Mike Kupfer wrote:
> stevel> So... do you guys think I should include mention of driver/modinfo 
> stevel> strings?
> 
> Not if there are already bugs for them that are marked with hg_trans.
> If we're going to call out specific subsets of the hg_trans bugs, I
> would focus on the ones that are likely to cause incorrect behavior,
> such as the class action script bugs.
> 
> stevel> Anything else I missed?
> 
> I'd mention what alias people should email if they have questions.  I
> guess that should be scm-migration-dev...?
> 
> In fact, if we can give a little guidance on "replacements", that might
> be helpful (and cut down on the amount of time we spend answering
> questions).  So, maybe the heads-up mail can include some text like
> 
>       If the module version can be determined by other means (e.g.,
>       showrev(1) output), you may simply wish to delete the keywords.
> 
> What do you think?

Sounds good, here is my new heads up:
---------------------
This is a heads up to let you all know of an important impending change
as we switch from Teamware to Mercurial.  Currently, onnv-gate has
multiple occurrences of user-visible SCCS keywords (%I%, %E%, etc.).
We've flagged and filed bugs for the ones that we've found[1].
Unfortunately, we (the SCM Migration project team) aren't experts in
that code, and can't know whether or not they can safely be replaced or
removed; what we *do* know is that when we switch to Mercurial, those
keywords will be visible in _unexpanded form_, so:

$ genmsg -v
genmsg version %I%

This is teh sux0r.  If you are responsible for any code that has user
visible SCCS keywords, or consume any such keywords (or have a
dependency in some form or another on this output), you should make
every effort to fix this code to either eliminate or replace the use of
SCCS keywords (and no, Mercurial does not have an equivalent keyword
feature).  If they aren't replaced, they will be carried through and
left unexpanded when we switch to Mercurial.

If the module or binary version can be determined by other means (e.g.:
showrev(1) output), you may simply wish to delete the keywords.

When a date is chosen for transitioning onnv-gate to Mercurial, we will
let you know so nobody is surprised.  We don't have an exact date yet,
but expect it sometime in the September-October timeframe.

If you have any questions, please email scm-migration-dev at opensolaris.org
If you have questions regarding source in usr/closed, please email
tonic-iteam at sun.com.

[1] flagged with the 'hg_trans' keyword (along with other bugs)
http://monaco.sfbay/list.jsf?product=solaris&keyword=hg_trans
(If you find other occurrences we haven't filed bugs for, please attach
the 'hg_trans' keyword to the bug you file to track it)
------------------------

> I'm assuming that once you've sent the heads-up mail, we should update
> the "user-visible" CRs that Rich filed, to indicate that we're not
> working on them.  This would include
> 
> - removing the oss-sponsor keyword (if set)
> - removing me as RE
> - changing RM from Bonnie back to the original RM
> - attaching the diffs that Rich sent out a few weeks ago
> 
> Let me know if you disagree.

Sounds like a good idea to me.

cheers,
steve

-- 
stephen lau // stevel at sun.com | 650.786.0845 | http://whacked.net
opensolaris // solaris kernel development

Reply via email to