1Git for CVS users 2================= 3v0.99.5, Aug 2005 4 5Ok, so you're a CVS user. That's ok, it's a treatable condition, and the 6first step to recovery is admitting you have a problem. The fact that 7you are reading this file means that you may be well on that path 8already. 9 10The thing about CVS is that it absolutely sucks as a source control 11manager, and you'll thus be happy with almost anything else. Git, 12however, may be a bit 'too' different (read: "good") for your taste, and 13does a lot of things differently. 14 15One particular suckage of CVS is very hard to work around: CVS is 16basically a tool for tracking 'file' history, while git is a tool for 17tracking 'project' history. This sometimes causes problems if you are 18used to doing very strange things in CVS, in particular if you're doing 19things like making branches of just a subset of the project. Git can't 20track that, since git never tracks things on the level of an individual 21file, only on the whole project level. 22 23The good news is that most people don't do that, and in fact most sane 24people think it's a bug in CVS that makes it tag (and check in changes) 25one file at a time. So most projects you'll ever see will use CVS 26'as if' it was sane. In which case you'll find it very easy indeed to 27move over to Git. 28 29First off: this is not a git tutorial. See 30link:tutorial.html[Documentation/tutorial.txt] for how git 31actually works. This is more of a random collection of gotcha's 32and notes on converting from CVS to git. 33 34Second: CVS has the notion of a "repository" as opposed to the thing 35that you're actually working in (your working directory, or your 36"checked out tree"). Git does not have that notion at all, and all git 37working directories 'are' the repositories. However, you can easily 38emulate the CVS model by having one special "global repository", which 39people can synchronize with. See details later, but in the meantime 40just keep in mind that with git, every checked out working tree will 41have a full revision control history of its own. 42 43 44Importing a CVS archive 45----------------------- 46 47Ok, you have an old project, and you want to at least give git a chance 48to see how it performs. The first thing you want to do (after you've 49gone through the git tutorial, and generally familiarized yourself with 50how to commit stuff etc in git) is to create a git'ified version of your 51CVS archive. 52 53Happily, that's very easy indeed. Git will do it for you, although git 54will need the help of a program called "cvsps": 55 56 http://www.cobite.com/cvsps/ 57 58which is not actually related to git at all, but which makes CVS usage 59look almost sane (ie you almost certainly want to have it even if you 60decide to stay with CVS). However, git will want 'at least' version 2.1 61of cvsps (available at the address above), and in fact will currently 62refuse to work with anything else. 63 64Once you've gotten (and installed) cvsps, you may or may not want to get 65any more familiar with it, but make sure it is in your path. After that, 66the magic command line is 67 68 git cvsimport -v -d <cvsroot> -C <destination> <module> 69 70which will do exactly what you'd think it does: it will create a git 71archive of the named CVS module. The new archive will be created in the 72subdirectory named <destination>; it'll be created if it doesn't exist. 73Default is the local directory. 74 75It can take some time to actually do the conversion for a large archive 76since it involves checking out from CVS every revision of every file, 77and the conversion script is reasonably chatty unless you omit the '-v' 78option, but on some not very scientific tests it averaged about twenty 79revisions per second, so a medium-sized project should not take more 80than a couple of minutes. For larger projects or remote repositories, 81the process may take longer. 82 83After the (initial) import is done, the CVS archive's current head 84revision will be checked out -- thus, you can start adding your own 85changes right away. 86 87The import is incremental, i.e. if you call it again next month it'll 88fetch any CVS updates that have been happening in the meantime. The 89cut-off is date-based, so don't change the branches that were imported 90from CVS. 91 92You can merge those updates (or, in fact, a different CVS branch) into 93your main branch: 94 95 git resolve HEAD origin "merge with current CVS HEAD" 96 97The HEAD revision from CVS is named "origin", not "HEAD", because git 98already uses "HEAD". (If you don't like 'origin', use cvsimport's 99'-o' option to change it.) 100 101 102Emulating CVS behaviour 103----------------------- 104 105 106So, by now you are convinced you absolutely want to work with git, but 107at the same time you absolutely have to have a central repository. 108Step back and think again. Okay, you still need a single central 109repository? There are several ways to go about that: 110 1111. Designate a person responsible to pull all branches. Make the 112repository of this person public, and make every team member 113pull regularly from it. 114 1152. Set up a public repository with read/write access for every team 116member. Use "git pull/push" as you used "cvs update/commit". Be 117sure that your repository is up to date before pushing, just 118like you used to do with "cvs commit"; your push will fail if 119what you are pushing is not up to date. 120 1213. Make the repository of every team member public. It is the 122responsibility of each single member to pull from every other 123team member. 124 125 126CVS annotate 127------------ 128 129So, something has gone wrong, and you don't know whom to blame, and 130you're an ex-CVS user and used to do "cvs annotate" to see who caused 131the breakage. You're looking for the "git annotate", and it's just 132claiming not to find such a script. You're annoyed. 133 134Yes, that's right. Core git doesn't do "annotate", although it's 135technically possible, and there are at least two specialized scripts out 136there that can be used to get equivalent information (see the git 137mailing list archives for details). 138 139Git has a couple of alternatives, though, that you may find sufficient 140or even superior depending on your use. One is called "git-whatchanged" 141(for obvious reasons) and the other one is called "pickaxe" ("a tool for 142the software archeologist"). 143 144The "git-whatchanged" script is a truly trivial script that can give you 145a good overview of what has changed in a file or a directory (or an 146arbitrary list of files or directories). The "pickaxe" support is an 147additional layer that can be used to further specify exactly what you're 148looking for, if you already know the specific area that changed. 149 150Let's step back a bit and think about the reason why you would 151want to do "cvs annotate a-file.c" to begin with. 152 153You would use "cvs annotate" on a file when you have trouble 154with a function (or even a single "if" statement in a function) 155that happens to be defined in the file, which does not do what 156you want it to do. And you would want to find out why it was 157written that way, because you are about to modify it to suit 158your needs, and at the same time you do not want to break its 159current callers. For that, you are trying to find out why the 160original author did things that way in the original context. 161 162Many times, it may be enough to see the commit log messages of 163commits that touch the file in question, possibly along with the 164patches themselves, like this: 165 166 $ git-whatchanged -p a-file.c 167 168This will show log messages and patches for each commit that 169touches a-file. 170 171This, however, may not be very useful when this file has many 172modifications that are not related to the piece of code you are 173interested in. You would see many log messages and patches that 174do not have anything to do with the piece of code you are 175interested in. As an example, assuming that you have this piece 176of code that you are interested in in the HEAD version: 177 178 if (frotz) { 179 nitfol(); 180 } 181 182you would use git-rev-list and git-diff-tree like this: 183 184 $ git-rev-list HEAD | 185 git-diff-tree --stdin -v -p -S'if (frotz) { 186 nitfol(); 187 }' 188 189We have already talked about the "\--stdin" form of git-diff-tree 190command that reads the list of commits and compares each commit 191with its parents. The git-whatchanged command internally runs 192the equivalent of the above command, and can be used like this: 193 194 $ git-whatchanged -p -S'if (frotz) { 195 nitfol(); 196 }' 197 198When the -S option is used, git-diff-tree command outputs 199differences between two commits only if one tree has the 200specified string in a file and the corresponding file in the 201other tree does not. The above example looks for a commit that 202has the "if" statement in it in a file, but its parent commit 203does not have it in the same shape in the corresponding file (or 204the other way around, where the parent has it and the commit 205does not), and the differences between them are shown, along 206with the commit message (thanks to the -v flag). It does not 207show anything for commits that do not touch this "if" statement. 208 209Also, in the original context, the same statement might have 210appeared at first in a different file and later the file was 211renamed to "a-file.c". CVS annotate would not help you to go 212back across such a rename, but GIT would still help you in such 213a situation. For that, you can give the -C flag to 214git-diff-tree, like this: 215 216 $ git-whatchanged -p -C -S'if (frotz) { 217 nitfol(); 218 }' 219 220When the -C flag is used, file renames and copies are followed. 221So if the "if" statement in question happens to be in "a-file.c" 222in the current HEAD commit, even if the file was originally 223called "o-file.c" and then renamed in an earlier commit, or if 224the file was created by copying an existing "o-file.c" in an 225earlier commit, you will not lose track. If the "if" statement 226did not change across such a rename or copy, then the commit that 227does rename or copy would not show in the output, and if the 228"if" statement was modified while the file was still called 229"o-file.c", it would find the commit that changed the statement 230when it was in "o-file.c". 231 232[ BTW, the current versions of "git-diff-tree -C" is not eager 233 enough to find copies, and it will miss the fact that a-file.c 234 was created by copying o-file.c unless o-file.c was somehow 235 changed in the same commit.] 236 237You can use the --pickaxe-all flag in addition to the -S flag. 238This causes the differences from all the files contained in 239those two commits, not just the differences between the files 240that contain this changed "if" statement: 241 242 $ git-whatchanged -p -C -S'if (frotz) { 243 nitfol(); 244 }' --pickaxe-all 245 246[ Side note. This option is called "--pickaxe-all" because -S 247 option is internally called "pickaxe", a tool for software 248 archaeologists.]