[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: Adding rev id's to source files (ITS#41)



At 02:22 AM 1/13/99 GMT, brad.rubenstein@gs.com wrote:
>Full_Name: Brad Rubenstein
>Version: 1.1.2
>OS: solaris 2.6
>URL: ftp://ftp.openldap.org/incoming/
>Submission from: (NULL) (208.168.16.140)
>
>
>A suggestion: if we put /* $Header$ */ at the top of all 
>source files, I can easily figure out the revision number of some problem code,

I had problem initially with CVS variable substitute causing conflicts
during merges.  However, this was primarily with preexisting RCS tags
from the U-Mich distribution.   I may try to add new tags to a couple
of a file or two, but if they cause conflicts they are gone (merges
are hard enough).  If they don't cause problems, then we can consider
adding them across the board.

>when I have no context except the source file (long after I've pulled a release
>and
>unzipped it,
>which release it is and what patchlevel, is information that I no longer have).
>Alternatively (and perhaps easier), when you
>build the tgz files for the distributions,
>leave the CVS directories in, so after the tgz is long gone, there is still
>enough information in the source code to figure out
>which release the sources came from.

The CVS directories are only useful if you have CVS.  If you have CVS,
you should really grab OpenLDAP via AnonCVS or CVSup.  And, if you applied
a patch, the patch (generated using CVSweb) the files in CVS wouldn't
be updated.