Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2014-02-25

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

All times shown according to UTC.

Time Nick Message
00:03 amgarching joined #darcs
01:35 gh_ joined #darcs
01:42 amgarching joined #darcs
01:59 alexei joined #darcs
05:53 mizu_no_oto joined #darcs
06:18 notdan joined #darcs
06:23 lelit joined #darcs
06:42 favonia joined #darcs
08:04 raichoo joined #darcs
10:30 alexsuraci__ joined #darcs
11:06 gal_bolle joined #darcs
11:28 alexei joined #darcs
13:22 bishboria joined #darcs
13:42 mizu_no_oto joined #darcs
15:28 mizu_no_oto joined #darcs
15:40 colDrMcBeardman joined #darcs
16:21 raichoo joined #darcs
17:20 lelit joined #darcs
17:23 favonia joined #darcs
18:01 mizu_no_oto joined #darcs
18:10 notdan_ joined #darcs
18:31 mizu_no_oto joined #darcs
19:02 mizu_no_oto joined #darcs
19:27 favonia joined #darcs
19:44 Heffalump sm: do you know off the top of your head why DarcsDen.State.Users has functionality for deleting and renaming users that isn't used anywhere?
19:47 whaletechno joined #darcs
19:58 sm Heffalump: eh.. no
19:59 Heffalump ok, I'm going to delete it if you don't mind, along with the commented out code for "getUsersWithRepos"
20:01 Heffalump also there are some patches for your consideration in http://hub.darcs.net/ganesh/darcsden-abstract-settings - they should make the test harness safe to run. A couple of the tests still fail, but they shouldn't destroy the prod instance any more. The patches do include removal of --port as previously discussed.
20:01 sm ok
20:02 * sm wonders if hub user rdesfo is here
20:05 sm Heffalump: why are we adding implicit parameters, again ? I just skimmed http://www.haskell.org/ghc/docs/latest/html/users_guide/other-type-extensions.html#implicit-parameters , pretty interesting. It gives us more dynamic typing ?
20:06 Heffalump not dynamic typing, dynamic configuration (or at least top-level controlled configuration)
20:06 sm so eg it lets us vary the db without recompiling ?
20:06 Heffalump I'm using them to replace global variables
20:06 Heffalump it would in principle, but I haven't written anything to do that yet
20:06 sm and without plumbing arguments through everything ?
20:07 Heffalump so its the plumbing avoidance that's the main goal
20:07 Heffalump btw my previous backend abstraction stuff already used implicit parameters, it just hid them a bit more
20:07 Heffalump we could quite easily replace DarcsDen.Settings with a configuration file now
20:07 sm right.. I haven't seen these in any other project, quite interested to see how it works out
20:08 sm great
20:08 Heffalump I've carefully avoided the semantically subtle cases like rebinding their value in a nested call (as per our discussion about --port the other day)
20:09 Heffalump biab
20:11 sm my pull from your repo succeeded and marked conflicts with my unrecorded local config even though I have "apply no-conflicts" in _darcs/prefs/defaults, is that expected ?
20:13 sm now I have a mess that's hard to undo
20:17 Heffalump apply is just for apply
20:17 Heffalump so yes it's expected, I think you need pull no-conflicts or perhaps pull skip-conflicts
20:18 Heffalump if you'd pushed from somewhere else or used apply directly then it would have fired
20:29 sm I see, thanks
20:30 sm have to go, I'll test/apply these nice patches soon
20:30 Heffalump any objections to lenses btw? I want to use them (in their simpler form, just replacing record access/update) to help refactor the state data types.
20:41 burp joined #darcs
21:15 sm Heffalump: ooh.. well darcsden dev is led by its contributors, if you think it's a win go for it
22:19 Heffalump I'll see how it looks once I've actually tried it
23:04 dolio So, I was talking to the bzr fan at work a week or two ago, and its scheme for doing branches and sharing repo state is pretty cool.
23:05 dolio Basically, if you have a directory tracked in bzr, and inside that a subdirectory tracked in bzr, the subdirectory can elect to store its changesets in the outer directory.
23:06 dolio And then any changesets common to any of the subdirectories will automatically only be stored once.
23:06 dolio So, for instance, you never get into a situation where 'darcs get' uses hard links, but then subsequent pulls don't, and duplicate information on disk.
23:07 dolio And different branches of the same overall repository don't duplicate stuff on disk, either, just like they wouldn't if you had mercurial or git style branches.
23:10 mizu_no_oto joined #darcs

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