Camelia, the Perl 6 bug

IRC log for #darcs, 2010-01-26

| Channels | #darcs index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:15 kpreid_ joined #darcs
00:22 thorkilnaur joined #darcs
00:30 sm joined #darcs
00:34 sm joined #darcs
01:22 twb joined #darcs
02:02 darcscommitbot joined #darcs
02:09 kpreid_ joined #darcs
03:39 tolkad joined #darcs
03:40 tolkad left #darcs
06:40 tomoj joined #darcs
06:43 lispy|web joined #darcs
06:57 lispy|web kowey: starting a benchmark now
07:34 balor joined #darcs
07:35 lispy|web joined #darcs
08:02 darcscommitbot joined #darcs
08:03 darcswikibot joined #darcs
08:03 lelit joined #darcs
08:07 merehap joined #darcs
08:10 merehap hey, I'm having a problem with darcs push:
08:10 merehap darcs failed:  Couldn't write `0000003931-950067a323f7a924e38a938bc32​623a4ccbc40c9ea7f917c06b560008ab6b098'
08:10 merehap in subdir inventories to sources:
08:11 merehap there a few google results for it, but nothing helpful that I could see
08:11 merehap I'm trying to push my repo to patch-tag
08:12 twb merehap: is it a single-user repo?
08:12 twb That is, has anyone else ever pushed to it?
08:13 merehap joined #darcs
08:13 merehap sorry about that
08:13 twb 19:12 <twb> merehap: is it a single-user repo?
08:13 twb 19:12 <twb> That is, has anyone else ever pushed to it?
08:13 merehap no, it has not been pushed before, I am the only user
08:13 twb I assume you're pushing via ssh.
08:13 merehap correct
08:13 twb If you just ssh in, do you get a shell?
08:14 merehap I've tried reuploading the key to patch-tag a few times
08:14 merehap what command are you suggesting?
08:14 twb Probably "ssh patch-tag.com"
08:15 merehap yes, that was successful
08:15 merehap I got the ubuntu prompt for the patch-tag server
08:16 merehap ssh merehap@patch-tag.com
08:18 twb What is "the ubuntu prompt"
08:18 merehap I've attempted repairing my repository which succeeds (this apparently fixed the problem for gwern as per google search results)
08:19 merehap Linux patch-tag.com 2.6.18.8-linode22 #1 SMP Tue Nov 10 16:12:12 UTC 2009 i686
08:19 merehap The programs included with the Ubuntu system are free software;
08:19 merehap the exact distribution terms for each program are described in the
08:19 merehap individual files in /usr/share/doc/*/copyright.
08:19 merehap Ubuntu comes with ABSOLUTELY NO WARRANTY, to the extent permitted by
08:19 merehap applicable law.
08:19 merehap sh-3.2$
08:20 merehap their ssh welcome message, I guess I should say
08:20 twb merehap: that's a shell.
08:20 twb I guess you aren't familiar with unix at the command line?
08:21 merehap I'm learning, I've used ssh before but clearly I don't know the terminology :)
08:21 twb We should probably remind the patch-tag.com developers to fix their /etc/issue and /etc/motd.
08:21 twb merehap: do you know how to inspect files from the command line?
08:21 merehap yes
08:21 merehap I have some ability with vim
08:21 twb merehap: you basically want to check that the ownership and permissions for your repo are sane
08:21 merehap ok
08:21 twb Particularly that _darcs/patches is writable and executable by your user
08:23 merehap drwxr-xr-x 2 root root  4096 Jan 25 03:29 patches
08:23 merehap I'm root on the machine and created the repo
08:24 merehap I'm guessing that I have the necessary permissions, judging my the output of ls -l there
08:25 twb You're root on patch-tag.com?!
08:25 merehap no, on my local machine :)
08:25 merehap I don't think they are that naive :)
08:25 twb The failure is about creating files on the destination
08:25 twb We need to know about the permissions on the destination
08:25 merehap all of the files in that location have the following permission set:
08:25 merehap -rw-r--r-- 2 root root 7712 Jan 25 03:21 0000020004-37874cda076bd6acd3e38b5eb5c​a4d15569e711d6ee1ebbb2c5382e52ed878ee
08:26 merehap oh
08:26 merehap just a second
08:27 merehap Public  Only project members; uploads disabled
08:27 merehap I may have been led astray as to what "uploads disabled" meant...
08:27 twb I dunno what you're citing there, but I meant you should ssh into patch-tag.com and inspect your repo using the CLI
08:29 merehap guess I may be an ssh noob, as apparently ls isn't enabled on the target machine
08:30 merehap what I displayed above was taken from their website, let me see if there is something there I can fix
08:31 twb That's probably because your shell is locked down
08:32 twb Which was what I really meant when I asked 19:13 <twb> If you just ssh in, do you get a shell?
08:32 twb You could try using an SFTP client to inspect it.  The sftp module of ssh is probably still enabled.
08:33 merehap help displays a considerable number of commands, would I see these if it were locked down?
08:34 twb They're part of the shell
08:34 merehap for example, it seems cd works
08:34 twb SFTP is going to be the least painful way to check, I think
08:35 merehap a shell that doesn't have ls but does have cd? very odd
08:35 merehap ok, I'll check it out
08:35 twb cd MUST be implemented inside the shell.  ls is an external program
08:36 merehap ah, well thanks for the unix 101, I guess that's not what you were hoping to teach by hanging out in this channel :)
08:36 twb Whatever.
08:36 twb If you really want Unix 101, you can google for "RUTE", which is an online textbook on the subject
08:36 sea-gull joined #darcs
08:37 merehap well I've been learning gradually, though I didn't expect to learn this to get darcs working with patch-tag
08:37 merehap anyways, thanks for your help
08:40 twb I strongly suspect what you're seeing is a wacky edge case that patch-tag hasn't coded for properly/yet.
08:41 twb Darcs doesn't (yet) make things easy for security-conscious hosting providers.
09:09 gal_bolle joined #darcs
09:26 Korusef joined #darcs
09:45 gh_ joined #darcs
09:49 tux_rocker joined #darcs
10:32 kowey joined #darcs
10:34 kowey morning
10:43 gh_ hi
10:47 kolibrie joined #darcs
11:22 JaffaCake joined #darcs
11:26 mornfall ZuriHac booked.
11:26 kowey yay!
11:27 kowey please add yourself to http://www.haskell.org/has​kellwiki/ZuriHac/Attendees
11:27 kowey I've been tacking (Darcs) on to people's affiliations to help tibbe out
11:27 kowey I'm assuming we're allowed to claim gal_bolle as one of our own? :-)
11:28 kowey oh, and we've passed the $1000 mark in our fundraising - announcement to follow
11:29 mornfall I have noticed. :) (due to identi.ca)
11:32 darcswikibot 26 Jan 11:31 - add myself to travel plans (mornfall)
11:42 twb Is "tibbe" a yiddish loanword or something?
11:42 twb Oh, I see, tibbe is someone's name.
11:43 twb I was reading it like "to help [us] divvy up the spoils"
11:44 kowey yeah, Johan Tibell
12:51 gal_bolle kowey: yes, i'm one of yours
12:51 gal_bolle (hence one of ours)
12:52 gal_bolle joined #darcs
12:58 kowey OK, I've added you (and Petr) to the list
13:16 malcolmw joined #darcs
14:30 kowey informal benchmark of darcs HEAD on darcs vs darcs with optimised pristine: http://hpaste.org/fastcgi/hpa​ste.fcgi/view?id=16582#a16582
14:33 mornfall Interesting.
14:34 kowey running the GHC one before I run off to a meeting
14:34 kowey we don't know if meaningful due to lack of stats-fu, but I guess the memory numbers don't lie
14:42 Korusef joined #darcs
14:45 gal_bolle left #darcs
14:54 twb kowey: those stats aren't meaningful without other numbers to compare them to
14:54 twb (Since the benchmark will be hardware-dependent.)
15:21 kowey yeah
15:24 kowey I was going for some sort of informal feasability study on generating benchmarks for optimize --pristine
15:24 kowey but I think that's probably out of our league, since we're still struggling with the basics
15:24 kowey anyway I just wanted to run one and see what happened
15:33 darcswikibot 26 Jan 15:32 - jot down some more stuff to work on during the sprint (kowey)
15:35 lispy joined #darcs
15:36 kpreid__ joined #darcs
15:38 gbeshers joined #darcs
15:40 gbeshers joined #darcs
15:41 kowey and here are those meaningless GHC numbers: http://hpaste.org/fastcgi/hpa​ste.fcgi/view?id=16582#a16586
15:42 lispy kowey: any ideas why my benchmark is not complete?
15:43 lispy kowey: also, what is the relationship between 2.3.98.1 and 2.4b1?  I think it's very confusing to refer to one version of darcs with two different version #s.
15:44 kowey lispy: ah! you weren't around for the design of that convention
15:45 kowey lispy: our beta numbering scheme uses .97.x for alphas, .98.x for beta and .99.x for release candidates (thx, mornfall)
15:45 kowey this is so that we can give sensible numbers to darcs-beta
15:45 kowey so the confusing thing is that darcs 2.3.98.2 would be darcs 2.4 beta 2
15:45 kowey because you want darcs beta to be < darcs
15:46 kowey perhaps a less confusing way would be if we switched to the odd/even scheme for the next release
15:47 kowey so that darcs 2.5.0.1 would be alpha 1 for darcs 2.6, and so on...
15:47 lispy My only complaint is the duplicity of version numbers
15:47 lispy If we could fix that, I'd probably be okay with it
15:48 lispy kowey: Different topic: Has anyone investigated using cabal to insert version numbers of dependencies in --version output?  so that we could say, "Run darcs --version and tell us which version of hashed-storage you're using."
15:50 lispy If it's easy to get those version #s into the --version output then it could be really handy
15:50 lispy If it's difficult, then I say lets not worry about it yet
15:50 mornfall lispy: You haven't been keeping up. Try --exact-version on darcs HEAD.
15:52 lispy mornfall: it will be a while before I can test that.  So it's implemented?
15:53 mornfall lispy: http://pastebin.ca/1766638 -- if that's what you mean, yes
15:53 lispy mornfall: cool, that's really nice
15:53 lispy I never think to use --exact-version I guess
15:53 kowey and once we've got people's flags down, future bug reports will be a lot easier to deal with
15:54 lispy kowey: flags?  as in the options to 'cabal configure'?
15:54 kowey yeah
15:55 kowey for example, it would be nice if darcs --exact-version could tell us if the user had -fcurl-pipelining or not
15:56 kowey here's the bug: http://bugs.darcs.net/issue849
15:56 mornfall kowey: That's harder, since I haven't figured a way to extract that data.
15:56 mornfall From Cabal that is.
15:57 mornfall Maybe it's doable in 1.8, haven't tried yet.
15:57 kowey we could bug dcoutts about it, perhaps when both of us have free moments
15:58 kpreid_ joined #darcs
15:58 kowey both meaning him and some general subset of darcs folks
15:58 kowey gh_ : perhaps a nice feature for you to work on during the sprint is that gzRead thingy you were asking for
15:59 kowey gh_: except I no longer remember what that was, only I remember you telling me about it over lunch in September and me replying that there was a ProbablyEasy bug for it :-)
16:03 darcswikibot 26 Jan 16:02 - copy over probablyeasy priorities from 2009-11 sprint (kowey)
16:09 gh_ ah yes, to apply a gzipped patch
16:09 lispy We use old-time still.  I wonder if that package is in risk of being deprecated
16:09 kowey @tell tux_rocker perhaps we should do the odd/even thing next release (darcs 2.5.1.x as darcs 2.6 alpha, 2.5.2 as beta; 2.5.99.x as rc) - see http://irclog.perlgeek.de/​darcs/2010-01-26#i_1929983
16:09 lambdabot Consider it noted.
16:10 kowey gh_: here's the bug number if you're interested - http://bugs.darcs.net/issue1427
16:11 kowey maybe I'll work a bit on plugins, myself
16:16 tux_rocker joined #darcs
16:20 gh_ "optimize --upgrade should do optimize - http://bugs.darcs.net/issue1643" are you sure it is still not in HEAD ?
16:22 darcswikibot 26 Jan 16:21 - more stuff to do (especially the new darcs annotate output) (kowey)
16:22 kowey oh! darcs tells me that Christian Berrer fixed it during the sprint
16:23 gh_ hehe
16:24 gh_ it's in the announcement of the 2.4 betas too...
16:25 kowey that's a good one to have in 2.4 too, since we're asking people to run darcs optimize --upgrade
16:26 darcswikibot 26 Jan 16:25 - issue1643 already fixed, thanks Christian B! (kowey)
16:28 gh_ yup
16:47 darcswikibot 26 Jan 16:45 - undarcs precision (Guillaume Hoffmann)
17:31 sm morning all
17:31 sm gh_, how do you edit the wiki.. over the web ?
17:38 kowey ah mornfall: I've worked out the reason I don't use apply --test on the staging repo, I think
17:39 kowey oh no, wait... no I haven't quite... nevermind
17:53 tux_rocker joined #darcs
17:53 lispy kowey: re: stage1 benchmarking, I'm skeptical that my problem was just slow benchmarks.  But, we'll see when I get back home.
17:58 mornfall kowey: No luck with the internal-library patch yet?
18:00 kowey internal-library patch?
18:00 mornfall cabal 1.8
18:00 kowey oh, I thought I replied to that
18:00 mornfall Hm. Have you?
18:00 kowey ah, no I haven't
18:01 kowey I must have thought the better of it, because mine was just a bureaucratic "ok, it looks like the best thing to do is just wait for cabal-install to support this"
18:03 kowey superficially: isn't there a lot of unrelated stuff in that patch?
18:04 mornfall Well, not really. But it needs sorting out.
18:05 mornfall (The Test bit is because GHC can't seem to handle overlapping instances across package boundaries... dunno why.)
18:05 * sm guesses no-one is pushing to the wiki, unless they log in as gitit, due to file permissions
18:05 mornfall The tools/ bit is because otherwise the internal library is never used.
18:05 * sm thinks setting it group-writable and -sticky and adding users to the gitit group would be a good way to go
18:06 kowey sm: the procedure is to push as gitit@
18:06 kowey sm: almost nobody has an account on darcs.net
18:06 sm I see
18:07 kowey what's the MyShow stuff for?
18:07 sm oh, it still picks up my name, from the patch. Excellent
18:08 sm gh_: nice cleanup on Related Software page
18:09 darcswikibot 26 Jan 18:03 - small clarification (Simon Michael)
18:17 kpreid_ joined #darcs
18:21 arjanb joined #darcs
19:04 gh_ joined #darcs
19:32 gh_ sm, i push to the wiki as kowey said
19:52 kpreid_ joined #darcs
20:18 sea-gull joined #darcs
20:35 darcswikibot 26 Jan 20:34 - Fixed lack of 'darcs' in 'cabal install darcs'. (robinbb)
20:39 nomeata joined #darcs
21:08 tux_rocker joined #darcs
21:42 kowey joined #darcs
22:29 tomoj joined #darcs
22:30 sshc joined #darcs
23:47 twb joined #darcs
23:48 kpreid_ joined #darcs
23:49 lispy BTW, the system that postgres uses for documentation has a very nice presentation.  See here: http://www.postgresql.org/docs/8.2/​interactive/functions-sequence.html
23:50 lispy It makes me wish our manual looked that nice, featured search, and allowed user comments.

| Channels | #darcs index | Today | | Search | Google Search | Plain-Text | summary