| Time |
S |
Nick |
Message |
| 00:25 |
|
|
mxc joined #darcs |
| 00:29 |
|
|
zooko joined #darcs |
| 00:31 |
|
|
twb joined #darcs |
| 01:47 |
|
|
lisppaste2 joined #darcs |
| 02:18 |
|
|
zooko joined #darcs |
| 02:24 |
|
juhp |
manpage seems to get installed rw------- |
| 02:24 |
|
juhp |
? |
| 02:24 |
|
juhp |
in 2.4 |
| 02:25 |
|
twb |
AFAIK there's no particular effort to specify permissions |
| 02:25 |
|
twb |
So I would guess it's because of the umask or because you did/didn't run it as root |
| 02:39 |
|
|
kpreid_ joined #darcs |
| 05:29 |
|
|
mxc joined #darcs |
| 05:34 |
|
|
lispy|web joined #darcs |
| 05:47 |
|
mxc |
sorry to keep bugging, but wondering if anyone has any idea where this error would come from: |
| 05:47 |
|
mxc |
No sources from which to fetch file `0000001654-8a5845b75d8a10fd4e3920158e528ef58325572baec1b33fb0f6ea7d48e50c27' |
| 05:47 |
|
lambdabot |
mxc: You have 1 new message. '/msg lambdabot @messages' to read it. |
| 05:47 |
|
mxc |
it comes when I run this, but isn't specific to that repo: |
| 05:47 |
|
mxc |
darcs get --lazy --debug http://code.haskell.org/yi |
| 05:48 |
|
lispy|web |
mxc: what version of darcs? |
| 05:49 |
|
mxc |
2.4 |
| 05:49 |
|
mxc |
but get the same thign with 2.3 |
| 05:49 |
|
lispy|web |
mxc: Hmm... |
| 05:49 |
|
mxc |
if you ahven't seen it, dont worry about it |
| 05:49 |
|
mxc |
i hate to suck up even more of people's time, but was wondering if someone had soem magic bullet |
| 05:50 |
|
lispy|web |
mxc: I think recently there were some other 2.4 bugs reported with files that were existent reporting as missing. I think one fix may have been to upgrade the mmap library and rebuild darcs? (this hasn't been confirmed as a fix yet) |
| 05:50 |
|
twb |
mxc: move your cache out of the way and try again |
| 05:50 |
|
twb |
(Just a guess) |
| 05:50 |
|
twb |
lispy|web: you're referring to http://bugs.darcs.net/patch169 |
| 05:51 |
|
lispy|web |
twb: yeah, that looks familiar |
| 05:52 |
|
lispy|web |
and the linked bug http://bugs.darcs.net/issue1753 |
| 06:00 |
|
twb |
mxc: fwiw, I cannot reproduce that issue with Darcs 2.4 built against mmap 0.5 |
| 06:03 |
|
mxc |
thanks anyway |
| 07:05 |
|
mornfall |
mxc: Have you tried building without curl? |
| 07:06 |
|
mornfall |
(cabal install darcs -f-curl [--reinstall]) |
| 07:08 |
|
twb |
I haven't tried that, FWIW |
| 07:09 |
|
mornfall |
twb: Wrt. issue1753? |
| 07:09 |
|
mxc |
trying now |
| 07:09 |
|
twb |
mornfall: yes |
| 07:09 |
|
mornfall |
The problem mxc has and issue1753 are quite unrelated. |
| 07:09 |
|
twb |
mornfall: or indeed anything |
| 07:10 |
|
mornfall |
twb: And -f-curl won't help you with the index issue. |
| 07:15 |
|
twb |
Okey dokey |
| 07:22 |
|
mxc |
rebuilding w/out curl didn't help |
| 07:22 |
|
mxc |
still get this: |
| 07:22 |
|
mxc |
Fetching over HTTP: http://code.haskell.org/yi/_da[…]fb0f6ea7d48e50c27 |
| 07:22 |
|
mxc |
No sources from which to fetch file `0000001654-8a5845b75d8a10fd4e3920158e528ef58325572baec1b33fb0f6ea7d48e50c27' |
| 07:22 |
|
mxc |
URL.waitNextUrl succeeded: http://code.haskell.org/yi/_da[…]fb0f6ea7d48e50c27 ./darcs23376 |
| 07:22 |
|
twb |
I can wget that |
| 07:23 |
|
mxc |
i can too |
| 07:23 |
|
mxc |
curl even |
| 07:23 |
|
twb |
Are you on Debian? |
| 07:23 |
|
mxc |
osx 10.6 |
| 07:23 |
|
twb |
Do you use an HTTP proxy? |
| 07:23 |
|
mxc |
no |
| 07:56 |
|
|
lelit joined #darcs |
| 07:59 |
|
|
npouillard joined #darcs |
| 08:00 |
|
|
npouillard joined #darcs |
| 08:02 |
|
|
darcscommitbot joined #darcs |
| 08:03 |
|
|
darcswikibot joined #darcs |
| 08:04 |
|
iaefai |
If I have completed a certain line of development and really only have an interest in keeping it for historical uses, is there a way to mark it as such? |
| 08:07 |
|
npouillard |
iaefai: is this branche merged into the main branch ? |
| 08:07 |
|
mornfall |
iaefai: You could stick a tag on it saying something witty. |
| 08:08 |
|
iaefai |
ok, I just did the tag thing |
| 08:08 |
|
iaefai |
Now does it refer to the last record or the one coming up? |
| 08:08 |
|
mornfall |
iaefai: Maybe edit the MOTD. |
| 10:32 |
|
|
gal_bolle joined #darcs |
| 10:38 |
|
|
kowey joined #darcs |
| 10:39 |
|
kowey |
morning |
| 10:43 |
|
kowey |
I would like to see if patch169 helps twb, even if it shouldn't by rights |
| 10:44 |
|
kowey |
(although there, you'd also want to make sure you're using hashed-storage 0.4.7 with mmap 0.4 as well) |
| 10:45 |
|
twb |
I have to jump through some hoops to get mmap 0.4, which I haven't bothered to do yet |
| 10:46 |
|
kowey |
same for derrida |
| 10:47 |
|
kowey |
(err, same curiosity, not same hoops) |
| 10:59 |
|
|
mxc joined #darcs |
| 11:15 |
|
|
mxc joined #darcs |
| 12:43 |
|
|
gh_ joined #darcs |
| 12:53 |
|
Korusef |
where is the documentation of the expression for --matches parameter ? |
| 12:58 |
|
kowey |
darcs help pattern |
| 12:59 |
|
kowey |
err, patterns |
| 13:02 |
|
Korusef |
ahh nice, was reading that text from sourcecode :) |
| 13:04 |
|
twb |
Pretty suboptimal, but better than nothing |
| 13:07 |
|
kowey |
there's no spot in the manual where it automagically appears? |
| 13:08 |
|
twb |
It's magicified into the manpage |
| 13:08 |
|
twb |
But it really deserves a good two- or three-page howto document, with lots of discussion and "real world" examples. |
| 13:09 |
|
twb |
e.g. using --match 'not name ^DRAFT:' in your .darcs/defaults |
| 13:10 |
|
kowey |
nod |
| 13:10 |
|
kowey |
I hope the random bits and pieces of documentation (I'm thinking HunkEditor) appearing on the wiki aren't disrupting your general plans |
| 13:11 |
|
kowey |
I guess that hunk editor page belongs on the wiki, given its informal style |
| 13:11 |
|
twb |
kowey: did I tell you I worked out how to turn the wiki (or subsets of it) into a PDF/static HTML manual? |
| 13:12 |
|
kowey |
oh, you might have! I remember something to that effect |
| 13:12 |
|
kowey |
via docutils? |
| 13:15 |
|
kowey |
do you have a clear idea what you want the Future to look like? (I see wiki, web, pdf, man, online help) |
| 13:15 |
|
kowey |
it sounds like you're saying that the Future will consist of a high-sobriety subset of the wiki => everything else |
| 13:18 |
|
twb |
Hmm? |
| 13:19 |
|
twb |
kowey: sorry, I wasn't listening because there was no "twb:" |
| 13:19 |
|
twb |
Via docutils and sphinx and rst2pdf, yes. |
| 13:20 |
|
twb |
The vision is to end up with a VCS-backed wiki, in rst format, of which a subsection is well-tended and is automagically converted into a printable manual. |
| 13:20 |
|
twb |
I'm not sure how that would best tie in with the manpage/online help. |
| 13:21 |
|
kowey |
thanks for reminding me! |
| 13:21 |
|
kowey |
ok, so potentially two divergent sources of documentation in the future wiki => pdf (no more need for web as subsumed by wiki) |
| 13:22 |
|
kowey |
plus online help / man page |
| 13:22 |
|
kowey |
maybe the two converge, maybe they don't |
| 13:23 |
|
kowey |
(I don't feel the need to flag you for this, just sort of idly absorbing the vision) |
| 13:24 |
|
twb |
The general idea is that most documentation should be editable by anyone |
| 13:25 |
|
twb |
But some "core" of it needs to be structured and assembled by someone with a view to the whole of the core |
| 13:26 |
|
kowey |
cohesion |
| 13:26 |
|
twb |
Anything that's changing fast gets pushed into the wiki, and anything that's really clear and helpful and unlikely to change might get absorbed from the wiki. |
| 13:27 |
|
kowey |
hmm, for example via a stable branch of the wiki? |
| 13:28 |
|
twb |
That could work, too. |
| 13:28 |
|
twb |
I hadn't really considered that, since darcs sucks so hard at long-term branches |
| 13:28 |
|
twb |
s/-term/-lived/ |
| 13:28 |
|
kowey |
nod |
| 13:29 |
|
kowey |
a definite +1 on strong editorial control for some subset of the documentation |
| 13:43 |
|
|
malcolmw joined #darcs |
| 14:02 |
|
twb |
1AM, time to knock off |
| 14:04 |
|
kowey |
good night |
| 14:10 |
|
mornfall |
:) |
| 14:10 |
|
mornfall |
Night. |
| 14:11 |
|
twb |
I still have to cycle home |
| 14:24 |
|
|
sm joined #darcs |
| 15:03 |
|
|
gal_bolle left #darcs |
| 15:29 |
|
Igloo |
Is there a way to convert a darcs-1 repo to a hashed repo without doing a "get" of it? |
| 15:41 |
|
kowey |
darcs optimize --upgrade |
| 15:41 |
|
kowey |
with Darcs 2.4 |
| 15:41 |
|
kowey |
it's not very well tested, I'm afraid :-( |
| 15:42 |
|
Igloo |
Ah, ta, I'll try upgrading |
| 15:44 |
|
Igloo |
Hmm, no darcs 2.4 even in unstable |
| 15:58 |
|
|
kpreid_ joined #darcs |
| 15:58 |
|
|
mxc joined #darcs |
| 16:10 |
|
|
gbeshers joined #darcs |
| 16:14 |
|
Igloo |
Hmm, darcs is using 2.2G of RAM |
| 16:15 |
|
Igloo |
but it claims to have worked |
| 16:25 |
|
kowey |
meaning it's still going, but it says it's done? |
| 16:27 |
|
Igloo |
It finished inbetween my two lines |
| 16:48 |
|
|
iaefai joined #darcs |
| 16:59 |
|
|
lispy joined #darcs |
| 17:42 |
|
|
arjanb joined #darcs |
| 17:57 |
|
|
kpreid_ joined #darcs |
| 18:09 |
|
|
gh_ joined #darcs |
| 19:12 |
|
|
volothamp joined #darcs |
| 20:19 |
|
Heffalump |
Igloo: what were you converting? |
| 20:21 |
|
Igloo |
The main ghc repo |
| 20:26 |
|
|
idnar joined #darcs |
| 20:27 |
|
* Heffalump |
wonders whether it's better with mornfall's de-slurpification patches |
| 20:28 |
|
kowey |
look, it's volothamp! (Luca) |
| 20:28 |
|
* kowey |
waves |
| 20:28 |
|
* Igloo |
looks |
| 20:29 |
|
kowey |
I'm slightly anxious about this missing index thing |
| 20:29 |
|
* Heffalump |
too |
| 20:29 |
|
kowey |
I hope we're not getting loads of people trying out darcs 2.4 only to find it explode on them |
| 20:30 |
|
Heffalump |
I hope we'd get at least a bit more noise |
| 20:31 |
|
sm |
I've often wished for a crash-reporting/talkback feature to gather such data more directly |
| 20:54 |
|
|
balor joined #darcs |
| 20:59 |
|
Igloo |
Argh. darcs1 can't understand the GHC repo, now |
| 21:00 |
|
Igloo |
darcs failed: Can't understand repository format: hashed |
| 21:01 |
|
Heffalump |
errm, isn't that obvious? |
| 21:01 |
|
* Heffalump |
wonders if he's missing something |
| 21:02 |
|
Igloo |
It would work fine if it just ignored the format type, wouldn't it? |
| 21:03 |
|
Heffalump |
huh? |
| 21:03 |
|
Igloo |
It's only pristine that's different isn't it? |
| 21:03 |
|
Heffalump |
the patches are stored in different filenames, the pristine is stored in different filenames, how could it? |
| 21:03 |
|
Heffalump |
patches are hashed too |
| 21:03 |
|
Igloo |
Oh, I didn't realise the patch filenames had changed too |
| 21:04 |
|
|
volothamp_ joined #darcs |
| 21:55 |
|
mornfall |
The inventory format has changed as well. |
| 22:01 |
|
|
sva- joined #darcs |
| 22:04 |
|
* lispy |
starts to worry about the 2.4 release as well |
| 22:04 |
|
lispy |
It seemed to work so well for us... |
| 22:05 |
|
Heffalump |
well, there hasn't been that much noise |
| 22:06 |
|
|
sva- joined #darcs |
| 22:11 |
|
|
sva- joined #darcs |
| 22:13 |
|
mornfall |
Hmm. |
| 22:13 |
|
mornfall |
Warning: This package indirectly depends on multiple versions of the same |
| 22:13 |
|
mornfall |
package. This is highly likely to cause a compile failure. |
| 22:13 |
|
mornfall |
package hashed-storage-0.4.7 requires mmap-0.4.1 |
| 22:13 |
|
mornfall |
package darcs-2.3.98.1 requires mmap-0.5.4 |
| 22:15 |
|
mornfall |
The problem is that in fact it does not cause a compile failure. But it does produce a crippled binary. |
| 22:15 |
|
mornfall |
Well, I haven't verified the latter, but it's highly likely. |
| 22:17 |
|
mornfall |
Hm. But now I have that binary and still can't reproduce the error. |
| 22:17 |
|
Igloo |
Can anyone explain what these numbers mean, OOI? |
| 22:17 |
|
Igloo |
http://paste.debian.net/62163/ |
| 22:18 |
|
mornfall |
Igloo: number of pristine files/dirs processed / total number of pristine files/dirs. |
| 22:19 |
|
mornfall |
And number of inventories in Writing inventory. |
| 22:19 |
|
mornfall |
It also means you have pre-2.4 darcs. :) |
| 22:19 |
|
Igloo |
Does it think my patches created 15 pristine files, then? |
| 22:19 |
|
mornfall |
Igloo: I think it only means it touched 15 pristine files. |
| 22:20 |
|
mornfall |
The cleanup might be done later than the synchronisation. |
| 22:20 |
|
Igloo |
Oh, is it because hashed pristine means there can be 15 more "files", even if no "files" are added? |
| 22:20 |
|
mornfall |
Well, it means that files are never modified, only added. |
| 22:20 |
|
mornfall |
And unreferenced files are removed (probably a bit later). |
| 22:34 |
|
Igloo |
So the numbers aren't the number that are actually used, but the number that exist. That seems a little odd. |
| 22:36 |
|
|
volothamp_ left #darcs |
| 22:36 |
|
|
volothamp_ joined #darcs |
| 22:37 |
|
|
volothamp joined #darcs |
| 22:54 |
|
|
JaffaCake joined #darcs |
| 23:03 |
|
|
gwern joined #darcs |
| 23:03 |
|
|
gwern joined #darcs |
| 23:15 |
|
|
JaffaCake joined #darcs |
