Camelia, the Perl 6 bug

IRC log for #darcs, 2009-04-11

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

All times shown according to UTC.

Time Nick Message
01:25 mark__ joined #darcs
02:21 [1]intripoon joined #darcs
04:30 intripoon joined #darcs
07:10 twb Can some GUI browser users please view http://twb.ath.cx/tmp (draft for darcs' home page)?
07:29 sm hi twb
07:30 sm it looks like the usual one without styles
07:30 twb Hi; you have about five minutes before I afk to deal with dinner
07:30 sm I wondered if you saw my post about converting moin with pandoc
07:31 twb Not yet
07:32 twb I doubt I'll get to the moin->gitit stuff before tomorrow.  Friday saw a HUGE lot of triage from Eric that I'm still trying to work through :-)
07:32 sm two posts in that same thread.. you should read them before working more on moin conversion !
07:33 sm I'm pretty sure pandoc will read moin markup for us soon
07:34 sm pandoc is what gitit uses
07:35 twb sm: OK.
07:35 twb My "conversion" is about preserving moin's edit history, NOT about converting from moin to rest/markdown markup.
07:35 sm ahh
07:36 sm so you'll get all that into darcs ?
07:36 twb Yes.
07:37 twb Unfortunately, that also currently includes spam and spam rollbacks, which constitute like 80% of the history
07:37 twb OK, I gotta afk now
07:37 sm ack.. it's probably not worth your time then eh ?
07:37 sm we can just cut over
07:37 sm well, good luck and good weekend
07:37 sm time to crash here
07:41 ppavelV6 joined #darcs
08:07 cinimod joined #darcs
09:15 lelit joined #darcs
09:20 darcscommitbot 29 Mar 21:38 - Add issue1327.sh
09:20 darcscommitbot 11 Apr 07:38 - Canonize Aaron Kaplan.
09:20 darcscommitbot 11 Apr 07:39 - Alphabetise .authorspellings.
09:20 darcscommitbot 11 Apr 03:04 - Code churn in Darcs.Commands.Add and Darcs.Commands.Remove.
09:20 darcscommitbot 11 Apr 04:27 - Follow autoconf convention of @VAR@ for .in files.
09:20 darcscommitbot 11 Apr 08:32 - Remove unused import in Darcs.Commands.Remove.
09:32 arjanb joined #darcs
09:45 gh_ joined #darcs
10:22 kowey joined #darcs
10:22 kowey whoops, hope I'm doing enough in the fire-fighting dept (CMS)
10:27 markstos joined #darcs
10:52 gh_ how can I build darcs' web page from the .in file ?
11:59 dancor joined #darcs
12:00 dancor i used tailor to copy the lambdabot darcs repo into a git repo.  i have added a commit to the git repo.  to get that commit back into the darcs repo, can i use tailor?
12:02 twb gh_: for the front page, "autoreconf && ./configure && make doc/index.html"
12:02 dancor i need the git repo for work, but my plan is to periodically transfer the changes back into the darcs repo and mail them for hopeful inclusion in the public lambdabot
12:02 twb Really the only thing in the .in file is that @VERSION@ and @RELEASE@ get turned into 2.2.0
12:02 twb dancor: theoretically tailor can be bidirectional, but I haven't done that myself.
12:03 twb lelit: you're the tailor maintainer, right?  Wanna help dancor?
12:04 dancor i can't quite tell if it's meant to be.  the help has "Update options" but maybe that's just to later pull more thing in the original direction
12:04 twb dancor: I think what you do is define two targets, or have two tailor.config files
12:04 twb One for each direction
12:04 dancor ah perhaps
12:05 twb There is also git and darcs fast-export, but I have not had much success going from git to darcs.
12:05 twb IOW darcs-fast-import is still buggy
12:05 gh_ twb: thanks!
12:08 dancor twb: one other thing, my git-copy was created in the same dir as the .tailor file, which was expected.  what was unexpected was that the .tailor file is _part_ of the git-copy repo (was "added" with the first commit in history).  that seems odd; should i have kept the .tailor file in the parent directory?
12:08 dancor where do you keep your .tailor file
12:08 gh_ twb: fine, but there is hardly any style or layout in the page i'm looking at
12:16 twb gh_: the styling (which is minimal) is done by darcs.css in the same directory, I believe.  But the link is absolute, so you will need to change it temporarily while editing
12:16 twb dancor: I don't actually use tailor much.
12:16 dancor ok, thanks :)
12:16 twb dancor: but yeah, you have to keep it outside the repo you're working on :-/
12:16 dancor ok
12:17 dancor is it true that darcs patches have strictly more information than git revisions?
12:25 twb That's difficult to say.
12:43 dancor it looks like http://vmiklos.hu/project/dar​cs-fast-export/git-darcs.html is going to be the answer to my prayers here
12:44 dancor i just need to add a small fix since lambda actually has ^H's in the change descriptions!
12:44 dancor s/lambda/\0bot
12:51 twb Haha
12:51 twb Sounds like someone did a "darcs record" on a dumb terminal
12:52 twb And got ^H and ^? mixed up, or something.
12:52 twb gh_: I replied to your post.
12:53 gh_ ok
12:54 * twb goes back to playing games.
12:56 lelit hi dancor
12:56 dancor lelit: hi
12:59 lelit unfortunately, the so called two-way (see http://progetti.arstecnica.it/tailor/ticket/1) setup is not complete as is
13:00 lelit zooko in the past had a working setup of such a thing, but it's tricky
13:00 lelit and I never had time to finish the feature :-\
13:00 dancor ah ok
13:00 dancor thanks
13:01 dancor i'm going to see if git-darcs can do it for me, otherwise i might come crawling back and see if i can add it to tailor :)
13:01 lelit I usually kept short-lived "darcs branches" out of upstream svn repositories
13:02 lelit please report back any conclusion :)
13:02 dancor ok
13:02 lelit on the configuration, maybe you wanna add the "subdir" option to the project section
13:04 lelit this is my (almost) usual tailor script: http://pastebin.com/d420260f4
13:06 lelit in that example, it will create a "billiards" directory in the CWD, containing tailor private stuff, and the real repository in a "billiards" subdir under that
13:06 lelit the two level let me follow several branches (say, b-trunk and b-stable)
13:07 dancor ok
13:07 lelit once you set +x attr on it, you just execute ./billiards.tailor
13:17 gwern 'It uses tables (instead of CSS) to enforce a
13:17 gwern particular layout on me, which happens to look very bad in my tty (w3m)
13:17 gwern and print media (html2ps) browsers.
13:18 gwern Greetings people of the future! I respect your impossibly fast computers, but kids thesedays! you just don't know how to make pretty webpages
13:18 gwern allow me to bring you homespun wisdom from the '80s so you might mend your ways!
13:25 arjanb :)
13:37 twb FSVO "pretty" = different for every damn site.
13:38 twb It's my browser, and I'll bloody well set the link colour and the font once.  Authors have no place telling my browser to use the wrong fonts, the wrong colours, or ragged right.
13:39 twb And assuming that my A4 printout is a scrollable 4:3 raster display is just plain annoying.
13:39 twb If you want to assuming that everyone on the planet using Firefox 3, then I guess I can't stop you.  But that is the absolute antithesis of the ideas that the "web" was founded on.
13:42 twb Then 80s had glass ttys, and I guess the 10s will have glass paper... :-(
13:43 twb s/then/the/
13:43 markstos left #darcs
13:43 dancor e-ink?  so with an even worse refresh than the ttys :)
13:59 dancor as of darcs 2.2.1, does (darcs repair) fix the bad-crc's problem?
14:20 gh_ i think this is an issue that is going to be fixed in the next hacking sprint
14:22 Heffalump if we can get enough of Duncan's time, anyway
14:41 dancor joined #darcs
14:42 dancor lelit: git-darcs worked.  i had to deal with crc bs tho and upgrade git to get git-fast-import with --import-mark
14:42 dancor so i think it mostly works but its error-checking could be better
14:55 lelit great
15:36 _ joined #darcs
16:08 ppavelV6 joined #darcs
16:23 ppavelV6 joined #darcs
18:20 sm joined #darcs
18:39 chessguy joined #darcs
19:26 sm morning all
19:35 DanielC joined #darcs
19:51 mark_ joined #darcs
20:09 DanielC I'm working on the Patch Theory documentation. Can someone give me the list of *primitive* patches? I have Hunk and token-replace. For example, are add-file, remove-file and rename-file considered primitive patches?
20:12 arjanb a list is in src/Darcs/Patch/Prim.lhs
20:12 DanielC thanks
20:12 arjanb but you need to know a little haskell to read that
20:13 DanielC I know how to make the factorial function in Haskell, but that's the limit of my Haskell knowledge.
20:14 * DanielC is downloading darcs sources
20:19 arjanb there's some latex code in the core sources but i have no idea where pdf of that can be found
20:23 DanielC Judging from the sources, I would say AddFile, RmFile, AddDir, RmDir, Hunk, Binary, TokenReplace, Move, ChangePref, and Split.
20:23 DanielC I don't know what the last two mean.
20:24 arjanb Split is an obsolete thing related to darcs1 mergers
20:25 DanielC ok
20:27 arjanb ChangePref is produced by darcs setpref
20:27 * DanielC looks up setpref
20:28 DanielC I see.
20:28 DanielC Thanks.
20:29 arjanb it's debatable if that should be considered a primitive patch
20:29 Heffalump what else would you call it?
20:31 arjanb a workaround
20:31 Heffalump for what?
20:31 Heffalump versioning metadata strikes me as a very good idea
20:31 arjanb for not treating the relevant preferences as real files
20:32 Heffalump but how would you decide what they were called?
20:32 Heffalump I guess darcs could have some special name for them, perhaps pretending to be in the _darcs directory.
20:35 arjanb that would have had my preference
20:35 Heffalump I prefer more specific patch types, personally.
21:38 ppavelV6 joined #darcs
22:29 mmmdonuts joined #darcs
23:05 twb arjanb: the latex gets turned into HTML at http://darcs.net/manual/
23:06 arjanb it doesn't include everything in Darcs.Patch
23:06 twb That's fixed last week, but the page hasn't been rebuilt
23:06 twb In your own repo, "autoreconf && ./configure && make pdf"
23:06 arjanb i see
23:09 twb Fuuuuuuck, I can't find my moin-to-gitit work
23:10 twb Hmm, I hope this is it
23:11 twb Phew
23:34 mark_ joined #darcs
23:35 gwern twb: how's it going?
23:35 twb gwern: hi
23:35 twb I'm just teaching moin2darcs how to preserve author and date metadata.
23:37 gwern what does it know so far?
23:38 twb That's a little hard to articulate
23:38 twb Do you have a shell on cho?
23:39 gwern yes
23:39 twb ~twb/public_html/DarcsWiki.DRAFT is where I got up to last time.
23:41 * gwern looks at the faq
23:42 gwern a few issues I can see there
23:42 gwern raw links ideally have <>
23:42 twb They do?
23:42 gwern and I'm not sure about {{{}}}
23:42 twb Yes, the syntax isn't being converted, as I've said.
23:42 gwern twb: yeah, that way they get hyperlinked
23:43 gwern oh, my bad then
23:43 gwern what is being converted? just the revision history so far?
23:43 twb IMO if raw links aren't hyperlinked, that's another reason to use reST.
23:43 twb Yes, my goal is to preserve revision history.  Once that's done, conversion of syntax can be done in the resulting Darcs repo.
23:43 gwern whoa, what're those summaries?
23:43 gwern some sort of unicode translation?
23:44 twb Yes, though most of the unicode is spam.
23:44 gwern hm. as long as every patch touches just the one file, presumably we can snip all of it out of the history right?
23:45 twb That's my hope.
23:45 gwern (but actually I need to do dye eggs right now, so I'll look some more in half an hour or an hour)
23:45 twb Unfortunately I think it's not that simple, because if a spammer adds a line and then Eric removes it, those will be dependencies for patches that later edit that file.  I think.
23:50 twb Heh, moinmoin chokes on files that contain colons
23:56 twb gwern: should http://test.darcs.net/ work at the moment?
23:56 twb Should .page files be in the root directory of a gitit darcs repo, or should they all be in a sub-dir (e.g. ./pages/)?

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