d316-1c4b-4281-b951-d872f2087c98
* Add an .order file to ensure that no global text symbol winds up at an addressmark@chromium.org2009-11-191-0/+41
higher than _ChromeMain, so that CrashReporter stacks and other symbolized stacks aren't so confusing. Add a tool to verify that nothing violates this ordering requirement. BUG=28257 TEST=verify_order, which is part of the build Review URL: http://codereview.chromium.org/414003 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@32539 0039d316-1c4b-4281-b951-d872f2087c98
generated by cgit v1.1 at 2025-02-12 10:36:29 +0000