[darcs-users] darcs weekly news #28

Eric Kow kowey at darcs.net
Wed May 13 01:46:01 UTC 2009


Trent said:
>> What can we do to make Darcs 2 (more) attractive to the GHC team, now
>> that we (presumably) have a chance of winning their patronage again?

I was a little bit worried earlier this week that I was not doing more
to seize the day on this (strike while the iron is hot).  I'm still a
bit worried :-)

Thanks to Trent for asking these kinds of questions.
 
Then Simon said:
> Resolve these open issues created or commented by GHC folk (add others if you know them):
>
> Simon Peyton-Jones (8):  
> http://bugs.darcs.net/issue?%40search_text=simonpj&title=&%40columns=title&topic=&id=&%40columns=id&%40sort=id&creation=&%40columns=creation&creator=&%40columns=creator&activity=&%40columns=activity&actor=&nosy=&priority=&%40group=priority&status=-1%2C1%2C2%2C3%2C4%2C5%2C6%2C16&%40columns=status&assignedto=&%40columns=assignedto&%40pagesize=50&%40startwith=0&%40sortdir=on&%40queryname=&%40old-queryname=&%40action=search
>
> Simon Marlow (3):
> http://bugs.darcs.net/issue?%40search_text=simonmar&title=&%40columns=title&topic=&id=&%40columns=id&%40sort=id&creation=&%40columns=creation&creator=&%40columns=creator&activity=&%40columns=activity&actor=&nosy=&priority=&%40group=priority&status=-1%2C1%2C2%2C3%2C4%2C5%2C6%2C16&%40columns=status&assignedto=&%40columns=assignedto&%40pagesize=50&%40startwith=0&%40sortdir=on&%40queryname=&%40old-queryname=&%40action=search
>
> Ian Lynagh (7):
> http://bugs.darcs.net/issue?%40search_text=igloo&title=&%40columns=title&topic=&id=&%40columns=id&%40sort=id&creation=&%40columns=creation&creator=&%40columns=creator&activity=&%40columns=activity&actor=&nosy=&priority=&%40group=priority&status=-1%2C1%2C2%2C3%2C4%2C5%2C6%2C16&%40columns=status&assignedto=&%40columns=assignedto&%40pagesize=50&%40startwith=0&%40sortdir=on&%40queryname=&%40old-queryname=&%40action=search

Some of these are quite long term.

Let's step back a little bit.  We know we have major lines of work to
do: darcs-3 and basic day-to-day performance issues.  These will take
quite a bit of time, although if we do things right, and are lucky, we
can roll out the file cache stuff and Petr's hashed-storage work within
a year.

In the meantime, it would make sense for us to adhere to our position
that darcs is not recommended for large repositories.

Nevertheless, if the GHC team are going to postpone their switch, what
can we do to make their current darcs experience more enjoyable?
Perhaps good enough to make it worth their while to upgrade to darcs-3
when it comes out?  It may take a lot of effort for them to upgrade to
darcs-2 format repositories, so maybe we could just focus on getting a
darcs-2 client on the darcs.haskell.org server and performance-testing
their ssh stuff to see if it makes any difference?  Maybe we could get
them to use hashed format repositories with --lazy so that they don't
have to ship tarballs around?  Maybe we could audit their darcs-all
script and see where things could go wrong?  We should keep in touch
with them.  So darcs-users: what can we realistically afford to offer
to do for them now with our resources?  It may be good if somebody would
name himself the GHC liaison and work on this kind of stuff :-)

-- 
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 194 bytes
Desc: not available
URL: <http://lists.osuosl.org/pipermail/darcs-users/attachments/20090512/31cc2563/attachment-0001.pgp>


More information about the darcs-users mailing list