{"thread":{"id":"26881","subject":"Reproducible crash in git merge","startedAt":"2011-03-27T16:54:32Z","lastAt":"2011-03-29T18:40:20Z","messageCount":5,"participants":["David Gilman","Jakub Narebski","Junio C Hamano"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"164390","messageId":"AANLkTimxHn_fwMKh9cbp9i5LkShUi=HK44nv2KJ7ENaQ@mail.gmail.com","threadId":"26881","inReplyTo":null,"subject":"Reproducible crash in git merge","fromName":"David Gilman","fromEmail":"davidgilman1@gmail.com","sentAt":"2011-03-27T16:54:32Z","receivedAt":"2011-03-27T16:54:32Z","isPatch":false,"sender":{"key":"davidgilman1@gmail.com","avatar":null},"body":"I can get git merge to crash during a git pull. I'm using git\n1.7.3.2. The repo (afaict) is fine and had worked previously.\nUnfortunately I can't share the repo itself because of its contents.\nHere's the crash log that MacOS X generated for the crash.\n\nProcess: git-merge [39769]\nPath: /opt/local/libexec/git-core/git-merge\nIdentifier: git-merge\nVersion: ??? (???)\nCode Type: X86 (Native)\nParent Process: git [39768]\n\nDate/Time: 2011-03-27 11:42:49.676 -0500\nOS Version: Mac OS X 10.5.8 (9L31a)\nReport Version: 6\nAnonymous UUID: 2A93C147-AB28-4ECD-B290-8F2DE59A5B4B\n\nException Type: EXC_BAD_ACCESS (SIGBUS)\nException Codes: 0x000000000000000a, 0x0000000001200000\nCrashed Thread: 0\n\nThread 0 Crashed:\n0 libSystem.B.dylib \t0xffff0e35 __memcpy + 1685\n(cpu_capabilities.h:246)\n1 libz.1.dylib \t0x0015ba3f inflate + 943\n2 git-merge \t0x000e4b57 git_inflate + 39\n3 git-merge \t0x000d0ead read_object + 445\n4 git-merge \t0x000d0fa8 read_sha1_file_repl + 40\n5 git-merge \t0x000973c7 write_entry + 183\n6 git-merge \t0x000979d6 checkout_entry + 582\n7 git-merge \t0x000e2b33 unpack_trees + 1923\n8 git-merge \t0x000422c0 checkout_fast_forward + 512\n9 git-merge \t0x00043588 cmd_merge + 2504\n10 git-merge \t0x00002a3d handle_internal_command + 317\n11 git-merge \t0x000030f3 main + 1251\n12 git-merge \t0x00002226 start + 54\n\nThread 0 crashed with X86 Thread State (32-bit):\n eax: 0xffff0e1d ebx: 0x0015b6a7 ecx: 0x0000001c edx: 0xffffdfc0\n edi: 0x0169ad80 esi: 0x0120200f ebp: 0xbfffc4f8 esp: 0xbfffc4f0\n ss: 0x0000001f efl: 0x00010286 eip: 0xffff0e35 cs: 0x00000017\n ds: 0x0000001f es: 0x0000001f fs: 0x00000000 gs: 0x00000037\n cr2: 0x01200000\n\nBinary Images:\n 0x1000 - 0x110fef +git-merge ??? (???)\n\n/opt/local/libexec/git-core/git-merge\n 0x153000 - 0x164feb +libz.1.dylib ??? (???)\n<14e90988b3fb4386b8fed7b6b13e5dc7> /opt/local/lib/libz.1.dylib\n 0x169000 - 0x261ff4 +libiconv.2.dylib ??? (???)\n<56e0608768d0bcfc4a75df4b5a299ad0> /opt/local/lib/libiconv.2.dylib\n 0x26e000 - 0x398fe7 +libcrypto.1.0.0.dylib ??? (???)\n<84a173ed04723f87b43fc6a5bccba7df>\n/opt/local/lib/libcrypto.1.0.0.dylib\n 0x3f8000 - 0x43afeb +libssl.1.0.0.dylib ??? (???)\n /opt/local/lib/libssl.1.0.0.dylib\n0x8fe00000 - 0x8fe2db43 dyld 97.1 (???)\n<458eed38a009e5658a79579e7bc26603> /usr/lib/dyld\n0x900b4000 - 0x9021bff3 libSystem.B.dylib ??? (???)\n /usr/lib/libSystem.B.dylib\n0x911e7000 - 0x911eefe9 libgcc_s.1.dylib ??? (???)\n /usr/lib/libgcc_s.1.dylib\n0x964b8000 - 0x964bcfff libmathCommon.A.dylib ??? (???)\n/usr/lib/system/libmathCommon.A.dylib\n0xffff0000 - 0xffff1780 libSystem.B.dylib ??? (???) /usr/lib/libSystem.B.dylib\n"},{"id":"164391","messageId":"m3pqpcv5ph.fsf@localhost.localdomain","threadId":"26881","inReplyTo":"AANLkTimxHn_fwMKh9cbp9i5LkShUi=HK44nv2KJ7ENaQ@mail.gmail.com","subject":"Re: Reproducible crash in git merge","fromName":"Jakub Narebski","fromEmail":"jnareb@gmail.com","sentAt":"2011-03-27T17:23:13Z","receivedAt":"2011-03-27T17:23:13Z","isPatch":false,"sender":{"key":"jnareb@gmail.com","avatar":"https://avatars.githubusercontent.com/u/2706?v=4"},"body":"David Gilman writes:\n\n> I can get git merge to crash during a git pull. I'm using git\n> 1.7.3.2. The repo (afaict) is fine and had worked previously.\n\n\"git fsck --all\" doesn't show any _errors_, doesn't it?\n\n> Unfortunately I can't share the repo itself because of its contents.\n\nDoes anyone remember the script that can be used to anonymize contents\nof repository that cannot be made public for debugging purposes?\n\nUnfortunately I didn't save this email (on git mailing list), nor do\nI remember enough from email to find it on one of git mailing list\narchives... \n\nIt would be good if it made it into 'contrib/' area, isn't it?\n-- \nJakub Narebski\nPoland\nShadeHawk on #git\n"},{"id":"164392","messageId":"7vr59s1lzu.fsf@alter.siamese.dyndns.org","threadId":"26881","inReplyTo":"m3pqpcv5ph.fsf@localhost.localdomain","subject":"Re: Reproducible crash in git merge","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2011-03-27T18:01:09Z","receivedAt":"2011-03-27T18:01:09Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Jakub Narebski writes:\n\n> David Gilman writes:\n>\n>> Unfortunately I can't share the repo itself because of its contents.\n>\n> Does anyone remember the script that can be used to anonymize contents\n> of repository that cannot be made public for debugging purposes?\n\nThat might help in certain narrow other cases, but I suspect that it is a\nway off tangent in this case, after looking at the stack trace.\n\nThe process is dying inside pthe latform implementation of memcpy of\ninflate from zlib; unless the bug is reproduceable independent of the\ncontents, any modification done by contents anonymizer is very likely to\nmake this particular issue go away, don't you think?\n"},{"id":"164394","messageId":"AANLkTinW6pndDmpi=mewxXT=vLSePoJ2meMozCSht6Ze@mail.gmail.com","threadId":"26881","inReplyTo":"m3pqpcv5ph.fsf@localhost.localdomain","subject":"Re: Reproducible crash in git merge","fromName":"David Gilman","fromEmail":"davidgilman1@gmail.com","sentAt":"2011-03-27T20:24:58Z","receivedAt":"2011-03-27T20:24:58Z","isPatch":false,"sender":{"key":"davidgilman1@gmail.com","avatar":null},"body":"OK, well I apologize for the bug report. I was able to reproduce it 2\nor 3 times but I can no longer do so. git fsck is showing no errors.\n\nOn Sun, Mar 27, 2011 at 12:23 PM, Jakub Narebski wrote:\n> David Gilman writes:\n>\n>> I can get git merge to crash during a git pull.  I'm using git\n>> 1.7.3.2.  The repo (afaict) is fine and had worked previously.\n>\n> \"git fsck --all\" doesn't show any _errors_, doesn't it?\n>\n>> Unfortunately I can't share the repo itself because of its contents.\n>\n> Does anyone remember the script that can be used to anonymize contents\n> of repository that cannot be made public for debugging purposes?\n>\n> Unfortunately I didn't save this email (on git mailing list), nor do\n> I remember enough from email to find it on one of git mailing list\n> archives...\n>\n> It would be good if it made it into 'contrib/' area, isn't it?\n> --\n> Jakub Narebski\n> Poland\n> ShadeHawk on #git\n>\n\n\n\n-- \nDavid Gilman\n"},{"id":"164606","messageId":"201103292040.21419.jnareb@gmail.com","threadId":"26881","inReplyTo":"7vr59s1lzu.fsf@alter.siamese.dyndns.org","subject":"Re: Reproducible crash in git merge","fromName":"Jakub Narebski","fromEmail":"jnareb@gmail.com","sentAt":"2011-03-29T18:40:20Z","receivedAt":"2011-03-29T18:40:20Z","isPatch":false,"sender":{"key":"jnareb@gmail.com","avatar":"https://avatars.githubusercontent.com/u/2706?v=4"},"body":"Junio C Hamano wrote:\n> Jakub Narebski writes:\n> > David Gilman writes:\n> >\n> > > Unfortunately I can't share the repo itself because of its contents.\n> >\n> > Does anyone remember the script that can be used to anonymize contents\n> > of repository that cannot be made public for debugging purposes?\n> \n> That might help in certain narrow other cases, but I suspect that it is a\n> way off tangent in this case, after looking at the stack trace.\n\nI'm sorry, I didn't examine stack trace...\n \n> The process is dying inside the platform implementation of memcpy of\n> inflate from zlib; unless the bug is reproduceable independent of the\n> contents, any modification done by contents anonymizer is very likely to\n> make this particular issue go away, don't you think?\n\nWell, assuming that issue is involved with sizes and not content, then\npossibly as long as anonymizer preserves sizes of objects. Which is not\nvery likely.\n\nAnyway, from what I remember this heisenbug vanished mysteriously...\n-- \nJakub Narebski\nPoland\n"}]}