Re: [Gnu-arch-users] Patch : precomimt hook

[Mark A. Flacy]

>>>>> "Jan" == Jan Hudec <address@hidden> writes:
Jan> 
Jan> I find the term "changeset is generated" a bit confusing for this. So
Jan> to be clear, it has to be called BEFORE actual diffing takes place (so
Jan> it has a chance to touch the files) and it might be reasonable to call
Jan> it after list of changes is computed.
Jan> 
Jan> It's not however always reasonable. What if the hook wanted to change
Jan> some file besides those already changed (eg. some kind of changelog).

Well, if you have access to the changeset then you can modify the changeset
any way that you wish.  

I believe that I'm thinking of an entirely different set of use cases that
you are.  I think it's fairly useless to call the hook prior to computing
the changeset.  You could add *THAT* functionality in a script that invokes
"tla commit" itself rather than adding a hook, could you not?

-- 
 Mark A. Flacy
 Any opinions expressed above are my own.  Any facts expressed above
 would imply that I know what I'm writing about.  Sometimes, I do!
 "We're in the hands of lunatics."
 "Of financiers.  Far worse." -- C.J. Cherryh, "Hellburner"