When you merge the activity history, why the date is not used as a condition. about git-cc

[charleso]

Hi hanjie0,

Actually I think that's what I started with originally, but it would miss things. I'm not sure I would accept a patch that does exactly that, but perhaps something like:

a.subject == b.subject and a.user == b.user and (b.date - a.date) < X

(Where I'm not exactly sure what X would be)

Out of curiosity is there any particular case where both the subject and user is matching up but it shouldn't be a single commit?

[hanjie0]

subject is the got from first part of the comments of the checkin operation. And most of time, the developers will not write the comments seriously. the simple comments will cause the same subject of a checkin/cs.

I just cannot understand why we do not just add a.date ==b.date as additional condition?
For CC, checkin operation is a atomic operation, and the date should be the same, right?

[charleso]

For CC, checkin operation is a atomic operation

You're joking right? Part of the reason it's so hard to integrate with Clearcase is that everything is done at the file level - it has to checkin one at a time. Even UCM has to synthesise an atomic commit by using tags for everything.

It's been a long time, but I'm pretty sure if you look at the lshistory of a reasonably large commit you may find the files have slightly different timestamps.

Feel free to fork git-cc and make the change, it's not like this is an actively maintained project... :)

[hanjie0]

Dear charleso, I am so sorry, I did not do the test myself, in fact, CC environment is not available for me up to now.
I mentioned/trusted "checkin operation is a atomic operation" since I got it from the CC user manual.

I will do the test and then back to you