Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2014-10-09

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

All times shown according to UTC.

Time Nick Message
00:08 siel joined #darcs
00:13 mizu_no_oto joined #darcs
00:24 byorgey joined #darcs
00:45 mornfall joined #darcs
01:04 mizu_no_oto joined #darcs
02:50 dolio joined #darcs
03:08 alexei_ joined #darcs
04:08 byorgey joined #darcs
05:04 c74d joined #darcs
05:21 Heffalump sm: I'm not sure if it fixes the immediate issue with patchname, but perhaps in general darcsden should build against darcs-reviewed rather than darcs-screened?
05:21 Heffalump at least production hub.darcs.net should
05:35 Heffalump I've made a note on issue/111 about what I think is the right fix, I'll update that/the channel if I get it tested
05:49 lelit joined #darcs
06:12 Heffalump sm: patch in ganesh's darcsden should fix it
07:06 gal_bolle joined #darcs
11:00 mizu_no_oto joined #darcs
11:50 Heffalump hi gal_bolle
13:56 gh_ joined #darcs
14:00 byorgey joined #darcs
14:31 mizu_no_oto joined #darcs
14:41 sm Heffalump: excellent! thank you. cf http://hub.darcs.net/simon/darcsden/issue/116
14:42 sm fix deployed
14:43 sm I have a resistance to darcs-reviewed, but I'm open to persuasion.. using it, wouldn't I have run into the same issue later ?
14:44 sm and libdarcs would have heard this feedback later ?
14:48 sm has anyone here seen problems logging in to darcs hub using github ? I'm looking for ideas about http://hub.darcs.net/simon/darcsden/issue/102
14:54 sm on another topic: I love darcs convert export | git fast-import , but I wonder why it converts . to _ in tag names
14:56 gh_ only in tag names?
14:56 gh_ let me see
14:56 gh_ sm, also beware that since the A-B-C to C patch hash change, incremental exports made with the A-B-C format are broken
14:56 gh_ in case you already did some
14:58 gh_ yeah I wonder why the function cleanTagName consider . a bad character.. mornfall?
14:58 gh_ maybe it's a git thing
15:08 sm gh_: thanks for the heads-up, I'll avoid incremental conversions
15:08 gh_ sm, it should not change anymore though
15:08 sm I seem to remember it was a problem for older git or git on older windows platforms, but . in git tags is fine nowadays
15:09 sm I'm converting the FunGEn repo, I think the game hackers are on git
15:10 gh_ I mean the hash should not change, the rest maybe yes :)
15:10 gh_ before the 2.10 release we should not hesitate breaking stuff in the git export
15:12 sm it's so useful and reassuring, I'd almost vote to make this a first-class command
15:13 sm or add such commands. darcs convert-to-git/convert-from-git
15:16 gh_ I think these are the forbidden sequences: https://www.kernel.org/pub/software/scm/git/docs/git-check-ref-format.html
15:25 byorgey joined #darcs
15:42 stepkut joined #darcs
15:43 sm just playing around, I converted the darcsden repo, and for some reason it converts only 3 of the 4 tags (it misses 0.5)
15:45 sm alias darcs-convert-to-git='git init && darcs convert export | git fast-import && deunderscore-git-tags && git reset'
15:45 sm alias deunderscore-git-tags='git tag | grep _ > t; for t in `cat t`; do git tag `echo $t | sed -e s/_/./g` $t; git tag -d $t; done'
15:46 lelix joined #darcs
15:49 alexei_ joined #darcs
16:04 mornfall joined #darcs
16:21 Heffalump sm: true re darcs-reviewed
16:25 alexei_ joined #darcs
16:57 mornfall joined #darcs
16:57 mizu_no_oto joined #darcs
17:10 stepkut` joined #darcs
17:15 stepkut joined #darcs
17:18 mornfall joined #darcs
17:22 Heffalump sm: big pile of patches in http://hub.darcs.net/ganesh/darcsden-testfix for making the test suite more reliable and nicer to use. Note that two of them actually change darcsden's HTML/Javascript to work around apparent bugs/problems in HTMLUnit, so you may rightly be skeptical of their appropriateness
17:29 alexei_ joined #darcs
17:31 sm Heffalump: very nice!
17:33 sm http://hub.darcs.net/ganesh/darcsden-testfix/patch/20141008060902-81bb2 is the only questionable thing I notice
17:33 sm darcs push --dry
17:34 sm (typo)
17:34 sm is the /../../.. needed ?
17:34 sm oh I see. nm, merged
17:35 sm I was wishing for this just yesterday of course - I see you used your time machine
18:13 Heffalump sm: I sort of thought a relative URL would make the whole setup easier to run under a subpath at some undefined time in the future, but there are absolute alternatives too, e.g. http://tosbourn.com/a-fix-for-window-location-origin-in-internet-explorer/
18:15 sm FYI, with my usual poor timing, darcs hub is being affected by a system upgrade just now
18:16 stepkut` joined #darcs
18:21 alexei_ joined #darcs
18:34 sm joined #darcs
18:44 sm joined #darcs
20:13 mizu_no_oto joined #darcs
21:55 * Heffalump wonders about making a Darcs.API.DarcsDen module in darcs and re-exporting all the stuff darcsden uses from that.
22:00 sm isn't all of Darcs.* the api ?
22:02 Heffalump well, that's the current problem, yes :-)
22:02 sm could you hide all of it and expose the bits used by darcsden and any other libdarcs users ?
22:02 Heffalump it's not trivial to actually hide it because libdarcs is used by the darcs exe too
22:02 sm I guess everything else might have to go under Darcs.Internal or something
22:03 Heffalump I think in practice if we make Darcs.API.someting we can just tell people to use that
22:03 sm maybe that's a good end goal, but Darcs.API would be a good start
22:03 Heffalump but my idea with Darcs.API.DarcsDen is not to put time into defining a proper API, just to codify what DarcsDen uses right now (and potentially make other modules for any other clients)
22:03 sm it sounds like a good start
22:05 Heffalump re the release thing, I think the only way we could put 2.9 on hackage would be as darcs-beta, and even then it's not necessarily at the beta stage yet. Or could releasing darcsden to hackage wait for darcs 2.10 to release to hackage?
22:06 sm it can certainly wait, that's my plan
22:06 sm noone is pressuring me for a darcsden release :)
22:06 sm just my own sense of "it's been too long since a release"
22:06 Heffalump ok, cool
22:07 sm which I will ignore
22:07 sm I don't think a lot of folks need to run darcsden as it stands
22:07 Heffalump yeah. And hopefully by release time it'll be usable as a personal UI for local repos
22:08 sm but what do I know, it might be big in China
22:08 sm (but I doubt it)
22:08 Heffalump :-)
22:08 siel joined #darcs
22:08 Heffalump oh, btw, I'm taking over redis on hackage from the current maintainer
22:11 sm cool. Is it not obsoleted by hedis ?
22:13 Heffalump I believe so, but I think doing basic maintenance for now is less effort than switching
22:18 sm I'm gonna do a bugfixing
22:18 sm a little
22:18 Heffalump of what?
22:19 sm I dunno.. something easy
22:19 sm soon as I get off the phone with IRS
22:19 Heffalump bugfixing sounds more fun
22:19 sm it's the perfect time :)
22:20 sm ouch.. wait time between 30 and 60 minutes
22:44 sm hrm, why is there a recently updated _darcs directory in the darcsden user's home directory, which is not a darcs repo
22:45 sm containing just inventories/ (300, from 2011 till today) and patches/ (30k, from 2008 till today)
22:47 sm updated eg when I view a repo's changes page
22:49 sm I expect it to be writing stuff in ~/users/USER/REPO/_darcs , or ~/cache/darcs, instead
22:51 sm we moved from ~/.darcs/cache to ~/.cache/darcs, right
23:31 mizu_no_oto joined #darcs
23:37 sm ganesh leads in number of files.. all those darcs repos
23:49 sm I lead in disk space.. the old ghc repo
23:50 sm phew! 1h30m exactly on the phone. I fixed no bugs, but cleaned up the darcs hub ops makefile & checked nothing is on fire

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