| Time |
S |
Nick |
Message |
| 00:29 |
|
|
ertai joined #darcs |
| 05:14 |
|
|
kpreid_ joined #darcs |
| 05:26 |
|
|
gwern joined #darcs |
| 06:41 |
|
|
adu joined #darcs |
| 07:42 |
|
|
lelit joined #darcs |
| 08:23 |
|
|
balor joined #darcs |
| 08:28 |
|
|
tux_rocker_ joined #darcs |
| 08:32 |
|
|
mamalujo joined #darcs |
| 08:53 |
|
|
sm_ joined #darcs |
| 08:54 |
|
mamalujo |
hello, Ive been trying to use darcs w etckeeper on my debian (squeeze, w a little bit of sid). unfortunately, this resulted in my apt freezing on the step where its to log the changes in VCS. it seems to freeze on 'Removing from pending' darcs version is 2.2.0-1, etckeeper is 0.37 . I didnt specify it should use darcs during installation, but changed /etc/etckeeper/etckeeper.conf after (and did etckeeper init), so i still have a .git ... |
| 08:55 |
|
mamalujo |
... (the default VCS) there. I believe I also made a darcs initialise manually some time before on /etc, but never did anything with it, before hearing about this neat script (ie etckeeper) Should probably also note that I dont know how to use darcs yet, thought to make this usage a motivation to learn, just hoped I wouldnt come into as major problem as a crashed apt so immediately :) so, am I stumbling into some familiar problem or ... |
| 08:55 |
|
mamalujo |
... misuse, or is this something more obscure? |
| 08:58 |
|
|
kolmodin joined #darcs |
| 09:00 |
|
mornfall |
mamalujo: Are you sure it's darcs that's hanging? Can you just run a few commands in your /etc? (darcs changes, darcs wh) |
| 09:05 |
|
mamalujo |
I cant say I am. the former command says nothing at all. the latter is spewed huge amounts of data after a short pause. |
| 09:08 |
|
mornfall |
Well, I guess darcs itself is working just fine. |
| 09:09 |
|
mornfall |
Although you may want to rm -rf _darcs and go with the etckeeper manual. |
| 09:09 |
|
mornfall |
s/go/start over/ or something |
| 09:12 |
|
mamalujo |
k, thx for the advice |
| 09:20 |
|
|
mamalujo left #darcs |
| 09:29 |
|
|
nomeata joined #darcs |
| 10:01 |
|
gh_ |
mornfall: darcs 2.3 beta 1 is often reporting that a file does not exist when I do a "darcs whatsnew" on it |
| 10:01 |
|
gh_ |
but still it does the whatsnew |
| 10:08 |
|
mornfall |
gh_: Oh? |
| 10:08 |
|
mornfall |
Can you paste example output? |
| 10:16 |
|
mornfall |
gh_: ? |
| 10:18 |
|
gh_ |
yep sorry |
| 10:18 |
|
gh_ |
pc was slow |
| 10:19 |
|
gh_ |
$ darcs w src/HTab/Formula.hs |
| 10:19 |
|
gh_ |
WARNING: File 'src/HTab/Formula.hs' does not exist! |
| 10:19 |
|
gh_ |
What's new in "src/HTab/Formula.hs": |
| 10:19 |
|
gh_ |
No changes! |
| 10:19 |
|
mornfall |
And the file is there, right? |
| 10:19 |
|
gh_ |
yes |
| 10:19 |
|
gh_ |
it does not say it for every file: |
| 10:19 |
|
mornfall |
That's curious. |
| 10:19 |
|
gh_ |
$ darcs w src/htab.hs |
| 10:19 |
|
gh_ |
What's new in "src/htab.hs": |
| 10:19 |
|
gh_ |
No changes! |
| 10:20 |
|
mornfall |
Could it match a boring regexp somehow? |
| 10:20 |
|
mornfall |
That is probably a bug (maybe a different one): we now say does not exist if the file is merely boring, too. |
| 10:20 |
|
mornfall |
Hm. |
| 10:20 |
|
gh_ |
hmm nope |
| 10:21 |
|
mornfall |
Although I don't see how it could be boring, either. |
| 10:21 |
|
mornfall |
Can you pack up the repo and publish? |
| 10:21 |
|
mornfall |
Also, can you, in ghci run: |
| 10:21 |
|
gh_ |
it seems to only affect files that are 2 levels deep |
| 10:21 |
|
mornfall |
:m Storage.Hashed |
| 10:21 |
|
gh_ |
2 and more |
| 10:22 |
|
mornfall |
:m + Storage.Hashed.Tree |
| 10:22 |
|
mornfall |
t <- expand =<< readPlainTree "." |
| 10:22 |
|
mornfall |
printPath t "src/HTab" |
| 10:22 |
|
mornfall |
? |
| 10:22 |
|
gh_ |
k |
| 10:23 |
|
mornfall |
gh_: Nevermind, reproduced. |
| 10:23 |
|
mornfall |
It always happens with 2 or deeper. Bug in hashed-storage, presumably. |
| 10:23 |
|
gh_ |
ah ok |
| 10:23 |
|
gh_ |
good thing we catch it now |
| 10:23 |
|
mornfall |
Yeah. That's why we do betas. :) |
| 10:24 |
|
mornfall |
Thanks for the report, too. |
| 10:25 |
|
gh_ |
you're welcome :) |
| 10:43 |
|
mornfall |
gh_: Fixed. Stupid typo it was... |
| 10:49 |
|
gh_ |
cool |
| 10:52 |
|
|
Jerri joined #darcs |
| 11:08 |
|
|
mamalujo joined #darcs |
| 11:26 |
|
mamalujo |
hm, removing /etc/_darcs and starting seemed to have worked, though it took such a long time that I considered it a hang and gave up on it - and since the previous try, everything was duplicated cuz i didnt delete residual .git, perhaps there wasnt any problem to begin with. but, I do hope apt will not continue to be that drastically slow on further usage.. |
| 11:28 |
|
mornfall |
I have no idea what etckeeper is doing, so I can't help you much. |
| 11:56 |
|
|
nomeata joined #darcs |
| 12:16 |
|
|
kpreid_ joined #darcs |
| 12:36 |
|
mornfall |
(wiki is down again) |
| 12:36 |
|
|
tux_rocker__ joined #darcs |
| 12:53 |
|
|
arjanb joined #darcs |
| 13:27 |
|
* tux_rocker__ |
realizes he doesn't like to work on darcs unless he's caffeinated |
| 13:36 |
|
|
intripoon joined #darcs |
| 14:05 |
|
Riastradh |
I can't recall whether I've asked about this before, but why does `darcs get http://...' spray what looks like wget output for every patch it downloads, for some repositories? (I don't know what repositories.) This is under Darcs 2.0.2 (ought I just to upgrade?). |
| 14:16 |
|
tux_rocker__ |
Riastradh: sometimes darcs actually uses wget to fetch those patches |
| 15:53 |
|
|
Jerri joined #darcs |
| 16:09 |
|
Riastradh |
Yes, tux_rocker__, but why does it spew the output of wget? |
| 16:44 |
|
|
klapaucjusz joined #darcs |
| 17:21 |
|
|
sm joined #darcs |
| 17:52 |
|
|
[1]intripoon joined #darcs |
| 19:17 |
|
|
kolmodin joined #darcs |
| 19:44 |
|
|
dcoutts__ joined #darcs |
| 19:52 |
|
mornfall |
Hmm. Lots of code for the endianity conversion business. |
| 19:52 |
|
mornfall |
Maybe I should outsource part of it to some other module (but, which). |
| 20:26 |
|
lispy |
mornfall: do you use Data.Binary? |
| 20:31 |
|
mornfall |
lispy: No, it's too slow. |
| 20:32 |
|
mornfall |
I use Foreign.Storable. |
| 20:33 |
|
mornfall |
(And I implemented linux-C-style xlate32/xlate64...) |
| 20:35 |
|
mornfall |
Btw. there's now http://repos.mornfall.net/hashed-storage/trac/ -- see "timeline" and "browse source" : - ) |
| 20:35 |
|
lispy |
Data.Binary is too slow? I find that surprising only because of all the sucess stories I've seen on haskell-cafe about it having good performance |
| 20:37 |
|
mornfall |
It is great, but it can't beat Foreign.Storable, only if because it has to copy the data (unlike Foreign.Storable). |
| 20:37 |
|
mornfall |
I need to access the bits inline in the mmap area. |
| 20:40 |
|
mornfall |
Don't worry, I originally used Data.Binary and profiled the thing. I gained nontrivial speedup from switching. |
| 20:41 |
|
Heffalump |
the reason being that you can modify the data inplace? |
| 20:41 |
|
mornfall |
Basically, yes. But reading is also faster this way. |
| 20:42 |
|
Heffalump |
do you know why that is? |
| 20:42 |
|
mornfall |
Well, I don't really know -- but even reading the index involves updating the timestamps and hashes in it, and using Data.Binary would mean rewriting the whole 7 megabyte binary blob. |
| 20:42 |
|
Heffalump |
ah, right |
| 20:42 |
|
mornfall |
So even if reading was on par, that wouldn't help. |
| 20:42 |
|
Heffalump |
so it's probably inplace-ness again |
| 20:42 |
|
Heffalump |
ah, ok |
| 20:44 |
|
mornfall |
(Well, purely theoretically, I could try doing partial Data.Binary based updates and then poke the parts to be updated in-place. I am not sure it would lead to shorter or better code though. Hmm.) |
| 20:46 |
|
mornfall |
However, with the xlate stuff, the common case gets no runtime penalty in the common case (little endian system). From looking at Data.Binary, my impression is, that even with getWord64le on a le system, there's a penalty, unless the compiler is super-smart... |
| 20:47 |
|
mornfall |
(Hmm, that's what you get for not reading before hitting enter. Whatever.) |
| 20:49 |
|
mornfall |
Anyway, I'm quite happy about today's progress. |
| 20:49 |
|
mornfall |
(Just that I have been neglecting Russian and piano. Bahmpf. : - P) |
| 20:58 |
|
|
fophillips joined #darcs |
| 21:58 |
|
|
ertai_ joined #darcs |
| 22:18 |
|
|
kpreid_ joined #darcs |
