* Richard Lowe <[EMAIL PROTECTED]> [2006-11-19 04:54]:
> Stephen Lau wrote:
> >Now that I've (hopefully) dug myself out of this onnv-gate Hg
> >regeneration hole, I'd like to start working on this SCM Migration thing.
> >
> >I believe these are our current tasks:
> >1) Finalise a priority list of developer/GK tools to convert
> > (This discussion has already started it looks like)
(Steve: please get your document into an opensolaris.org repository,
so that you can accept these as patches...)
> o add-gateling
> Would probably be best ignoring anonymous bringover, with 'real'
> users could map user->email via the webapp?
My thought is that this operation should be part of being added as a
committer, and that the anonymous aspect should be dropped.
> o backout
> Should just be 'hg backout' (perhaps wrapped to generate an
> appropriate message?)
+1.
> o buglist
> hg log -r${LAST_BUILD}:${THIS_BUILD} --template '{desc|strip}'
> Filter it just for IDs/case numbers if appropriate.
> (backouts?)
Sounds fine.
> o daily_update
> remove.
> o lock-gate/unlock-gate
> Lock via a hook rather than using mercurial's locking? should make it
> possible to exempt gate staff from the lock.
This feature is already integrated in the SCM management on the site:
project leaders can activate a lock that blocks operations not
initiated by project leaders.
> o pbcheck
> Would be better if this and wx pbchk ran exactly the same checks,
> exactly the same way (sharing implementation).
> Should refuse to take an obviously bogus putback, should probably
> only notify on others (various files in ON don't pass the checks in
> their natural state, especially true for cstyle).
Danek also mentioned making the putback checks run in the local commit
hook. This means there's a workspace preparation command, to install
the ON client-side hooks, in addition to running these on the incoming
or other hook on the server side.
> RTI check is hard, assuming os.o can't make use of webrticli.
> Bug ID check (as in wx, assumed wanted here) is also hard, since that
> uses monaco.
RTI for development release will have to be adjusted/reimplemented to
run on opensolaris.org...
> o pbconfirm
> Simple hook, assuming we can map user->email via the webapp backend
Yes, the hooks already have this function. (Is this the "your putback
will appear in" message? Is it really accurate for the non-SWAN
contributor?)
- Stephen
--
Stephen Hahn, PhD Solaris Kernel Development, Sun Microsystems
[EMAIL PROTECTED] http://blogs.sun.com/sch/
_______________________________________________
tools-discuss mailing list
[email protected]
