1Git for CVS users 2================= 3 4Ok, so you're a CVS user. That's ok, it's a treatable condition, and the 5first step to recovery is admitting you have a problem. The fact that 6you are reading this file means that you may be well on that path 7already. 8 9The thing about CVS is that it absolutely sucks as a source control 10manager, and you'll thus be happy with almost anything else. Git, 11however, may be a bit _too_ different (read: "good") for your taste, and 12does a lot of things differently. 13 14One particular suckage of CVS is very hard to work around: CVS is 15basically a tool for tracking _file_ history, while git is a tool for 16tracking _project_ history. This sometimes causes problems if you are 17used to doing very strange things in CVS, in particular if you're doing 18things like making branches of just a subset of the project. Git can't 19track that, since git never tracks things on the level of an individual 20file, only on the whole project level. 21 22The good news is that most people don't do that, and in fact most sane 23people think it's a bug in CVS that makes it tag (and check in changes) 24one file at a time. So most projects you'll ever see will use CVS 25_as_if_ it was sane. In which case you'll find it very easy indeed to 26move over to Git. 27 28First off: this is not a git tutorial. See Documentation/tutorial.txt 29for how git actually works. This is more of a random collection of 30gotcha's and notes on converting from CVS to git. 31 32Second: CVS has the notion of a "repository" as opposed to the thing 33that you're actually working in (your working directory, or your 34"checked out tree"). Git does not have that notion at all, and all git 35working directories _are_ the repositories. However, you can easily 36emulate the CVS model by having one special "global repository", which 37people can synchronize with. See details later, but in the meantime 38just keep in mind that with git, every checked out working tree will 39have a full revision control history of its own. 40 41 42Importing a CVS archive 43----------------------- 44 45Ok, you have an old project, and you want to at least give git a chance 46to see how it performs. The first thing you want to do (after you've 47gone through the git tutorial, and generally familiarized yourself with 48how to commit stuff etc in git) is to create a git'ified version of your 49CVS archive. 50 51Happily, that's very easy indeed. Git will do it for you, although git 52will need the help of a program called "cvsps": 53 54 http://www.cobite.com/cvsps/ 55 56which is not actually related to git at all, but which makes CVS usage 57look almost sane (ie you almost certainly want to have it even if you 58decide to stay with CVS). However, git will want at _least_ version 2.1 59of cvsps (available at the address above), and in fact will currently 60refuse to work with anything else. 61 62Once you've gotten (and installed) cvsps, you may or may not want to get 63any more familiar with it, but make sure it is in your path. After that, 64the magic command line is 65 66 git cvsimport -d <cvsroot> <module> <destination> 67 68which will do exactly what you'd think it does: it will create a git 69archive of the named CVS module. The new archive will be created in the 70subdirectory named <destination>; it'll be created if it doesn't exist. 71Default is the local directory. 72 73It can take some time to actually do the conversion for a large archive 74since it involves checking out from CVS every revision of every file, 75and the conversion script can be reasonably chatty, but on some not very 76scientific tests it averaged about twenty revisions per second, so a 77medium-sized project should not take more than a couple of minutes. For 78larger projects or remote repositories, the process may take longer. 79 80After the import is done, do this: 81 82 cp .git/refs/heads/<branch> .git/refs/heads/master 83 git-read-tree 84 git-checkout-cache -q -f -u -a 85 86The head branch is named "origin" by default; you can change that using 87the '-o' option to "git cvsimport". 88 89The import is incremental, i.e. if you call it again next month it'll 90fetch any CVS updates that have been happening in the meantime. You can 91then merge those updates into your main branch: 92 93 cg-merge <branch> 94 95 96Emulating CVS behaviour 97----------------------- 98 99 100FIXME! Talk about setting up several repositories, and pulling and 101pushing between them. Talk about merging, and branches. Some of this 102needs to be in the tutorial too. 103 104 105 106CVS annotate 107------------ 108 109So, something has gone wrong, and you don't know whom to blame, and 110you're an ex-CVS user and used to do "cvs annotate" to see who caused 111the breakage. You're looking for the "git annotate", and it's just 112claiming not to find such a script. You're annoyed. 113 114Yes, that's right. Core git doesn't do "annotate", although it's 115technically possible, and there are at least two specialized scripts out 116there that can be used to get equivalent information (see the git 117mailing list archives for details). 118 119Git has a couple of alternatives, though, that you may find sufficient 120or even superior depending on your use. One is called "git-whatchanged" 121(for obvious reasons) and the other one is called "pickaxe" ("a tool for 122the software archeologist"). 123 124The "git-whatchanged" script is a truly trivial script that can give you 125a good overview of what has changed in a file or a directory (or an 126arbitrary list of files or directories). The "pickaxe" support is an 127additional layer that can be used to further specify exactly what you're 128looking for, if you already know the specific area that changed. 129 130Let's step back a bit and think about the reason why you would 131want to do "cvs annotate a-file.c" to begin with. 132 133You would use "cvs annotate" on a file when you have trouble 134with a function (or even a single "if" statement in a function) 135that happens to be defined in the file, which does not do what 136you want it to do. And you would want to find out why it was 137written that way, because you are about to modify it to suit 138your needs, and at the same time you do not want to break its 139current callers. For that, you are trying to find out why the 140original author did things that way in the original context. 141 142Many times, it may be enough to see the commit log messages of 143commits that touch the file in question, possibly along with the 144patches themselves, like this: 145 146 $ git-whatchanged -p a-file.c 147 148This will show log messages and patches for each commit that 149touches a-file. 150 151This, however, may not be very useful when this file has many 152modifications that are not related to the piece of code you are 153interested in. You would see many log messages and patches that 154do not have anything to do with the piece of code you are 155interested in. As an example, assuming that you have this piece 156code that you are interested in in the HEAD version: 157 158 if (frotz) { 159 nitfol(); 160 } 161 162you would use git-rev-list and git-diff-tree like this: 163 164 $ git-rev-list HEAD | 165 git-diff-tree --stdin -v -p -S'if (frotz) { 166 nitfol(); 167 }' 168 169We have already talked about the "--stdin" form of git-diff-tree 170command that reads the list of commits and compares each commit 171with its parents. The git-whatchanged command internally runs 172the equivalent of the above command, and can be used like this: 173 174 $ git-whatchanged -p -S'if (frotz) { 175 nitfol(); 176 }' 177 178When the -S option is used, git-diff-tree command outputs 179differences between two commits only if one tree has the 180specified string in a file and the corresponding file in the 181other tree does not. The above example looks for a commit that 182has the "if" statement in it in a file, but its parent commit 183does not have it in the same shape in the corresponding file (or 184the other way around, where the parent has it and the commit 185does not), and the differences between them are shown, along 186with the commit message (thanks to the -v flag). It does not 187show anything for commits that do not touch this "if" statement. 188 189Also, in the original context, the same statement might have 190appeared at first in a different file and later the file was 191renamed to "a-file.c". CVS annotate would not help you to go 192back across such a rename, but GIT would still help you in such 193a situation. For that, you can give the -C flag to 194git-diff-tree, like this: 195 196 $ git-whatchanged -p -C -S'if (frotz) { 197 nitfol(); 198 }' 199 200When the -C flag is used, file renames and copies are followed. 201So if the "if" statement in question happens to be in "a-file.c" 202in the current HEAD commit, even if the file was originally 203called "o-file.c" and then renamed in an earlier commit, or if 204the file was created by copying an existing "o-file.c" in an 205earlier commit, you will not lose track. If the "if" statement 206did not change across such rename or copy, then the commit that 207does rename or copy would not show in the output, and if the 208"if" statement was modified while the file was still called 209"o-file.c", it would find the commit that changed the statement 210when it was in "o-file.c". 211 212[ BTW, the current versions of "git-diff-tree -C" is not eager 213 enough to find copies, and it will miss the fact that a-file.c 214 was created by copying o-file.c unless o-file.c was somehow 215 changed in the same commit.] 216 217You can use the --pickaxe-all flag in addition to the -S flag. 218This causes the differences from all the files contained in 219those two commits, not just the differences between the files 220that contain this changed "if" statement: 221 222 $ git-whatchanged -p -C -S'if (frotz) { 223 nitfol(); 224 }' --pickaxe-all 225 226[ Side note. This option is called "--pickaxe-all" because -S 227 option is internally called "pickaxe", a tool for software 228 archaeologists.]