Created on 2009-08-28.16:13:33 by kowey, last changed 2010-06-15.21:30:48 by admin.
msg8573 (view) |
Author: kowey |
Date: 2009-08-28.16:13:22 |
|
Darcs-2.3 converts to --hashed by default which is slow.
When Darcs is fetching from old-fashioned, it ought to say that this is possible.
|
msg8576 (view) |
Author: kowey |
Date: 2009-08-28.18:01:47 |
|
Hmm, Duncan pointed out that msg8573 wasn't very clear.
1. I think darcs get of a hashed repo will be faster than darcs get
old-fashioned with the current Darcs 2.3 behaviour of converting it to hashed.
At least, if the server side was hashed we'd be able to use --lazy plus the caches
2. Darcs should notice that it's in the middle of performing such a conversion
and say something to the effect of "if you could persuade the maintainer of
REMOTE-REPO-PATH to darcs optimize --hashed, darcs get will be faster"
It's best if issue1584 was implemented so that the message in #2 can be very
simple, but you can work on this independently of it
|
msg8578 (view) |
Author: twb |
Date: 2009-08-29.01:24:12 |
|
On Fri, Aug 28, 2009 at 06:13:22AM +0000, Eric Kow wrote:
> Darcs-2.3 converts to --hashed by default which is slow.
This was actioned due to issue1435, and (indirectly) particularly to
avoid data loss as in issue1453. It also means stuff like inter-repo
caching (~/.darcs/cache) work.
|
msg8579 (view) |
Author: duncan |
Date: 2009-08-29.02:01:40 |
|
On Sat, 2009-08-29 at 01:24 +0000, Trent W. Buck wrote:
> Trent W. Buck <trentbuck@gmail.com> added the comment:
>
> On Fri, Aug 28, 2009 at 06:13:22AM +0000, Eric Kow wrote:
> > Darcs-2.3 converts to --hashed by default which is slow.
>
> This was actioned due to issue1435, and (indirectly) particularly to
> avoid data loss as in issue1453. It also means stuff like inter-repo
> caching (~/.darcs/cache) work.
Note that I was never complaining that 2.3 defaults to --hashed. On the
contrary, it's clearly the sensible thing, even if the initial
conversion is a tad slower.
The issue is how to encourage people to migrate and to make it easy for
them to do so. This is more important now that 2.3 makes the hashed
format really quick, and the traditional format really slow.
|
msg8912 (view) |
Author: kowey |
Date: 2009-10-06.08:48:18 |
|
My patch has made it into the darcs repo. We need a volunteer to write a test case:
Quoting Jason:
local repo supports hashed, remote supports hashed
local repo supports hashed, remote does not support hashed
local repo does not support hashed, remote supports hashed
local repo does not support hashed, remote does not support hashed
|
msg9285 (view) |
Author: kowey |
Date: 2009-11-15.08:48:13 |
|
I claim that it's really not worthwhile to write a test for this.
Please kick me if I'm mistaken...
|
|
Date |
User |
Action |
Args |
2009-08-28 16:13:34 | kowey | create | |
2009-08-28 18:01:57 | kowey | set | nosy:
kowey, darcs-devel, duncan.coutts, jaredj, dmitry.kurochkin superseder:
+ optimize --upgrade messages:
+ msg8576 |
2009-08-29 01:24:15 | twb | set | nosy:
+ twb messages:
+ msg8578 |
2009-08-29 01:40:22 | twb | set | nosy:
- twb |
2009-08-29 02:01:43 | duncan.coutts | set | nosy:
+ twb messages:
+ msg8579 |
2009-09-15 22:42:52 | kowey | set | status: needs-implementation -> has-patch nosy:
kowey, darcs-devel, duncan.coutts, twb, jaredj, dmitry.kurochkin assignedto: kowey |
2009-10-06 08:48:27 | kowey | set | status: has-patch -> needs-reproduction nosy:
kowey, darcs-devel, duncan.coutts, twb, jaredj, dmitry.kurochkin messages:
+ msg8912 assignedto: kowey -> |
2009-10-24 00:31:50 | admin | set | nosy:
+ duncan, - duncan.coutts |
2009-11-15 08:48:21 | kowey | set | status: needs-reproduction -> resolved messages:
+ msg9285 |
2010-06-15 21:30:46 | admin | set | milestone: 2.4.x |
2010-06-15 21:30:48 | admin | set | topic:
- Target-2.4 |
|