[PATCH 2 of 3] move store walking from streamclone.py into store.py

Adrian Buehlmann adrian at cadifra.com
Wed Jul 23 14:03:23 CDT 2008


On 23.07.2008 17:51, Paul Moore wrote:
> 2008/7/23 Adrian Buehlmann <adrian at cadifra.com>:
>>> Can you summarise the discussion over patch 3, describing your
>>> approach and the issues Matt has with it? If you can do this, I'll
>>> look at whether it is something I would be able to pick up. I'd like
>>> to see the problem resolved, but I'm afraid I don't have time to go
>>> back through the various emails trying to pick out which issues have
>>> been addressed and which are outstanding.
>> No, I sure won't repeat here that discussion and summarising
>> is really the wrong thing.
> 
> OK, well I probably won't be able to pick up this work then.

I'm sorry if I sounded a bit harsh here and I sure would be glad for
any help you can offer and also would appreciate your work
and comments on this, but the devil is really in the detail. I really see
no point in trying to simplify matters. Matt and I are in disagreement,
I can't summarise more than that.

I also currently don't want to restart here the discussion
we've already held. I'm sorry, but I think you should at least read
very closely all of Matt's arguments to learn what he wants.

I think it might currently be a good idea to let this calm down
for a while und just concentrate on saving the things where we
have reached consensus, namely patch 1 and 2 alone.

Maybe we can later extend that a little bit more towards solving
the real goals, without bringing in the full patch 3 thing.

I have some ideas and questions about Matt's views and goals for verify,
but please let's not tack that into this thread here.

> Matt - in case you are interested in seeing this implemented, can you
> summarise the current status of patch 3 in a way that might allow me
> to pick it up?
> 
>> Do whatever you like. I'm currently only interested to save the
>> work I did for patch 1 and 2, as that part seems to be ok for
>> Matt.
> 
> Fine. Are you at least willing to post your final version of patch 3,
> so that it's available for reference?

I have already done so, the very last state I have for the patch series
is:

intro: http://selenic.com/pipermail/mercurial-devel/2008-July/007223.html
patch 1: http://selenic.com/pipermail/mercurial-devel/2008-July/007224.html
patch 2: http://selenic.com/pipermail/mercurial-devel/2008-July/007225.html
patch 3: http://selenic.com/pipermail/mercurial-devel/2008-July/007227.html

The state before, which has seen some extensive testing is:

intro: http://selenic.com/pipermail/mercurial-devel/2008-July/007204.html
patch 1: http://selenic.com/pipermail/mercurial-devel/2008-July/007206.html
patch 2: http://selenic.com/pipermail/mercurial-devel/2008-July/007205.html
patch 3: http://selenic.com/pipermail/mercurial-devel/2008-July/007207.html

Also helpful is to see the threaded archive:
http://selenic.com/pipermail/mercurial-devel/2008-July/thread.html#7204
(July 2008 Archives by thread)

I'm going to do some minor work now on patch 2 as per request of Dirkjan,
in order to get patch 1 and 2 finished independently of patch 3. But if you
want, I can revise patch 3 to make sure it will still work on top of that.
But Matt has already said that he doesn't want it, so you would have to
do a fresh start for patch 3 yourself anyway, after having understood all
of Matt's arguments. Of course you could reuse some things of patch 3
like the encoding functions. But the rest is pretty much null and void
per what Matt has said.

If you have any specific request or question, just ask, maybe also off list.
But just don't expect that I will be able to simplify things here.


More information about the Mercurial-devel mailing list