Is this right?
Vincent Ladeuil
v.ladeuil+lp at free.fr
Wed Nov 5 07:35:35 GMT 2008
>>>>> "Stephen" == Stephen J Turnbull <stephen at xemacs.org> writes:
Stephen> Guillermo Gonzalez writes:
>> On Tue, Nov 4, 2008 at 8:55 PM, Russel Winder
>> > To be honest all that is needed is a small announcement email on the
>> > email list and a note on the Lauchpad pages. The issue is public
>> > information presented in places where people can find it easily.
Stephen> I'm sorry, but that's 1995 thinking. In 2008, the
Stephen> issue is not where "people can find it easily"---as
Stephen> you point out, people generally won't go farther
Stephen> than "trivially"---it is where "machines can follow
Stephen> it automatically". We know how to do this.
You're talking about the other side of the problem there, Russel
was talking about the subscribers of a branch, you're talking
about the publisher.
On that side there are indeed several means to address the
problem but none can be fully automated, the publisher *has* to
do *something*.
Stephen> (1) HTTP redirect.
Already supported (modulo some known bugs I'm aware of and plan
to address ASAP (for medium values of soon)).
Stephen> (2) Any URL schema private to bzr could grow a
Stephen> redirect status.
Well, lp itself use the 'location' file trick, I don't remember
the details nor the policy about support for that in bzr but it's
protocol neutral.
Stephen> (3) repo formats could grow a redirect
Stephen> configuration. (2) could probably be implemented
Stephen> this way, but for robustness (the case when a repo
Stephen> gets completely removed) there should be a net
Stephen> protocol that could be implemented by a lightweight
Stephen> server on the well-known port.
YAGNI. You still need the publisher to *do* something.
A cheap trick than can be used *today* is to make a commit with a
comment containing the new location.
But note that this doesn't address the case we're talking about
because the old branch just doesn't exist anymore so there is *no
place* where you can put a redirection mechanism in place.
<snip/>
>> I'll [send out an announcement] ASAP
>> If you think that announcement should be in other places, please let
>> me know and I'll post there too.
Stephen> Of course as long as the technology is broken, this
Stephen> is indeed the courteous thing to do.
Well, even when the branch still exists if the publisher does
nothing, subscribers stay in the dark... so it's not a
technological problem only.
Vincent
More information about the bazaar
mailing list