Mike Kupfer <mike.kupfer at sun.com> writes:

>>>>>> "Ken" == Ken Erickson <ken.erickson at sun.com> writes:
>
> Ken> OK - so I have sort of asked this already - maybe I was unclear 
>
> Sorry, I hadn't realized we needed to provide more input.  Too much
> confusion/fallout from Steve leaving, I guess.  Anyway...
>
> Ken> Will my modified pbconfirm get email from mercurial?  
>
> No.  Or at least, if you want pbconfirm to be mail-driven, you'll have
> to write the code to generate the email.

It's worth noting here that the only mechanism TeamWare provided for
hooks is via the various kinds of mail notification.  The existing
tools trigger via mail messages because that's the only option they
have, not necessarily because that's the right thing, or a better
idea, or anything of the sort.  You'd have to evaluate whether that's
really a good thing to do on a case-by-case basis.

> What you get from Mercurial is the ability to specify a command
> (probably a shell script) or a Python function that will be run when a
> particular event happens.  The Mercurial book
> (http://hgbook.red-bean.com/hgbook.html) has a good chapter on hooks
> (how to use them, how to write them).  
>
> Ken> What I was hoping for is a minimally functional mercurial workspace
> Ken> that has been set up the way people have decided that the SFW and
> Ken> ON workspaces are going to be set up.
>
> I think that's a good idea.  I don't know of any such workspace, though
> I would expect that the people who worked on some of the checker hooks
> (see below) would have had to set up something like that.  

That's what we were going to use tijuana for.

> Anyone else on the team want to chime in here?

The hooks are as yet not existant in the public code.  Steve tested
some out on the gate-side via a sfwnv tree he could make junk changes
to, I tested the stuff on our side out via a clone of onnv I could
make junk changes to.

The specifics of what allowed you'd need to discuss with the rest of
us.  The specifics of how the gate would behave you'd need to figure
out, or discuss with the gks.

> Ken> Or, is it my job, as "script monkey", to also invent the interfaces
> Ken> that the scripts will use?
>
> The event hook interface is defined for both commands (i.e., what
> environment variables are passed in) and Python.  Some hooks have
> already been written, especially the conformance check hooks (copyright,
> hdrchk, etc).  They're in the project gate under usr/src/tools/onbld.
>
> But beyond that, you get to decide how to get from the event hook to the
> end functionality.  If you get stuck, holler.

Yes.

> Ken> I'd just like to start making some progress here, but so far I
> Ken> still know almost nothing about how the "open" gates will operate.
> Ken> Does anybody know?

> I don't know of any write-up.  I've got a vague understanding, but
> probably not as detailed as Rich's or Steve's.  Let me see what
> information I can turn up, and I'll try putting it on the genunix wiki.

Ok.

-- Rich

Reply via email to