1git-bisect(1) 2============= 3 4NAME 5---- 6git-bisect - Find the change that introduced a bug by binary search 7 8 9SYNOPSIS 10-------- 11'git bisect' <subcommand> <options> 12 13DESCRIPTION 14----------- 15The command takes various subcommands, and different options depending 16on the subcommand: 17 18 git bisect help 19 git bisect start [<bad> [<good>...]] [--] [<paths>...] 20 git bisect bad [<rev>] 21 git bisect good [<rev>...] 22 git bisect skip [(<rev>|<range>)...] 23 git bisect reset [<branch>] 24 git bisect visualize 25 git bisect replay <logfile> 26 git bisect log 27 git bisect run <cmd>... 28 29This command uses 'git rev-list --bisect' to help drive the 30binary search process to find which change introduced a bug, given an 31old "good" commit object name and a later "bad" commit object name. 32 33Getting help 34~~~~~~~~~~~~ 35 36Use "git bisect" to get a short usage description, and "git bisect 37help" or "git bisect -h" to get a long usage description. 38 39Basic bisect commands: start, bad, good 40~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 41 42The way you use it is: 43 44------------------------------------------------ 45$ git bisect start 46$ git bisect bad # Current version is bad 47$ git bisect good v2.6.13-rc2 # v2.6.13-rc2 was the last version 48 # tested that was good 49------------------------------------------------ 50 51When you give at least one bad and one good versions, it will bisect 52the revision tree and say something like: 53 54------------------------------------------------ 55Bisecting: 675 revisions left to test after this 56------------------------------------------------ 57 58and check out the state in the middle. Now, compile that kernel, and 59boot it. Now, let's say that this booted kernel works fine, then just 60do 61 62------------------------------------------------ 63$ git bisect good # this one is good 64------------------------------------------------ 65 66which will now say 67 68------------------------------------------------ 69Bisecting: 337 revisions left to test after this 70------------------------------------------------ 71 72and you continue along, compiling that one, testing it, and depending 73on whether it is good or bad, you say "git bisect good" or "git bisect 74bad", and ask for the next bisection. 75 76Until you have no more left, and you'll have been left with the first 77bad kernel rev in "refs/bisect/bad". 78 79Bisect reset 80~~~~~~~~~~~~ 81 82Oh, and then after you want to reset to the original head, do a 83 84------------------------------------------------ 85$ git bisect reset 86------------------------------------------------ 87 88to get back to the original branch, instead of being on the bisection 89commit ("git bisect start" will do that for you too, actually: it will 90reset the bisection state). 91 92Bisect visualize 93~~~~~~~~~~~~~~~~ 94 95During the bisection process, you can say 96 97------------ 98$ git bisect visualize 99------------ 100 101to see the currently remaining suspects in 'gitk'. `visualize` is a bit 102too long to type and `view` is provided as a synonym. 103 104If 'DISPLAY' environment variable is not set, 'git log' is used 105instead. You can even give command line options such as `-p` and 106`--stat`. 107 108------------ 109$ git bisect view --stat 110------------ 111 112Bisect log and bisect replay 113~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 114 115The good/bad input is logged, and 116 117------------ 118$ git bisect log 119------------ 120 121shows what you have done so far. You can truncate its output somewhere 122and save it in a file, and run 123 124------------ 125$ git bisect replay that-file 126------------ 127 128if you find later you made a mistake telling good/bad about a 129revision. 130 131Avoiding to test a commit 132~~~~~~~~~~~~~~~~~~~~~~~~~ 133 134If in a middle of bisect session, you know what the bisect suggested 135to try next is not a good one to test (e.g. the change the commit 136introduces is known not to work in your environment and you know it 137does not have anything to do with the bug you are chasing), you may 138want to find a near-by commit and try that instead. 139 140It goes something like this: 141 142------------ 143$ git bisect good/bad # previous round was good/bad. 144Bisecting: 337 revisions left to test after this 145$ git bisect visualize # oops, that is uninteresting. 146$ git reset --hard HEAD~3 # try 3 revs before what 147 # was suggested 148------------ 149 150Then compile and test the one you chose to try. After that, tell 151bisect what the result was as usual. 152 153Bisect skip 154~~~~~~~~~~~~ 155 156Instead of choosing by yourself a nearby commit, you may just want git 157to do it for you using: 158 159------------ 160$ git bisect skip # Current version cannot be tested 161------------ 162 163But computing the commit to test may be slower afterwards and git may 164eventually not be able to tell the first bad among a bad and one or 165more "skip"ped commits. 166 167You can even skip a range of commits, instead of just one commit, 168using the "'<commit1>'..'<commit2>'" notation. For example: 169 170------------ 171$ git bisect skip v2.5..v2.6 172------------ 173 174would mean that no commit between `v2.5` excluded and `v2.6` included 175can be tested. 176 177Note that if you want to also skip the first commit of a range you can 178use something like: 179 180------------ 181$ git bisect skip v2.5 v2.5..v2.6 182------------ 183 184and the commit pointed to by `v2.5` will be skipped too. 185 186Cutting down bisection by giving more parameters to bisect start 187~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 188 189You can further cut down the number of trials if you know what part of 190the tree is involved in the problem you are tracking down, by giving 191paths parameters when you say `bisect start`, like this: 192 193------------ 194$ git bisect start -- arch/i386 include/asm-i386 195------------ 196 197If you know beforehand more than one good commits, you can narrow the 198bisect space down without doing the whole tree checkout every time you 199give good commits. You give the bad revision immediately after `start` 200and then you give all the good revisions you have: 201 202------------ 203$ git bisect start v2.6.20-rc6 v2.6.20-rc4 v2.6.20-rc1 -- 204 # v2.6.20-rc6 is bad 205 # v2.6.20-rc4 and v2.6.20-rc1 are good 206------------ 207 208Bisect run 209~~~~~~~~~~ 210 211If you have a script that can tell if the current source code is good 212or bad, you can automatically bisect using: 213 214------------ 215$ git bisect run my_script 216------------ 217 218Note that the "run" script (`my_script` in the above example) should 219exit with code 0 in case the current source code is good. Exit with a 220code between 1 and 127 (inclusive), except 125, if the current 221source code is bad. 222 223Any other exit code will abort the automatic bisect process. (A 224program that does "exit(-1)" leaves $? = 255, see exit(3) manual page, 225the value is chopped with "& 0377".) 226 227The special exit code 125 should be used when the current source code 228cannot be tested. If the "run" script exits with this code, the current 229revision will be skipped, see `git bisect skip` above. 230 231You may often find that during bisect you want to have near-constant 232tweaks (e.g., s/#define DEBUG 0/#define DEBUG 1/ in a header file, or 233"revision that does not have this commit needs this patch applied to 234work around other problem this bisection is not interested in") 235applied to the revision being tested. 236 237To cope with such a situation, after the inner 'git bisect' finds the 238next revision to test, with the "run" script, you can apply that tweak 239before compiling, run the real test, and after the test decides if the 240revision (possibly with the needed tweaks) passed the test, rewind the 241tree to the pristine state. Finally the "run" script can exit with 242the status of the real test to let the "git bisect run" command loop to 243determine the outcome. 244 245EXAMPLES 246-------- 247 248* Automatically bisect a broken build between v1.2 and HEAD: 249+ 250------------ 251$ git bisect start HEAD v1.2 -- # HEAD is bad, v1.2 is good 252$ git bisect run make # "make" builds the app 253------------ 254 255* Automatically bisect a broken test suite: 256+ 257------------ 258$ cat ~/test.sh 259#!/bin/sh 260make || exit 125 # this "skip"s broken builds 261make test # "make test" runs the test suite 262$ git bisect start v1.3 v1.1 -- # v1.3 is bad, v1.1 is good 263$ git bisect run ~/test.sh 264------------ 265+ 266Here we use a "test.sh" custom script. In this script, if "make" 267fails, we "skip" the current commit. 268+ 269It's safer to use a custom script outside the repo to prevent 270interactions between the bisect, make and test processes and the 271script. 272+ 273And "make test" should "exit 0", if the test suite passes, and 274"exit 1" (for example) otherwise. 275 276* Automatically bisect a broken test case: 277+ 278------------ 279$ cat ~/test.sh 280#!/bin/sh 281make || exit 125 # this "skip"s broken builds 282~/check_test_case.sh # does the test case passes ? 283$ git bisect start HEAD HEAD~10 -- # culprit is among the last 10 284$ git bisect run ~/test.sh 285------------ 286+ 287Here "check_test_case.sh" should "exit 0", if the test case passes, 288and "exit 1" (for example) otherwise. 289+ 290It's safer if both "test.sh" and "check_test_case.sh" scripts are 291outside the repo to prevent interactions between the bisect, make and 292test processes and the scripts. 293 294Author 295------ 296Written by Linus Torvalds <torvalds@osdl.org> 297 298Documentation 299------------- 300Documentation by Junio C Hamano and the git-list <git@vger.kernel.org>. 301 302GIT 303--- 304Part of the linkgit:git[1] suite