6

This seems like it should be so simple, but I can't find any solution that appears to work...

I need a CVS command that given the name of a tag that you have applied to a file, it will give you the revision number.

CVS Tree structure:

(filename)
    |
    +--> 1.1-----(branch)
          |         |
          |      1.1.1.1---(tag1)
          |         |
          |      1.1.1.2---(tag2)
          |         |
          |      1.1.1.3---(tag3)
          |         |
          |         :
         1.2
          |
          |
          :

For example: Using a CVS command, given the tag name "tag2", how can I get CVS to give me the revision number "1.1.1.2"?

The closest thing I can find is using the log command with the -Q flag, but that still gives me much more information than I need.

ex: cvs -Q log -h filename

Passing the tagname to the log command seems to have no effect.

CVS Version information:

enter image description here

My current solution is to use a perl script to parse the output from the log command but there has to be a simpler way...

tjwrona1992
  • 8,614
  • 8
  • 35
  • 98

1 Answers1

1

Passing a tag name (with a -r option) to the log command does have an effect, just not a particularly useful one and it's effect is hidden by "-h".

Usually the easiest way to get a revision number for your VERSION file (the normal use-case for this) is to include a keyword in it; ie:

# Thu 21 May 08:40:59 BST 2015
THISREV="$Revision$"

Note: to get a repository version number this VERSION file must be committed every time you make a commit to the repo.

If you need a revision for a specific file then you're basically falling back on scripting from the "symbolic names" part of the log. So for r-1-0-0 you do this:

cvs -Q log -h VERSION | awk '/^\tr-1-0-0:/ {print $NF;}'

There's no direct equivalent of the git describe command.

user3710044
  • 2,261
  • 15
  • 15
  • Unfortunately I am not the one creating the files that are going into the repository so I have no way of enforcing that a version line be added within each file. Right now I am using a Perl script to parse the log command output, but I was just really hoping there was a faster easier way. – tjwrona1992 May 21 '15 at 12:39
  • Although you didn't fully answer my question you did come up with a solution so I have awarded the bounty to you. For now I will stick to using a Perl script since your suggestion isn't possible for me. – tjwrona1992 May 28 '15 at 13:07
  • I've accepted this answer because the awk solution does work for me on Linux and its highly unlikely that this will get any more responses. I would like a solution that works on windows though. – tjwrona1992 Sep 18 '15 at 20:07
  • The `git-bash` that's installed as a standard part of the plain git tools comes with a compile of `gawk` as should all the `msys` compilations. The `mawk.exe` program has been available as a native win32 program for many years. Then of course there's the `cygwin` stuff, which IMO would be serious overkill (and complicated) for "just an awk command". – user3710044 Sep 19 '15 at 09:52