49976 2010-11-23 08:14:27 -0800 [Qt] Fix crashes in debug mode 2010-12-10 02:20:36 -0800 1 1 1 Unclassified WebKit JavaScriptCore 528+ (Nightly build) All All RESOLVED FIXED Qt, QtTriaged P2 Normal --- 1 ossy webkit-unassigned barraclough darin loki oliver robert webkit.review.bot zherczeg oldest_to_newest 312854 0 74668 ossy 2010-11-23 08:14:27 -0800 Created attachment 74668 GDB backtrace On trunk 8 tests crash in debug mode: (r72592) dom/html/level1/core/hc_attrappendchild3.html dom/xhtml/level2/html/table07.xhtml dom/xhtml/level2/html/table22.xhtml fast/js/sputnik/Conformance/15_Native_Objects/15.3_Function/15.3.3/S15.3.3_A2_T2.html fast/js/sputnik/Conformance/15_Native_Objects/15.4_Array/15.4.4/15.4.4.9_Array_prototype_shift/S15.4.4.9_A1.1_T1.html fast/js/sputnik/Conformance/15_Native_Objects/15.7_Number/15.7.3/15.7.3.5_Number.NEGATIVE_INFINITY/S15.7.3.5_A1.html fast/js/sputnik/Conformance/15_Native_Objects/15.10_RegExp/15.10.2/15.10.2.12_CharacterClassEscape/S15.10.2.12_A6_T2.html ietestcenter/Javascript/15.4.4.15-5-8.html These tests don't crash if we run only one test, but we can reproduce the crashes. for example I tried with the following tests: dom/xhtml/level2/html/object11.xhtml dom/xhtml/level2/html/object12.xhtml dom/xhtml/level2/html/object13.xhtml dom/xhtml/level2/html/object14.xhtml dom/xhtml/level2/html/object15.xhtml dom/xhtml/level2/html/table01.xhtml dom/xhtml/level2/html/table02.xhtml dom/xhtml/level2/html/table03.xhtml dom/xhtml/level2/html/table04.xhtml dom/xhtml/level2/html/table06.xhtml dom/xhtml/level2/html/table07.xhtml GDB backtrace is attached. 312855 1 ossy 2010-11-23 08:15:42 -0800 I tried it with interpreter and JIT too, and I got same results, so I think it is a general JavaScriptCore bug. 312858 2 ossy 2010-11-23 08:17:27 -0800 Have you got any idea what caused these crashes and how can we fix it? 313109 3 barraclough 2010-11-23 18:46:26 -0800 If I'm reading the backtrace correctly, it looks like this assert is failing: > result == CallTypeNone || value.isValidCallee() Which means value is not a valid callee. We now require all callable objects in JSC (i.e. host functions) to be derived from JSObjectWithGlobalObject, which is what isValidCallee() tests. Objects created via the JSC API should all derive from JSObjectWithGlobalObject (well, bar global objects created via the API, but they'll still be okay – isValidCallee() actually checks the contents of anonymous slot 0, GOs will pass this check too). So my guess would be something like, that the Qt DRT is exposing some host functions to JS by directly deriving from a JSObject type that is not a JSObjectWithGlobalObject, and directly implementing the getCallData() callback itself. If so, a fix would be to use the NativeFunctionWrapper type when wrapping C functions. My next step in debugging this would be to see what function was being called when this ASSERT fired. Hope this helps. G. 313416 4 robert 2010-11-24 11:57:17 -0800 This is due to a problem with m_cacheableBindingRootObject in ScriptController. The call that is causing the crash is on LayoutTestController which is a runtime object and is a JSObjectWithGlobalObject. Darin predicted this in https://bugs.webkit.org/show_bug.cgi?id=48758#c13 unfortunately. This particular crash seems to be due to a stale reference to the globalObject in the runtime object. This happens because m_cacheableBindingRootObject persists between page loads. The globalObject associated with the root object is updated between page loads since bug 48758. However there is still a reference to the old global object in the runtime object associated with the root object's/JSC object's instance. So it looks like that needs to be updated as well. I'm not sure why this is only a problem on debug builds. 314244 5 oliver 2010-11-28 19:11:27 -0800 > I'm not sure why this is only a problem on debug builds. I assume it's an assertion (eg. debug only), and so more-or-less deterministic in debug builds, whereas in release builds i suspect it works through luck. 315032 6 zherczeg 2010-11-30 07:25:38 -0800 This is a Garbage Collector issue. The m_structure of QtRuntimeMetaMethod is also a JSValue (JSDOMWindow). This object is freed by GC during the execution of table07.xhtml. QtRuntimeMetaMethod is an InternalFunction. How an InternalFunction (JSObject descendant) should mark its asObject(this)->getAnonymousValue(0) ( == propertyStorage()[0] )? Would be good to know how QtRuntimeMetaMethod itself is marked. 315249 7 robert 2010-11-30 12:46:04 -0800 Hi Zoltan, Thanks for looking at this. (In reply to comment #6) > This is a Garbage Collector issue. The m_structure of QtRuntimeMetaMethod is also a JSValue (JSDOMWindow). This object is freed by GC during the execution of table07.xhtml. How can you tell? I'd like to know how I could have spotted this! >QtRuntimeMetaMethod is an InternalFunction. How an InternalFunction (JSObject descendant) should mark its asObject(this)->getAnonymousValue(0) ( == propertyStorage()[0] )? Would be good to know how QtRuntimeMetaMethod itself is marked. Would a RefPtr protect(value) not do it? If so, why not? 315309 8 oliver 2010-11-30 14:20:04 -0800 (In reply to comment #6) > This is a Garbage Collector issue. The m_structure of QtRuntimeMetaMethod is also a JSValue (JSDOMWindow). This object is freed by GC during the execution of table07.xhtml. QtRuntimeMetaMethod is an InternalFunction. How an InternalFunction (JSObject descendant) should mark its asObject(this)->getAnonymousValue(0) ( == propertyStorage()[0] )? Would be good to know how QtRuntimeMetaMethod itself is marked. Structures aren't GC allocated, i assume you mean the global object reference? 317466 9 zherczeg 2010-12-05 23:44:06 -0800 > Structures aren't GC allocated, i assume you mean the global object reference? Yeah, I was not precise: static PassRefPtr<Structure> createStructure(JSValue proto). { return Structure::create(proto, TypeInfo(ObjectType, StructureFlags), AnonymousSlotCount);. } The "proto" object is freed. 317467 10 zherczeg 2010-12-05 23:51:10 -0800 > How can you tell? I'd like to know how I could have spotted this! Requires a lot of JSC internal knowledge. First, you should "feel" this is a GC issue, and know which object will be freed (but still referenced). Then you need to put a breakpoint where this object must be exists (i.e: its constructor), and execute a "watch *(int*)this" inside GDB. This put a write watchpoint to the vptr (virtual ptr) of an object, which never changes, except if the memory is overwritten. In case of GC, it overwrites this are during a sweep, since it maintains a chan list for free cells. And GDB will stops when this particular object is sweeped by GC. > Would a RefPtr protect(value) not do it? If so, why not? No, cells live an entiriely different life cycle than anny other objects in WebKit. The GC should tell which cells are referenced, and which are not. 317479 11 zherczeg 2010-12-06 00:48:29 -0800 (In reply to comment #9) > > Structures aren't GC allocated, i assume you mean the global object reference? (In reply to comment #9) > > Structures aren't GC allocated, i assume you mean the global object reference? > > Yeah, I was not precise: > > static PassRefPtr<Structure> createStructure(JSValue proto). > { > return Structure::create(proto, TypeInfo(ObjectType, StructureFlags), AnonymousSlotCount);. > } > > The "proto" object is freed. Ah hell, no. So you were right: JSObjectWithGlobalObject::JSObjectWithGlobalObject (base class of InternalFunction) putAnonymousValue(GlobalObjectSlot, globalObject); Hm, shall this class should mark its "globalObject" ? 317626 12 oliver 2010-12-06 07:58:51 -0800 (In reply to comment #11) > (In reply to comment #9) > > > Structures aren't GC allocated, i assume you mean the global object reference? > > (In reply to comment #9) > > > Structures aren't GC allocated, i assume you mean the global object reference? > > > > Yeah, I was not precise: > > > > static PassRefPtr<Structure> createStructure(JSValue proto). > > { > > return Structure::create(proto, TypeInfo(ObjectType, StructureFlags), AnonymousSlotCount);. > > } > > > > The "proto" object is freed. > > Ah hell, no. So you were right: > > JSObjectWithGlobalObject::JSObjectWithGlobalObject (base class of InternalFunction) > putAnonymousValue(GlobalObjectSlot, globalObject); > > Hm, shall this class should mark its "globalObject" ? Its global object should be marked through the base JSObject::markChildren method (which marks all of the properties slots, including anon. storage) 317746 13 robert 2010-12-06 11:31:57 -0800 Needs the patch at https://bugs.webkit.org/show_bug.cgi?id=50091 in order to recreate the problem 317771 14 zherczeg 2010-12-06 12:05:30 -0800 > Its global object should be marked through the base JSObject::markChildren method (which marks all of the properties slots, including anon. storage) Looks reasonable. However, the InernalFunction object exits after GC (its vptr is still valid, thus it is not GC'ed) while the JSDOMWindow is invalid. How this can be happen? Seems black magic. 318122 15 zherczeg 2010-12-07 04:08:36 -0800 Hopefully I know what is happening know. A new "feature" was introduced to the GarbageCollector (new at least for me), which defers the destruction of Cells until they reused (thus, immediate sweep is unneeded, and distributing the sweep time during the actual allocation). Thus, "dead" objects can still seems living, and that was confused me. All InternalFunctions allocated by QtClass are sweeped out, and the name/method hashmap cache becomes invalid. Two possible solutions: 1) Something should somehow mark these objects. Is there an easy way to do this? 2) Drop this cache (this would follow the cacheless ObjectiveC implemetation) Since these objects are probably only used by DumpRenderTree, I would vote to 2, but would be interested in your opinion. 318169 16 75810 zherczeg 2010-12-07 06:58:09 -0800 Created attachment 75810 fix Really tired of this bug... 318228 17 webkit.review.bot 2010-12-07 09:04:26 -0800 Attachment 75810 did not pass style-queue: Failed to run "[u'git', u'reset', u'--hard', u'refs/remotes/trunk']" exit_code: 128 error: Could not write new index file. fatal: Could not reset index file to revision 'refs/remotes/trunk'. If any of these errors are false positives, please file a bug against check-webkit-style. 318280 18 webkit.review.bot 2010-12-07 10:05:10 -0800 Attachment 75810 did not pass style-queue: Failed to run "[u'git', u'reset', u'--hard', u'refs/remotes/trunk']" exit_code: 128 error: Could not write new index file. fatal: Could not reset index file to revision 'refs/remotes/trunk'. If any of these errors are false positives, please file a bug against check-webkit-style. 318326 19 robert 2010-12-07 11:05:27 -0800 Does this allow the tests unskipped in https://bugs.webkit.org/show_bug.cgi?id=37725 to still pass? 318327 20 webkit.review.bot 2010-12-07 11:06:05 -0800 Attachment 75810 did not pass style-queue: Failed to run "[u'git', u'reset', u'--hard', u'refs/remotes/trunk']" exit_code: 128 error: Could not write new index file. fatal: Could not reset index file to revision 'refs/remotes/trunk'. If any of these errors are false positives, please file a bug against check-webkit-style. 318380 21 webkit.review.bot 2010-12-07 12:07:31 -0800 Attachment 75810 did not pass style-queue: Failed to run "[u'git', u'reset', u'--hard', u'refs/remotes/trunk']" exit_code: 128 error: Could not write new index file. fatal: Could not reset index file to revision 'refs/remotes/trunk'. If any of these errors are false positives, please file a bug against check-webkit-style. 318602 22 webkit.review.bot 2010-12-07 21:38:02 -0800 Attachment 75810 did not pass style-queue: Failed to run "['WebKitTools/Scripts/update-webkit']" exit_code: 2 Updating OpenSource Incomplete data: Delta source ended unexpectedly at /usr/lib/git-core/git-svn line 5061 Died at WebKitTools/Scripts/update-webkit line 132. If any of these errors are false positives, please file a bug against check-webkit-style. 319992 23 75810 eric 2010-12-10 00:08:07 -0800 Comment on attachment 75810 fix OK. 320038 24 75810 webkit.review.bot 2010-12-10 02:20:29 -0800 Comment on attachment 75810 fix Clearing flags on attachment: 75810 Committed r73699: <http://trac.webkit.org/changeset/73699> 320039 25 webkit.review.bot 2010-12-10 02:20:36 -0800 All reviewed patches have been landed. Closing bug. 74668 2010-11-23 08:14:27 -0800 2010-11-23 08:14:27 -0800 GDB backtrace bt.txt text/plain 7489 ossy JCBnZGIgV2ViS2l0QnVpbGQvRGVidWcvYmluL0R1bXBSZW5kZXJUcmVlDQoNCmRvbS94aHRtbC9s ZXZlbDIvaHRtbC9vYmplY3QxMS54aHRtbA0KZG9tL3hodG1sL2xldmVsMi9odG1sL29iamVjdDEy LnhodG1sDQpkb20veGh0bWwvbGV2ZWwyL2h0bWwvb2JqZWN0MTMueGh0bWwNCmRvbS94aHRtbC9s ZXZlbDIvaHRtbC9vYmplY3QxNC54aHRtbA0KZG9tL3hodG1sL2xldmVsMi9odG1sL29iamVjdDE1 LnhodG1sDQpkb20veGh0bWwvbGV2ZWwyL2h0bWwvdGFibGUwMS54aHRtbA0KZG9tL3hodG1sL2xl dmVsMi9odG1sL3RhYmxlMDIueGh0bWwNCmRvbS94aHRtbC9sZXZlbDIvaHRtbC90YWJsZTAzLnho dG1sDQpkb20veGh0bWwvbGV2ZWwyL2h0bWwvdGFibGUwNC54aHRtbA0KZG9tL3hodG1sL2xldmVs Mi9odG1sL3RhYmxlMDYueGh0bWwNCmRvbS94aHRtbC9sZXZlbDIvaHRtbC90YWJsZTA3LnhodG1s DQoNCkFTU0VSVElPTiBGQUlMRUQ6IHJlc3VsdCA9PSBDYWxsVHlwZU5vbmUgfHwgdmFsdWUuaXNW YWxpZENhbGxlZSgpDQooLi4vLi4vLi4vSmF2YVNjcmlwdENvcmUvcnVudGltZS9KU0NlbGwuaDoy MjAgSlNDOjpDYWxsVHlwZSBKU0M6OmdldENhbGxEYXRhKEpTQzo6SlNWYWx1ZSwgSlNDOjpDYWxs RGF0YSYpKQ0KDQpQcm9ncmFtIHJlY2VpdmVkIHNpZ25hbCBTSUdTRUdWLCBTZWdtZW50YXRpb24g ZmF1bHQuDQpbU3dpdGNoaW5nIHRvIFRocmVhZCAweGYyZmM0NmMwIChMV1AgMTQwMTEpXQ0KMHhm NWQ5OWQ4MyBpbiBKU0M6OmdldENhbGxEYXRhICh2YWx1ZT17dSA9IHthc0VuY29kZWRKU1ZhbHVl ID0gLTEzMTM3MzI3ODA4LCBhc0RvdWJsZSA9IC1uYW4oMHhmZmZmY2YwZjQ0OTQwKSwgYXNCaXRz ID0ge3BheWxvYWQgPSAtMjUyNDI1OTIwLCB0YWcgPSAtNH19fSwgY2FsbERhdGE9QDB4ZmY5MDgy ZDQpIGF0IC4uLy4uLy4uL0phdmFTY3JpcHRDb3JlL3J1bnRpbWUvSlNDZWxsLmg6MjIwDQoyMjAg ICAgICAgICAgICAgQVNTRVJUKHJlc3VsdCA9PSBDYWxsVHlwZU5vbmUgfHwgdmFsdWUuaXNWYWxp ZENhbGxlZSgpKTsNCihnZGIpDQooZ2RiKSBidA0KIzAgIDB4ZjVkOTlkODMgaW4gSlNDOjpnZXRD YWxsRGF0YSAodmFsdWU9e3UgPSB7YXNFbmNvZGVkSlNWYWx1ZSA9IC0xMzEzNzMyNzgwOCwgYXNE b3VibGUgPSAtbmFuKDB4ZmZmZmNmMGY0NDk0MCksIGFzQml0cyA9IHtwYXlsb2FkID0gLTI1MjQy NTkyMCwgdGFnID0gLTR9fX0sIGNhbGxEYXRhPUAweGZmOTA4MmQ0KSBhdCAuLi8uLi8uLi9KYXZh U2NyaXB0Q29yZS9ydW50aW1lL0pTQ2VsbC5oOjIyMA0KIzEgIDB4ZjY5ZDliMDAgaW4gY3RpX29w X2NhbGxfTm90SlNGdW5jdGlvbiAoYXJncz0weGZmOTA4MzMwKSBhdCAuLi8uLi8uLi9KYXZhU2Ny aXB0Q29yZS9qaXQvSklUU3R1YnMuY3BwOjIxODkNCiMyICAweGY2OWNlZmRmIGluIGRvdWJsZUhh c2ggKGtleT00MTUxNTU2NTYwKSBhdCAuLi8uLi8uLi9KYXZhU2NyaXB0Q29yZS93dGYvSGFzaFRh YmxlLmg6NDQ3DQojMyAgMHhmNjljYmYxNCBpbiBKU0M6OkpJVENvZGU6OmV4ZWN1dGUgKHRoaXM9 MHhhMWRhN2RjLCByZWdpc3RlckZpbGU9MHg5ZmM3YTNjLCBjYWxsRnJhbWU9MHhmMGZiNjA0MCwg Z2xvYmFsRGF0YT0weDlmYWMwMjApIGF0IC4uLy4uLy4uL0phdmFTY3JpcHRDb3JlL2ppdC9KSVRD b2RlLmg6NzcNCiM0ICAweGY2OWM0NmJjIGluIEpTQzo6SW50ZXJwcmV0ZXI6OmV4ZWN1dGVDYWxs ICh0aGlzPTB4OWZjN2EzMCwgY2FsbEZyYW1lPTB4YTFlNTllYywgZnVuY3Rpb249MHhmMGY0MWQ4 MCwgY2FsbFR5cGU9SlNDOjpDYWxsVHlwZUpTLCBjYWxsRGF0YT1AMHhmZjkwODYwNCwgdGhpc1Zh bHVlPQ0KICAgICAgICAgIHt1ID0ge2FzRW5jb2RlZEpTVmFsdWUgPSAtMTMxMzczNDY1NjAsIGFz RG91YmxlID0gLW5hbigweGZmZmZjZjBmNDAwMDApLCBhc0JpdHMgPSB7cGF5bG9hZCA9IC0yNTI0 NDQ2NzIsIHRhZyA9IC00fX19LCBhcmdzPUAweGZmOTA4NjRjKSBhdCAuLi8uLi8uLi9KYXZhU2Ny aXB0Q29yZS9pbnRlcnByZXRlci9JbnRlcnByZXRlci5jcHA6ODQ5DQojNSAgMHhmNjlmMTM5MSBp biBKU0M6OmNhbGwgKGV4ZWM9MHhhMWU1OWVjLCBmdW5jdGlvbk9iamVjdD17dSA9IHthc0VuY29k ZWRKU1ZhbHVlID0gLTEzMTM3MzM5MDA4LCBhc0RvdWJsZSA9IC1uYW4oMHhmZmZmY2YwZjQxZDgw KSwgYXNCaXRzID0ge3BheWxvYWQgPSAtMjUyNDM3MTIwLCB0YWcgPSAtNH19fSwgY2FsbFR5cGU9 SlNDOjpDYWxsVHlwZUpTLCBjYWxsRGF0YT1AMHhmZjkwODYwNCwgdGhpc1ZhbHVlPQ0KICAgICAg ICAgIHt1ID0ge2FzRW5jb2RlZEpTVmFsdWUgPSAtMTMxMzczNDY1NjAsIGFzRG91YmxlID0gLW5h bigweGZmZmZjZjBmNDAwMDApLCBhc0JpdHMgPSB7cGF5bG9hZCA9IC0yNTI0NDQ2NzIsIHRhZyA9 IC00fX19LCBhcmdzPUAweGZmOTA4NjRjKSBhdCAuLi8uLi8uLi9KYXZhU2NyaXB0Q29yZS9ydW50 aW1lL0NhbGxEYXRhLmNwcDozOA0KIzYgIDB4ZjVkOTlmOTQgaW4gV2ViQ29yZTo6SlNNYWluVGhy ZWFkRXhlY1N0YXRlOjpjYWxsIChleGVjPTB4YTFlNTllYywgZnVuY3Rpb25PYmplY3Q9e3UgPSB7 YXNFbmNvZGVkSlNWYWx1ZSA9IC0xMzEzNzMzOTAwOCwgYXNEb3VibGUgPSAtbmFuKDB4ZmZmZmNm MGY0MWQ4MCksIGFzQml0cyA9IHtwYXlsb2FkID0gLTI1MjQzNzEyMCwgdGFnID0gLTR9fX0sIGNh bGxUeXBlPUpTQzo6Q2FsbFR5cGVKUywNCiAgICBjYWxsRGF0YT1AMHhmZjkwODYwNCwgdGhpc1Zh bHVlPXt1ID0ge2FzRW5jb2RlZEpTVmFsdWUgPSAtMTMxMzczNDY1NjAsIGFzRG91YmxlID0gLW5h bigweGZmZmZjZjBmNDAwMDApLCBhc0JpdHMgPSB7cGF5bG9hZCA9IC0yNTI0NDQ2NzIsIHRhZyA9 IC00fX19LCBhcmdzPUAweGZmOTA4NjRjKSBhdCAuLi8uLi8uLi9XZWJDb3JlL2JpbmRpbmdzL2pz L0pTTWFpblRocmVhZEV4ZWNTdGF0ZS5oOjQ4DQojNyAgMHhmNWRkMDU2YyBpbiBXZWJDb3JlOjpK U0V2ZW50TGlzdGVuZXI6OmhhbmRsZUV2ZW50ICh0aGlzPTB4YTIwOTk3MCwgc2NyaXB0RXhlY3V0 aW9uQ29udGV4dD0weGExZDYwYzAsIGV2ZW50PTB4OWZhNWUxOCkgYXQgLi4vLi4vLi4vV2ViQ29y ZS9iaW5kaW5ncy9qcy9KU0V2ZW50TGlzdGVuZXIuY3BwOjEyNA0KIzggIDB4ZjVmZDkyODkgaW4g V2ViQ29yZTo6RXZlbnRUYXJnZXQ6OmZpcmVFdmVudExpc3RlbmVycyAodGhpcz0weGExZjlmOTgs IGV2ZW50PTB4OWZhNWUxOCwgZD0weGExZmEwMjQsIGVudHJ5PUAweGExZWMxZDgpIGF0IC4uLy4u Ly4uL1dlYkNvcmUvZG9tL0V2ZW50VGFyZ2V0LmNwcDozNDINCiM5ICAweGY1ZmRhYTljIGluIFdl YkNvcmU6OkV2ZW50VGFyZ2V0OjpmaXJlRXZlbnRMaXN0ZW5lcnMgKHRoaXM9MHhhMWY5Zjk4LCBl dmVudD0weDlmYTVlMTgpIGF0IC4uLy4uLy4uL1dlYkNvcmUvZG9tL0V2ZW50VGFyZ2V0LmNwcDoz MTENCiMxMCAweGY2MzZhZmY1IGluIFdlYkNvcmU6OkRPTVdpbmRvdzo6ZGlzcGF0Y2hFdmVudCAo dGhpcz0weGExZjlmOTgsIHBycEV2ZW50PXttX3B0ciA9IDB4ZmY5MDg3ZTh9LCBwcnBUYXJnZXQ9 e21fcHRyID0gMHhmZjkwODdlNH0pIGF0IC4uLy4uLy4uL1dlYkNvcmUvcGFnZS9ET01XaW5kb3cu Y3BwOjE1NDcNCiMxMSAweGY2MzZiMTU1IGluIFdlYkNvcmU6OkRPTVdpbmRvdzo6ZGlzcGF0Y2hU aW1lZEV2ZW50ICh0aGlzPTB4YTFmOWY5OCwgZXZlbnQ9e21fcHRyID0gMHhmZjkwODg1NH0sIHRh cmdldD0weGExZDYwODgsIHN0YXJ0VGltZT0weGEyMDNhOWMsIGVuZFRpbWU9MHhhMjAzYWE0KSBh dCAuLi8uLi8uLi9XZWJDb3JlL3BhZ2UvRE9NV2luZG93LmNwcDoxNTU5DQojMTIgMHhmNjM2Yzgz YyBpbiBXZWJDb3JlOjpET01XaW5kb3c6OmRpc3BhdGNoTG9hZEV2ZW50ICh0aGlzPTB4YTFmOWY5 OCkgYXQgLi4vLi4vLi4vV2ViQ29yZS9wYWdlL0RPTVdpbmRvdy5jcHA6MTUxMw0KIzEzIDB4ZjVm OWY4OTYgaW4gV2ViQ29yZTo6RG9jdW1lbnQ6OmRpc3BhdGNoV2luZG93TG9hZEV2ZW50ICh0aGlz PTB4YTFkNjA4OCkgYXQgLi4vLi4vLi4vV2ViQ29yZS9kb20vRG9jdW1lbnQuY3BwOjM0NTcNCiMx NCAweGY1ZmEzMGFhIGluIFdlYkNvcmU6OkRvY3VtZW50OjppbXBsaWNpdENsb3NlICh0aGlzPTB4 YTFkNjA4OCkgYXQgLi4vLi4vLi4vV2ViQ29yZS9kb20vRG9jdW1lbnQuY3BwOjIxMDYNCiMxNSAw eGY2MmZhOGNiIGluIFdlYkNvcmU6OkZyYW1lTG9hZGVyOjpjaGVja0NhbGxJbXBsaWNpdENsb3Nl ICh0aGlzPTB4OWZiMTUzNCkgYXQgLi4vLi4vLi4vV2ViQ29yZS9sb2FkZXIvRnJhbWVMb2FkZXIu Y3BwOjkwMA0KIzE2IDB4ZjYyZmY1MGIgaW4gV2ViQ29yZTo6RnJhbWVMb2FkZXI6OmNoZWNrQ29t cGxldGVkICh0aGlzPTB4OWZiMTUzNCkgYXQgLi4vLi4vLi4vV2ViQ29yZS9sb2FkZXIvRnJhbWVM b2FkZXIuY3BwOjg0OA0KIzE3IDB4ZjYyZmY2YzcgaW4gV2ViQ29yZTo6RnJhbWVMb2FkZXI6Omxv YWREb25lICh0aGlzPTB4OWZiMTUzNCkgYXQgLi4vLi4vLi4vV2ViQ29yZS9sb2FkZXIvRnJhbWVM b2FkZXIuY3BwOjc5NQ0KIzE4IDB4ZjYyZDgxZmIgaW4gV2ViQ29yZTo6Q2FjaGVkUmVzb3VyY2VM b2FkZXI6OnNldExvYWRJblByb2dyZXNzICh0aGlzPTB4YTFkZjIzOCwgbG9hZD1mYWxzZSkgYXQg Li4vLi4vLi4vV2ViQ29yZS9sb2FkZXIvY2FjaGUvQ2FjaGVkUmVzb3VyY2VMb2FkZXIuY3BwOjM0 MQ0KIzE5IDB4ZjYzMTgzY2QgaW4gV2ViQ29yZTo6TG9hZGVyOjpkaWRGaW5pc2hMb2FkaW5nICh0 aGlzPTB4OWZjN2UzMCwgbG9hZGVyPTB4YTIxYTFlMCkgYXQgLi4vLi4vLi4vV2ViQ29yZS9sb2Fk ZXIvbG9hZGVyLmNwcDoyMDYNCiMyMCAweGY2MzM0YjUyIGluIFdlYkNvcmU6OlN1YnJlc291cmNl TG9hZGVyOjpkaWRGaW5pc2hMb2FkaW5nICh0aGlzPTB4YTIxYTFlMCwgZmluaXNoVGltZT0wKSBh dCAuLi8uLi8uLi9XZWJDb3JlL2xvYWRlci9TdWJyZXNvdXJjZUxvYWRlci5jcHA6MTgwDQojMjEg MHhmNjMyODM3ZiBpbiBXZWJDb3JlOjpSZXNvdXJjZUxvYWRlcjo6ZGlkRmluaXNoTG9hZGluZyAo dGhpcz0weGEyMWExZTAsIGZpbmlzaFRpbWU9MCkgYXQgLi4vLi4vLi4vV2ViQ29yZS9sb2FkZXIv UmVzb3VyY2VMb2FkZXIuY3BwOjQzNQ0KIzIyIDB4ZjY2NzRhMjAgaW4gV2ViQ29yZTo6UU5ldHdv cmtSZXBseUhhbmRsZXI6OmZpbmlzaCAodGhpcz0weGExYWQwZTgpIGF0IC4uLy4uLy4uL1dlYkNv cmUvcGxhdGZvcm0vbmV0d29yay9xdC9RTmV0d29ya1JlcGx5SGFuZGxlci5jcHA6MzA5DQojMjMg MHhmNjY3NTMwMSBpbiBXZWJDb3JlOjpRTmV0d29ya1JlcGx5SGFuZGxlcjo6cXRfbWV0YWNhbGwg KHRoaXM9MHhhMWFkMGU4LCBfYz1RTWV0YU9iamVjdDo6SW52b2tlTWV0YU1ldGhvZCwgX2lkPTEs IF9hPTB4ZmY5MDhjNjgpIGF0IC4vbW9jX1FOZXR3b3JrUmVwbHlIYW5kbGVyLmNwcDo4NA0KIzI0 IDB4ZjNiY2E2MDUgaW4gUU1ldGFPYmplY3Q6Om1ldGFjYWxsIChvYmplY3Q9MHgwLCBjbD1RTWV0 YU9iamVjdDo6SW52b2tlTWV0YU1ldGhvZCwgaWR4PTUsIGFyZ3Y9MHhmZjkwOGM2OCkgYXQga2Vy bmVsL3FtZXRhb2JqZWN0LmNwcDoyMzcNCiMyNSAweGYzYmQ5NmQxIGluIFFNZXRhT2JqZWN0Ojph Y3RpdmF0ZSAoc2VuZGVyPTB4YTIwOTNkMCwgbT0weGY0OGQ5ZTZjLCBsb2NhbF9zaWduYWxfaW5k ZXg9MSwgYXJndj0weDApIGF0IGtlcm5lbC9xb2JqZWN0LmNwcDozMjg3DQojMjYgMHhmNDg4M2I3 NyBpbiBRTmV0d29ya1JlcGx5OjpmaW5pc2hlZCAodGhpcz0weGEyMDkzZDApIGF0IC5tb2MvcmVs ZWFzZS1zaGFyZWQvbW9jX3FuZXR3b3JrcmVwbHkuY3BwOjE1MA0KIzI3IDB4ZjQ4ODNjZTMgaW4g UU5ldHdvcmtSZXBseTo6cXRfbWV0YWNhbGwgKHRoaXM9MHhhMjA5M2QwLCBfYz1RTWV0YU9iamVj dDo6SW52b2tlTWV0YU1ldGhvZCwgX2lkPTEsIF9hPTB4YTIxOGZlMCkgYXQgLm1vYy9yZWxlYXNl LXNoYXJlZC9tb2NfcW5ldHdvcmtyZXBseS5jcHA6MTI4DQojMjggMHhmNDdmYmUwMiBpbiBRRmls ZU5ldHdvcmtSZXBseTo6cXRfbWV0YWNhbGwgKHRoaXM9MHhhMjA5M2QwLCBfYz1RTWV0YU9iamVj dDo6SW52b2tlTWV0YU1ldGhvZCwgX2lkPTksIF9hPTB4YTIxOGZlMCkgYXQgLm1vYy9yZWxlYXNl LXNoYXJlZC9tb2NfcWZpbGVuZXR3b3JrcmVwbHlfcC5jcHA6NjQNCiMyOSAweGYzYmNhNjA1IGlu IFFNZXRhT2JqZWN0OjptZXRhY2FsbCAob2JqZWN0PTB4MCwgY2w9UU1ldGFPYmplY3Q6Okludm9r ZU1ldGFNZXRob2QsIGlkeD05LCBhcmd2PTB4YTIxOGZlMCkgYXQga2VybmVsL3FtZXRhb2JqZWN0 LmNwcDoyMzcNCiMzMCAweGYzYmQ0YzU2IGluIFFNZXRhQ2FsbEV2ZW50OjpwbGFjZU1ldGFDYWxs ICh0aGlzPTB4YTFhZWVmMCwgb2JqZWN0PTB4YTIwOTNkMCkgYXQga2VybmVsL3FvYmplY3QuY3Bw OjU2MQ0KIzMxIDB4ZjNiZDYxZTMgaW4gUU9iamVjdDo6ZXZlbnQgKHRoaXM9MHhhMjA5M2QwLCBl PTB4YTFhZWVmMCkgYXQga2VybmVsL3FvYmplY3QuY3BwOjEyNDANCiMzMiAweGYzZTM3YTRjIGlu IFFBcHBsaWNhdGlvblByaXZhdGU6Om5vdGlmeV9oZWxwZXIgKHRoaXM9MHg5ZjRkMjIwLCByZWNl aXZlcj0weGEyMDkzZDAsIGU9MHhhMWFlZWYwKSBhdCBrZXJuZWwvcWFwcGxpY2F0aW9uLmNwcDo0 MzAyDQojMzMgMHhmM2UzZTRiMiBpbiBRQXBwbGljYXRpb246Om5vdGlmeSAodGhpcz0weGZmOTA5 NTU4LCByZWNlaXZlcj0weGEyMDkzZDAsIGU9MHhhMWFlZWYwKSBhdCBrZXJuZWwvcWFwcGxpY2F0 aW9uLmNwcDozNzA2DQojMzQgMHhmM2JjNTEzYiBpbiBRQ29yZUFwcGxpY2F0aW9uOjpub3RpZnlJ bnRlcm5hbCAodGhpcz0weGZmOTA5NTU4LCByZWNlaXZlcj0weGEyMDkzZDAsIGV2ZW50PTB4YTFh ZWVmMCkgYXQga2VybmVsL3Fjb3JlYXBwbGljYXRpb24uY3BwOjcyNg0KIzM1IDB4ZjNiYzdhNGMg aW4gUUNvcmVBcHBsaWNhdGlvblByaXZhdGU6OnNlbmRQb3N0ZWRFdmVudHMgKHJlY2VpdmVyPTB4 MCwgZXZlbnRfdHlwZT0wLCBkYXRhPTB4OWY0ZDUzMCkgYXQgLi4vLi4vaW5jbHVkZS9RdENvcmUv Li4vLi4vc3JjL2NvcmVsaWIva2VybmVsL3Fjb3JlYXBwbGljYXRpb24uaDoyMTUNCiMzNiAweGYz YmM3YmZkIGluIFFDb3JlQXBwbGljYXRpb246OnNlbmRQb3N0ZWRFdmVudHMgKHJlY2VpdmVyPTB4 MCwgZXZlbnRfdHlwZT0wKSBhdCBrZXJuZWwvcWNvcmVhcHBsaWNhdGlvbi5jcHA6MTI2MA0KIzM3 IDB4ZjNiZjFlZGYgaW4gcG9zdEV2ZW50U291cmNlRGlzcGF0Y2ggKHM9MHg5ZjRjMTgwKSBhdCAu Li8uLi9pbmNsdWRlL1F0Q29yZS8uLi8uLi9zcmMvY29yZWxpYi9rZXJuZWwvcWNvcmVhcHBsaWNh dGlvbi5oOjIyMA0KIzM4IDB4ZjMwNTAxZDggaW4gZ19tYWluX2NvbnRleHRfZGlzcGF0Y2ggKCkg ZnJvbSAvdXNyL2xpYi9saWJnbGliLTIuMC5zby4wDQojMzkgMHhmMzA1Mzg3MyBpbiA/PyAoKSBm cm9tIC91c3IvbGliL2xpYmdsaWItMi4wLnNvLjANCiM0MCAweDA5ZjRkMWEwIGluID8/ICgpDQoj NDEgMHgwMDAwMDAwMCBpbiA/PyAoKQ== 75810 2010-12-07 06:58:09 -0800 2010-12-10 02:20:29 -0800 fix 0001-debug-crash.patch text/plain 1633 zherczeg RnJvbSA2ODkwZmU5ZmQ5MjYxYTYxYjI3ODdmZmI2OGJlYTM5ZDI5ZTgyZTc3IE1vbiBTZXAgMTcg MDA6MDA6MDAgMjAwMQpGcm9tOiBab2x0YW4gSGVyY3plZyA8emhlcmN6ZWdAaW5mLnUtc3plZ2Vk Lmh1PgpEYXRlOiBUdWUsIDcgRGVjIDIwMTAgMTU6NTY6NDEgKzAxMDAKU3ViamVjdDogW1BBVENI XSBkZWJ1Zy1jcmFzaAoKU2lnbmVkLW9mZi1ieTogWm9sdGFuIEhlcmN6ZWcgPHpoZXJjemVnQGlu Zi51LXN6ZWdlZC5odT4KLS0tCiBXZWJDb3JlL0NoYW5nZUxvZyAgICAgICAgICAgICAgICAgfCAg IDE0ICsrKysrKysrKysrKysrCiBXZWJDb3JlL2JyaWRnZS9xdC9xdF9pbnN0YW5jZS5jcHAgfCAg ICAxICsKIDIgZmlsZXMgY2hhbmdlZCwgMTUgaW5zZXJ0aW9ucygrKSwgMCBkZWxldGlvbnMoLSkK CmRpZmYgLS1naXQgYS9XZWJDb3JlL0NoYW5nZUxvZyBiL1dlYkNvcmUvQ2hhbmdlTG9nCmluZGV4 IGQ0ODdmMjYuLjQwMDQwOWMgMTAwNjQ0Ci0tLSBhL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1dl YkNvcmUvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTcgQEAKKzIwMTAtMTItMDcgIFpvbHRhbiBIZXJj emVnICA8emhlcmN6ZWdAd2Via2l0Lm9yZz4KKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkg KE9PUFMhKS4KKworICAgICAgICBbUXRdIEZpeCBjcmFzaGVzIGluIGRlYnVnIG1vZGUKKyAgICAg ICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTQ5OTc2CisKKyAgICAg ICAgVGhlIG1fbWV0aG9kcyBoYXNobWFwIG9mIFF0SW5zdGFuY2UgY29udGFpbnMgSW50ZXJuYWxG dW5jdGlvbnMKKyAgICAgICAgd2hvc2UgZGVwZW5kIG9uIHRoZSBjdXJyZW50IFJ1bnRpbWVPYmpl Y3QuIFdoZW4gd2UgcmVjcmVhdGUgdGhlCisgICAgICAgIFJ1bnRpbWVPYmplY3QsIHdlIHNob3Vs ZCBhbHNvIHJlc2V0IHRoaXMgaGFzaG1hcC4KKworICAgICAgICAqIGJyaWRnZS9xdC9xdF9pbnN0 YW5jZS5jcHA6CisgICAgICAgIChKU0M6OkJpbmRpbmdzOjpRdEluc3RhbmNlOjpuZXdSdW50aW1l T2JqZWN0KToKKwogMjAxMC0xMi0wNiAgUGF2ZWwgRmVsZG1hbiAgPHBmZWxkbWFuQGNocm9taXVt Lm9yZz4KIAogICAgICAgICBOb3QgcmV2aWV3ZWQuIFJvbGxpbmcgb3V0IHI3Mjk4OCBmb3IgYnJl YWtpbmcgaWZyYW1lcy4KZGlmZiAtLWdpdCBhL1dlYkNvcmUvYnJpZGdlL3F0L3F0X2luc3RhbmNl LmNwcCBiL1dlYkNvcmUvYnJpZGdlL3F0L3F0X2luc3RhbmNlLmNwcAppbmRleCBjMmUyZGNlLi43 ODI2M2U5IDEwMDY0NAotLS0gYS9XZWJDb3JlL2JyaWRnZS9xdC9xdF9pbnN0YW5jZS5jcHAKKysr IGIvV2ViQ29yZS9icmlkZ2UvcXQvcXRfaW5zdGFuY2UuY3BwCkBAIC0xODEsNiArMTgxLDcgQEAg Q2xhc3MqIFF0SW5zdGFuY2U6OmdldENsYXNzKCkgY29uc3QKIFJ1bnRpbWVPYmplY3QqIFF0SW5z dGFuY2U6Om5ld1J1bnRpbWVPYmplY3QoRXhlY1N0YXRlKiBleGVjKQogewogICAgIEpTTG9jayBs b2NrKFNpbGVuY2VBc3NlcnRpb25zT25seSk7CisgICAgbV9tZXRob2RzLmNsZWFyKCk7CiAgICAg cmV0dXJuIG5ldyAoZXhlYykgUXRSdW50aW1lT2JqZWN0KGV4ZWMsIGV4ZWMtPmxleGljYWxHbG9i YWxPYmplY3QoKSwgdGhpcyk7CiB9CiAKLS0gCjEuNi4wLjQKCg==