Why not use a taginfo trigger to capture tag data. Below is a sample
output of the perl script I use.
Repository: /cvs_repositories/m2k_cvsroot/m2k/PCP_DEV/Players/AirCAT/src
Tag: RadCAT_Version_0_2
Operation: add
Date: Fri Nov 5 15:55:01 1999
Tagged-By: larkinbr
File: AirCAT.cdf Revision: 1.1.1.1
File: AirCAT.h Revision: 1.1.1.1
File: AirCAT_Platform.cdf Revision: 1.1.1.1
File: AirCAT_Platform.h Revision: 1.1.1.1
File: AirCAT_Platform_func.cpp Revision: 1.1.1.1
File: AirCAT_func.cpp Revision: 1.1.1.1
File: files.make Revision: 1.1.1.1
File: makefile Revision: 1.1.1.1
Mike Sutton | public class
SAIC | software_failure : management_failure
Beavercreek, OH |
[EMAIL PROTECTED] | These are MY opinions, not SAIC's
On 11/28/00 17:24:57, Derek R. Price wrote:
> Donald Sharp wrote:
>
> > Here's a first crack at it. Probably still needs some work, but wanted
> > to get some feedback ;) It's attached to the end of this email...
>
> That's not a complete fix. Here's what I noticed:
>
> 1) There's no filename or directory saved to the history log, leaving the
> user with no idea which files were tagged. You need to list both filename
> and directory since, unlike rtag, tag can specify individual files, each
> tagged file must be listed in the history separately, like a commit history
> entry. Note that this will require a new record type.
>
> 2) 'A' is not really appropriate in lieu of a revision number or symbolic
> "base" tag. A user would have to run a second command to find which revision
> number the tag was attached to. This information should be added for the tag
> delete log too since the user would not have to trace back through the log to
> find what the last tag op to that file was to know where the delete happened
> - information which might not be available if the history log was ever reset
> and which would be hard to calculate based on the current rtag logs. I would
> advocate that the symtag should be listed if in the original tag command, but
> whatever the final affected numerical tag is should be listed too. If the
> tag was a move that should probably be listed too, perhaps as a "delete" then
> a "tag", but a single entry might be more succinct and there should still be
> plenty of namespace, especially if you make "delete", "add", and "move" sub
> record types of "tag". I'm thinking it might be appropriate to alter rtag to
> use the new log format as well.
>
> 3) Don't forget that a final patch submission should contain a ChangeLog
> entry and any necessary doc changes, but I'm sure I didn't need to remind you
> about that. :)
>
> Derek
>
_______________________________________________
Info-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/info-cvs
