ba9f118869366fc9774d991497780a0fdc91ca61: Bug 1630383: Remove direct access to wrapped Value in v8::Object r=tcampbell
Iain Ireland <iireland@mozilla.com> - Fri, 17 Apr 2020 17:46:08 +0000 - rev 524858
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1630383: Remove direct access to wrapped Value in v8::Object r=tcampbell In preparation for the next patch, which changes the internal representation of v8::Object, this patch removes all direct access (in derived classes) to the `value_` member of v8::Object. Instead, we now use accessor methods. This patch also changes a few places that were using `JS::Value` as a parameter to take `const JS::Value&` instead. Differential Revision: https://phabricator.services.mozilla.com/D71360
b4bd410fbcf6470b5d139f5d833dea03777e1f16: Bug 1630383: Fix CheckNotBackReferenceIgnoreCase r=jandem
Iain Ireland <iireland@mozilla.com> - Mon, 20 Apr 2020 16:15:29 +0000 - rev 524857
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1630383: Fix CheckNotBackReferenceIgnoreCase r=jandem In each iteration of the inner loop, we compare two characters. If they are not the same, but the first character is a letter, then we convert both characters to lower case and see if they are the same. At least, that's what we intend to do. We are accidentally loading a second copy of the first letter instead. This means that any two letters will compare equal in a case-insensitive backreference. Computers are hard. Differential Revision: https://phabricator.services.mozilla.com/D71359
a468a05ee7b527c39879b4cc0745c9a21b11848d: Bug 1630383: Fix off-by-one error in CheckPosition r=jandem
Iain Ireland <iireland@mozilla.com> - Mon, 20 Apr 2020 11:45:53 +0000 - rev 524856
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1630383: Fix off-by-one error in CheckPosition r=jandem When changing from inputStartMinusOne to inputStart, I got the boundary condition wrong for backwards-looking CheckPosition. This broke some test262 regexp tests. Depends on D71357 Differential Revision: https://phabricator.services.mozilla.com/D71358
82240b0220855ba49de30ea5fddedcf7d1517758: Bug 1630383: Make successHandler optional r=jandem
Iain Ireland <iireland@mozilla.com> - Mon, 20 Apr 2020 11:40:15 +0000 - rev 524855
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1630383: Make successHandler optional r=jandem We can compile regexps that we know ahead of time will not match. (For example, compiling a regexp containing a two-byte character to match latin-1 strings.) Depends on D71356 Differential Revision: https://phabricator.services.mozilla.com/D71357
22127e9ef811c656d012728845e4d4fd80d4799d: Bug 1630383: Add irregexp to ignore list for implicit constructor static analysis r=sfink
Iain Ireland <iireland@mozilla.com> - Fri, 17 Apr 2020 17:18:57 +0000 - rev 524854
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1630383: Add irregexp to ignore list for implicit constructor static analysis r=sfink The V8 code in the new engine uses implicit constructors with wild abandon, and it doesn't seem like a good use of time to try upstreaming a patch to remove them. Depends on D71355 Differential Revision: https://phabricator.services.mozilla.com/D71356
0c7019227612c142d14b5f5e18277606b45d0975: Bug 1630383: Appease hazard analysis r=sfink
Iain Ireland <iireland@mozilla.com> - Fri, 17 Apr 2020 21:01:58 +0000 - rev 524853
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1630383: Appease hazard analysis r=sfink For a variety of reasons, some better than others, we call into irregexp with unrooted GC pointers. (For example, the input string is unrooted, at least in part because rooting it would complicate Ion.) Everything is properly guarded with AutoCheckCannotGC, which means that we need to convince the hazard analysis that nothing untoward is going on. There are three obstacles to this: 1. ICU uses a pile of function pointers internally. To keep things interesting, they change the names of their functions each release (so ICU 66 has "icu_66::UnicodeSet::applyFilter()" and "uprv_malloc_66", but a newer version would update the number.) I handled these in annotations.js by adding an `isICU` check and throwing regexps at the problem. 2. The call to the generated code happens via function pointer, so we need to get the hazard analysis to ignore it. I used the solution from the old import (adding an AutoSuppressGCAnalysis). (See ExecuteRaw in RegExpAPI.cpp.) 3. The RegExpCompileData struct that V8 uses to communicate information between the parser and the compiler has one member that is an Object (~= JS::Value) and another than is a Handle. Because of this, the hazard analysis also wants to make sure that we don't GC during compilation. This is mostly not a problem, except for one place where we call a merge sort with a function pointer as comparator. I added SortConsecutiveAtoms to the whitelist to fix this. (See https://searchfox.org/mozilla-central/rev/97cb0a90bd053de87cd1ab7646d5565809166bb1/js/src/new-regexp/regexp-compiler-tonode.cc#498) Depends on D71354 Differential Revision: https://phabricator.services.mozilla.com/D71355
026a6f7e37ef3d158e698e992f3a7438bd896c17: Bug 1630383: Shim simplifications r=mgaudet
Iain Ireland <iireland@mozilla.com> - Mon, 20 Apr 2020 13:01:40 +0000 - rev 524852
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1630383: Shim simplifications r=mgaudet A couple of cleanups that have been floating around my patch stack for a while. Depends on D71353 Differential Revision: https://phabricator.services.mozilla.com/D71354
4c0cbbdaa2400f96ba7c154692f464d29e90cc80: Bug 1630383: Fix off-thread parsing r=mgaudet
Iain Ireland <iireland@mozilla.com> - Mon, 20 Apr 2020 13:01:22 +0000 - rev 524851
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1630383: Fix off-thread parsing r=mgaudet I had this right in my original patch stack, but 'fixed' it before putting the patches up for review. The old engine only needs to initialize the regexp stack for the main thread, because it's only used when executing regexps, but we need the isolate for off-thread parsing, so it needs to be initialized even for helper threads. Depends on D71352 Differential Revision: https://phabricator.services.mozilla.com/D71353
fe9af68999c4dcaaa130a9da1b6e4aa14641c40a: Bug 1630383: Fix broken builds r=mgaudet
Iain Ireland <iireland@mozilla.com> - Mon, 20 Apr 2020 13:00:07 +0000 - rev 524850
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1630383: Fix broken builds r=mgaudet While getting my patches organized for review and addressing comments, I managed to accidentally break a few build configurations on try. 1. I wasn't building binast locally, so it was just broken. 2. I missed an underscore in ARM-specific masm code. 3. The new regexp-error.h file that I imported needs to #include regexp-shim.h to avoid breaking non-unified builds. Differential Revision: https://phabricator.services.mozilla.com/D71352
6f9b5bafc63fe3fd0d5be1d6ac75a3154c963f7e: Bug 1379232 - Assert that ChunkPools are empty in their destructor r=jandem
Jon Coppeard <jcoppeard@mozilla.com> - Mon, 20 Apr 2020 11:59:06 +0000 - rev 524849
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1379232 - Assert that ChunkPools are empty in their destructor r=jandem This appears to pass on try now, probably due to the memory leak reporting changes around bug 1407593. Differential Revision: https://phabricator.services.mozilla.com/D71508
bdb63d73a58d7636c3be6bfeec4b08ccc7875898: Bug 1605320 - Test that Use in console context menu entry selects the right context. r=jlast.
Nicolas Chevobbe <nchevobbe@mozilla.com> - Mon, 20 Apr 2020 16:21:37 +0000 - rev 524848
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1605320 - Test that Use in console context menu entry selects the right context. r=jlast. Differential Revision: https://phabricator.services.mozilla.com/D71551
0cc9ef01b1cfec865df32edc76a779a851445064: Bug 1596812 Part 7 - Add documentation for the stub installer web UI. r=agashlin,nalexander
Molly Howell <mhowell@mozilla.com> - Mon, 20 Apr 2020 03:06:08 +0000 - rev 524847
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1596812 Part 7 - Add documentation for the stub installer web UI. r=agashlin,nalexander Depends on D56582 Differential Revision: https://phabricator.services.mozilla.com/D56583
c75138f9f26780eb33581f9b13588c55c9225da5: Bug 1596812 Part 6 - Replace the stub installer UI code with calls to the web plugin. r=agashlin
Molly Howell <mhowell@mozilla.com> - Mon, 20 Apr 2020 03:06:08 +0000 - rev 524846
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1596812 Part 6 - Replace the stub installer UI code with calls to the web plugin. r=agashlin Depends on D56581 Differential Revision: https://phabricator.services.mozilla.com/D56582
7dd034d966f21a33b636fa3c50396902fa701918: Bug 1596812 Part 5 - Add the web content files and include them in the installer build. r=agashlin,nalexander,mconley
Molly Howell <mhowell@mozilla.com> - Mon, 20 Apr 2020 03:06:08 +0000 - rev 524845
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1596812 Part 5 - Add the web content files and include them in the installer build. r=agashlin,nalexander,mconley Depends on D56580 Differential Revision: https://phabricator.services.mozilla.com/D56581
c6debaebcb1fac5852e03efc873635e6048e7dd0: Bug 1596812 Part 4 - Add the WebBrowser plugin to the installer build files. r=agashlin,nalexander
Molly Howell <mhowell@mozilla.com> - Mon, 20 Apr 2020 03:06:07 +0000 - rev 524844
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1596812 Part 4 - Add the WebBrowser plugin to the installer build files. r=agashlin,nalexander Differential Revision: https://phabricator.services.mozilla.com/D56580
860657b3ae14bb727bb8efa7115d9981b53bbaba: Bug 1596812 Part 3 - Compiled binary for the WebBrowser plugin. r=agashlin
Molly Howell <mhowell@mozilla.com> - Mon, 20 Apr 2020 03:06:07 +0000 - rev 524843
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1596812 Part 3 - Compiled binary for the WebBrowser plugin. r=agashlin Depends on D56577 Differential Revision: https://phabricator.services.mozilla.com/D56579
1c80f6c6d43bc4e39fe11150ae2a92ac0bd45a1d: Bug 1596812 Part 2 - NSIS WebBrowser plugin. r=agashlin,nalexander
Molly Howell <mhowell@mozilla.com> - Mon, 20 Apr 2020 03:06:07 +0000 - rev 524842
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1596812 Part 2 - NSIS WebBrowser plugin. r=agashlin,nalexander This is all the code and build files for an NSIS plugin that enables rendering a web page as the content of an NSIS dialog. Documentation and the compiled binary are in later commits in this series. Depends on D56576 Differential Revision: https://phabricator.services.mozilla.com/D56577
32eba50a6ae13027827ecb119cbf733de63718bb: Bug 1596812 Part 1 - Update our custom nsisui.exe. r=agashlin
Molly Howell <mhowell@mozilla.com> - Mon, 20 Apr 2020 03:06:07 +0000 - rev 524841
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1596812 Part 1 - Update our custom nsisui.exe. r=agashlin Minify this file by removing the dialogs we don't need and hiding all the unnecessary controls in the one we do need, so the stub installer code doesn't have to do that manually (I would have removed those controls altogether, but the NSIS compiler errors out if you do that). The dialog resources were edited directly in-place using Resource Hacker <http://angusj.com/resourcehacker/>. Differential Revision: https://phabricator.services.mozilla.com/D56576
82284d55b34d9dcb238f05400497f4abbbad1698: Bug 1631139 - GuardToClass folding for ArrayObject. r=jandem
Tom Schuster <evilpies@gmail.com> - Mon, 20 Apr 2020 14:13:28 +0000 - rev 524840
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1631139 - GuardToClass folding for ArrayObject. r=jandem Differential Revision: https://phabricator.services.mozilla.com/D71425
cb0840a3587a92ba74003af7d841d8fca7f0401d: Bug 1624391 - Add capability to parse update XMLs to the update agent r=agashlin,mhowell
Kirk Steuber <ksteuber@mozilla.com> - Fri, 17 Apr 2020 20:48:10 +0000 - rev 524839
Push 37332 by ncsoregi@mozilla.com at Tue, 21 Apr 2020 02:08:37 +0000
Bug 1624391 - Add capability to parse update XMLs to the update agent r=agashlin,mhowell Differential Revision: https://phabricator.services.mozilla.com/D69064
(0) -300000 -100000 -30000 -10000 -3000 -1000 -300 -100 -50 -20 +20 +50 +100 +300 +1000 +3000 +10000 +30000 tip