On Tue, 2011-03-22 at 10:54 -0400, Vitezslav Humpa wrote:
- I'd support changing test_vars['hook'] to test_vars['event'] or
something similar, since we'd like to get rid of the 'hook' term overall, IIRC.
- Connects to (1) - if you decide to change that, then you'll also
need to change the control.autoqa files (specifically the "if hook not in [xyz, abc]" lines).
- I know it's not a programming issue, but changing README to
reflect current state of watcher/event parsing would be nice.
We've just went through the all the changes with Kamil and decided to go with all the three things above just like you recommended - basically to go "deeper" and really remove the occurrence of "hook" everywhere :)
- couple more things
- do not modify the clean behavior of Makefile (clean just the src
dir)
- edit specfile accordingly (Something I didn't do before)
- rename the watchers all to watcher.py and rename their dirs as well
I have just done what is mentioned above. Basically the "hook" term was finally eliminated from entire project. Considering the spread usage in code about everywhere, I hope I did not screw up some functionality.
James, If you can, please check, whether the changes to the SPEC file are fine, I have ignored it before.
Double check the Makefile as well. If you can run 'make rpms' successfully, that should be good.
Only place where I did not mess with the "hook" term is the git-post-receive watcher,as I was not confident with my understanding of various uses of "hook" in it, so if you guys don't mind, I'd keep refactoring of watchers/git-post-receive/* files (+related parts of Makefile and specfile) on it's creator (James?).
All has been pushed to origin/vhumpa again.
Maybe this was mentioned earlier, apologies if I missed it. Why the renaming of the events to drop the "post-" string? The 'post/pre' string seems to be important to the event still?
Thanks, James