CVS and viewcvs Quick Memo

Original Japanese version (By Suzuki) is here.

• CVS memo
• viewcvs quick memo

CVS memo

• Initialize
  

  Generating a cvsroot directory and its managing files. It is only once for each CVSROOT. Usually, you only need to do this once for your system.

       % cvs init      
       

• Submit a directory structure. (The first add command.)
  

       % cvs import (position in the repository) vender-tag release-tag     
       Ex.: % cvs import project-1/foo susuki SNAP19980610
       

• Add files

       % cvs add filename
       and
       % cvs commit
       

  Notice: You can only add files under your current directory.

• Commit/update your change to repository.

  If you already add files to repository, just do cvs commit.

• Delete files.

       % rm file-a
       % cvs remove file-a
       

  If you haven't rm file-a, cvs complains it.

• Get the newest version

       % cvs checkout (the position of top directory in the repository)
       

  This makes a project under your current directory. And Using with the module function is more useful.

  Or, If you've already checkouted once,

       % cvs update
       

  makes update current directory recursively. But this is only effecting existing directory. If you want to update directory structure,

       % cvs update -Pd
       

  Each option means: P: Prune empty directories, d: Create any directories that exist in the repository if they're missing from the working directory.

• Get an old version. If you want to get the specified-day version (e.g., 2005-05-31).

       % cvs checkout -D 2005-05-31 module
       

• Making snapshot with tags.

       % cvs tag tag-name     
       

  adds tags under current directory. If you want to get this by tag name, do next:

       % cvs checkout -r tag-name (the position of top directory in the repository)
       

  If you want to delete tag use -d. Sometimes I misspelled tags name...

       % cvs tag -d tag-name-you-want-to-delete
       

• Using $Id$

  If this string sequence is found in your file, CVS(Actually RCS) replace this with Id information. Use this in comments.

• Making module

  This is useful! First, you should checkout the managing files of cvs via:

       % cvs checkout CVSROOT
       

  Now you get the directory CVSROOT under your current directory. Then you edit the modules in this directory. For example,

       knuthchap2 project/knuth/Vol1/taocp1/chap2
       

  After this, you can checkout chap2 with only next:

       % cvs checkout chap2
       

• A notice of subcommands recursive ability.

  CVS manual said, ``Almost all'' of the subcommands of CVS work recursively when you specify a directory as an argument. Almost all means, for example, the add subcommand is not effect recursively.

• Diff the files between in the current directory and in the repository
  

  You can get the difference between a current file and the file in the repository which is committed last time by:

       % cvs diff filename
       

• Check the updating status

       % cvs status
       

  You can only know the your own change by cvs diff. If you want to know the others change, use this.

• -n option
  

  Same as make's -n command. This do not change any files. You can check which files are changed without any change with:

       % cvs -n update
       

  Sometimes it is useful using status.

• Conflicts

  When you've get a heavy conflict, your file becomes totally mess. The only solution is read the file. Before updating, you can check file status with cvs -n update. One the other way is that you return to one before version.

• Making a branch
  

  You need to add tags to your file before making branch. It seems that, first making tags, and next branch information is added to the tags.

       % cvs rtag -b -r tag-name branch-name module-name
       

  The difference with tag is that you need only to specify the branch to get the newest revision of the branch.

• Ignoring specific files.
  • General

    cvs read .cvsignore in each directory. If you do not want commit the files, e.g., ppm files and jpeg files, then you put this file list in your .cvsignore file.

    	  *.ppm
    	  *.jpg
    	  

  • $HOME/.cvsignore

    This file globally affects in your environment. You need to take care to use this, but useful.

• Delete sticky tags

       % cvs update -A
       

• From RCS to CVS (this information comes from CVS info file.): CVS uses RCS to manage each file revision information. Then, you can just move to files to the repository.
  • Imports files without RCS related files. Next is an example.
    1. Assume next files you have.

       	       % ls
       	       RCS  foo.C  bar.hh
       	       

    2. Move RCS related files to somewhere and import the files.

       	       % mv RCS ../RCS.bak	# move
       	       % cvs -d /repository/cvsroot import proj yamauchi snap1
       	       

    3. Remove RCS files created by cvs, and replace RCS's ,v files.

       	       % rm /repository/cvsroot/proj/foo.C,v
       	       % rm /repository/cvsroot/proj/bar.hh,v
       	       % cp ../RCS.bak/foo.C,v /repository/cvsroot/proj/
       	       % cp ../RCS.bak/bar.hh,v /repository/cvsroot/proj/
       	       

    4. Make sure to the revision information is OK with checking out. For the safety, you should keep original files to somewhere. In this example, revnum is just an adequate revision number. (here, 1.1)

       	       % mkdir tmp; cd tmp
       	       % cvs -d /repository/cvsroot checkout proj
       	       % cvs update -r[revnum] foo.C
       	       

view cvs memo

• viewcvs : view cvs repository with web browser. (I think Web browser is a good GUI of cvs repository browsing.)
• Install
  • download from http://viewcvs.sourceforge.net/
  • expand tarball, cd viewcvs-version, and run ./viewcvs-install
• Standalone mode.
  • go installed dir.
  • ./standalone.py -g -r /your/repository/directory