| Article: |
Running Arbitrary Scripts Under CVS | |
| Subject: | Using python script with verifymsg | |
| Date: | 2003-08-04 03:21:22 | |
| From: | anonymous2 | |
|
I'm using python script to check log files.
|
||
Showing messages 1 through 3 of 3.
-
Using python script with verifymsg
2003-08-04 15:00:58 anonymous2 [Reply | View]
-
Using python script with verifymsg
2003-08-05 01:57:13 anonymous2 [Reply | View]
So I guess it's not possible to open editor and re-edit temp log message if original message has been noticed to be incomplete ?? Is this due to client/server model ?? -
Using python script with verifymsg
2003-08-05 15:52:02 anonymous2 [Reply | View]
It is possible - you are, after all, writing a separate program that could do just about anything.
But:
* Some client-server models don't allow such interaction.
* You are adding arbitrary data to the socket CVS is using for its data, and could accidentally trigger some other CVS command.
* You could generate a deadlock, as one person's commit can't proceed because it's waiting on another person doing their editing, but the other person is waiting on the first to do editing. This might or might not be a problem in a loginfo script, which runs after the commit is processed, but would be a problem in a verifymsg script.
In addition, CVS doesn't re-read the log message after a loginfo script, so any changes you make are ignored. It can re-read the log message after a verifymsg script, however. (See my previous response.)
Jenn Vesperman.




You may find that verifymsg is more useful to you when checking whether the user has written a proper log file entry. If you combine that with the 'ReReadLogAfterVerify' option in the CVSROOT/config file, your script can change the log message that CVS stores.
If there is simply not enough information in the log file, you can refuse to allow the commit to proceed.