Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2013-12-13

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

All times shown according to UTC.

Time Nick Message
01:04 amgarchIn9 joined #darcs
01:36 arpunk joined #darcs
01:56 whaletechno joined #darcs
02:13 mizu_no_oto joined #darcs
02:28 markstos joined #darcs
02:35 mizu_no_oto joined #darcs
02:46 mizu_no_oto joined #darcs
04:09 preflex joined #darcs
04:28 mizu_no_oto joined #darcs
05:01 arpunk_ joined #darcs
06:11 DarkFox joined #darcs
06:12 DarkFox How do you take an existing repository and remove files from the first recorded set?
06:12 DarkFox I.e. splitting a game with modifications to just the server and no client.
06:12 DarkFox With no modifications done to the client.
06:18 lelit joined #darcs
06:37 Heffalump DarkFox: so you want a new incompatible repository that's just for the server?
06:40 DarkFox Heffalump: Not quite incompatible? Just stripped for what is required for the server component.
06:40 DarkFox The server is compatible and doesn't need any of the client code / images / audio / etc
06:41 Heffalump by "incompatible" I mean you wouldn't be able to push/pull patches between the two
06:41 DarkFox Oh
06:42 Heffalump there's no way you can avoid that - I think it's a property of any DVCS with atomic commits that you can't split a repository like that and keep compatibility
06:42 DarkFox Other copies of the repo will be easier to be just destroyed and re-pull... It's on a private repo with a friend and myself.
06:42 Heffalump and you don't want to pull/push with the client repository in future?
06:42 DarkFox Destroy over ~40M and pull ~3M
06:43 DarkFox I restarted the repo last night to re-write all the modifications, the only problem is that it still has the mass of client data.
06:43 DarkFox I can recreate the whole repo again, doesn't affect anyone here.
06:43 DarkFox Just myself.
06:44 Heffalump ok - if you want to do this, are you comfortable with grabbing the latest development tree of darcs? Unfortunately you need the "rebase" feature which isn't released yet.
06:44 DarkFox Can I export the changes to patches and darcs init on the minimal base, then apply the patches?
06:45 Heffalump in theory you could do that with darcs diff, yes. That's roughly what rebase will automate for you.
06:45 DarkFox Personally I think I'd rather have a series of patch files
06:45 DarkFox Heffalump: How would I use darcs diff to create patches for each record?
06:45 Heffalump fair enough
06:46 Heffalump off the top of my head, it's just darcs diff --patch "patch name" for each one
06:47 DarkFox That compares to current?
06:47 DarkFox Ah -to and from look nice
06:47 Heffalump --patch should compare just after that patch against just before (I just tested it)
06:47 DarkFox Oh
06:47 DarkFox So running on the first is pretty useless :P
06:48 * DarkFox looked at the output wrong lol
06:48 * DarkFox should probally sleep haha
06:48 Heffalump it's still useful because it'll diff against empty
06:48 Heffalump which is what you want for your first replacement patch
06:48 DarkFox Sweet
06:48 Heffalump only thing is I'm not sure if it'll work nicely for binary data
06:48 DarkFox diff doesn't touch binary
06:48 DarkFox Just says it changed
06:48 Heffalump also if you have any "token replace" or "move" patches you'll have a problem
06:49 DarkFox Well.. for default behaviour at least
06:49 Heffalump (these are the things rebase will do better than the diff-and-patch approach)
06:49 Heffalump afk, back in a few mins
06:49 DarkFox So.... cabal-dev install -j darcs-devel?
06:49 * DarkFox runs cabal update - 47 days apparently.
06:51 DarkFox Heffalump: What tokens?
06:51 * DarkFox doesn't think he has any token replace or move patches?
06:56 Heffalump DarkFox: darcs get http://darcs.net darcs-devel ; cd darcs-devel ; cabal-dev install -j
06:56 * DarkFox vim's his patches
06:56 Heffalump the dev version isn't on hackage
06:57 Heffalump DarkFox: token replace is what you get from "darcs replace", move is a file move or rename recorded in darcs via "darcs move"
06:57 Heffalump or a directory move for that matter
06:57 DarkFox Noted, I'll give this a go soon, but just for now I think I'll go with the patch method because here, I can make modifications / merge each patch.... Not normal thing to do but I want it to be clean :P
06:57 Heffalump yeah, no problem
06:57 DarkFox Oh yea, makes sense :)
06:57 DarkFox But yea, I haven't done either so this should be easy
06:57 Heffalump you can actually do the same with rebase but then there's a learning curve and it is still experimental/rough around the edges
06:58 DarkFox What? Edit all the patches?
06:58 DarkFox Move things around?
06:58 Heffalump yes, basically it lets you change anything one patch at a time and then manages the conflicts with the next patch for you
06:58 * DarkFox wonders why the server keeps segfaulting....
06:58 DarkFox [1]    826 segmentation fault (core dumped)  ./teeworlds_srv_d
06:58 DarkFox Lol
06:58 Heffalump or at least helps with that
06:59 Heffalump did you write it in Haskell? :-)
06:59 DarkFox Heffalump: So tempted to port the game server to haskell :D
06:59 DarkFox But ... Vanilla -> Mod -> My extra modding. :P
06:59 DarkFox c++ and my modifications shouldn't affect the stability :|
07:00 Heffalump it's not actually that hard to write segfaulting haskell
07:00 DarkFox Except for maybe... the mass of projectiles - in such case wouldn't be a problem in production.
07:00 Heffalump either use a dodgy external C library like some versions of curl, or exploit a type system bug
07:00 DarkFox Heffalump: It is, unless you just use C libraries that segfault themselves?
07:00 DarkFox What type system bugs?
07:02 * DarkFox knew zero c++ until a few days ago. Now, I can re-write a lot of the whole game. :D
07:02 Heffalump https://ghc.haskell.org/trac/ghc/ticket/1496
07:02 DarkFox Server side at least
07:02 Heffalump it's being fixed in the next GHC though
07:02 DarkFox Heffalump: So that would be - using compiler extentions that aren't proven to be stable :P
07:03 DarkFox Probally C backed.
07:03 DarkFox Or no doubt? Lol
07:03 Heffalump well, most Haskell nowadays uses compiler extensions :-)
07:03 Heffalump and GHC does have a C-based RTS. But this bug doesn't exploit that.
07:04 DarkFox Maybe so, personally I don't use much more than LambdaCase, PatternView, and a few others but don't go an mass have things that don't make sense for the usage :D
07:04 * DarkFox doesn't see why a lot of people don't like case O_o
07:05 Heffalump darcs uses GADTs pretty heavily
07:06 DarkFox Also - "Haskell wouldn't be good for anything realtime anyway" - Someone in #C++ last night.
07:06 DarkFox GADT?
07:08 Heffalump Generalized Algebraic Datatypes
07:09 DarkFox Oh
07:09 DarkFox Yea
07:13 * DarkFox <3 darcs diff; forgot to record some changes before stipping everything down before - now I can just diff anything that has modified instead of being deleted :D
07:14 DarkFox darcs diff $(darcs status | awk '/^M/{print $2}')
07:14 DarkFox :-)
07:35 raichoo joined #darcs
07:40 DarkFox Heffalump: Just decided to attempt to apply the patches - been doing a number of things... And when I try darcs apply ../new/add-cwave.patch; darcs tells me .. darcs failed:  Bad patch bundle!
07:41 DarkFox Any idea what that is meant to mean?
07:43 DarkFox Do the timestamps in a patch file matter/
07:52 DarkFox :|
08:00 raichoo joined #darcs
10:00 arpunk joined #darcs
10:13 donri joined #darcs
10:14 DarkFox Tip for anyone who wants to compare the diff -> darcs diff | vim -
10:14 DarkFox Be cool if vim knew to read patch files and show like vim -d
10:26 gh_ joined #darcs
11:52 donri joined #darcs
11:54 owst joined #darcs
11:59 mizu_no_oto joined #darcs
13:14 nomeata joined #darcs
13:30 arpunk_ joined #darcs
14:11 mizu_no_oto joined #darcs
14:33 markstos joined #darcs
14:53 gh__ joined #darcs
14:53 gh__ hi
14:55 notdan Hello
15:16 sm hello
15:17 notdan Hey, what's up?
15:19 sm notdan: not much, just getting started
15:27 sm darcs' use of a pager drives me nuts in emacs shells
15:28 sm I really must sort this out
15:29 notdan Hm, is there a flag to force it not to use a pager?
15:29 sm setting PAGER=cat would do
15:40 sm that would require dealing with my multi-decade shell configuration catastrophe :)
15:40 mizu_no_oto joined #darcs
15:54 owst joined #darcs
15:56 dleverton joined #darcs
15:57 sm http://hub.darcs.net/darcs/darcs-screened/patch/20131119223227-5ef8f says it uses the pager when output is more than 20 lines, though I don't see where 20 is hard coded
15:58 gh_ sm, you may want to set DARCS_PAGER instead of PAGER
15:59 sm is a fixed limit standard practice ? I vaguely though programs would use a pager if their output is larger than the height of the terminal window (and in an ideal world, in an emacs shell the height would be infinite)
16:00 gh_ sm, the 20 lines stuff is done in Darcs.UI.External.viewDocWith
16:00 sm thx gh
16:01 gh_ sm, it's just a not-so-bad limit that is hardcoded because detecting the terminal height would require adding another dependency to darcs. this is the same limit as for the output of darcs help.
16:01 sm actually I don't want anything using a pager in emacs shells, so PAGER is better
16:02 mornfall eh
16:02 sm yes mornfall :)
16:02 mornfall less -F anyone? :)
16:03 gh_ less -XF
16:03 mornfall gh_: fix your terminfo :D
16:04 gh_ too much black magic for me :)
16:05 raichoo joined #darcs
16:06 sm echo $EMACS
16:06 sm oops
16:07 sm ah.. $INSIDE_EMACS is a way to detect an emacs shell window
16:08 sm unconditionally disable pager when INSIDE_EMACS is set, anyone ?
16:12 sm it's a special case, but actually might be a reasonable and safe way to add a bit of usability
16:14 pseudolio Is there an objection to adding the dependency to darcs that would let it determine the terminal size?
16:14 pseudolio Or, just that it would be more work?
16:15 sm pseudolio: I assume it makes installation harder eg on windows, because it requires c libs
16:16 sm though how great an impact that has on the world is arguable
16:18 sm ack, INSIDE_EMACS is also set in emacs's terminal emulation modes, where you would want the pager
16:18 pseudolio You could add it to just the unix deps, I guess.
16:19 pseudolio I guess that's pretty annoying.
16:20 sm when maintainers and packagers are scarce, I can see why any added complications are to be avoided
16:21 sm on the other hand, fixed 20 lines is a bit clunky looking :)
16:26 byorgey_ joined #darcs
16:26 gh_ pseudolio, "Or, just that it would be more work?" yes. but patches are always welcome, if you want to have a look at it
16:26 pseudolio Maybe.
16:28 gh_ also, we may waste too much attention and energy trying to make perfect something that is good enough (I suspect the "small output -> no pager" case to be seldom)
16:29 pseudolio I have tall terminals a lot. :)
16:30 sm gh_ I wonder if that now applies to darcs as a whole
16:30 pseudolio Does emacs really report infinite height?
16:30 sm it'd be good to know if darcs usage is growing or shrinking
16:31 sm pseudolio: I suspect "it depends" eg on $TERM, probably not
16:34 gh_ pseudolio, have you tried setting DARCS_PAGER="less -XF" ? is it useful for you?
16:34 dixie joined #darcs
16:34 burp joined #darcs
16:35 sm http://stackoverflow.com/questions/263890/how-do-i-find-the-width-height-of-a-terminal-window/7575044#7575044 <- doing it without c libs on windows
16:38 arpunk_ joined #darcs
16:38 pseudolio Works in unix-like environments.
16:38 pseudolio Doesn't seem to work in powershell.
16:39 pseudolio Maybe I just don't know how to set the variable correctly there, though.
16:41 pseudolio gh_: My ulterior motive is to also get width information, and make sure darcs help output fits that. :)
16:44 markstos joined #darcs
16:45 arpunk joined #darcs
16:46 pseudolio I actually don't use less anyway. I use vimpager, which detects the height itself. So I don't really need this.
16:49 pseudolio sm: The note on that answer says it returns the number of lines in the buffer, not the window size.
17:14 donri joined #darcs
17:17 amgarchIn9 joined #darcs
17:21 owst joined #darcs
17:29 lelit joined #darcs
17:39 sm oh right you are
18:12 mizu_no_oto joined #darcs
18:29 arpunk_ joined #darcs
18:51 raichoo joined #darcs
19:41 raichoo joined #darcs
20:25 alexei_ joined #darcs
20:55 donri joined #darcs
21:02 gh_ pseudolio, re: darcs help, a global approach to all darcs' terminal output sound good
21:26 owst joined #darcs
21:30 mizu_no_oto joined #darcs
21:59 addos joined #darcs
21:59 arpunk joined #darcs
22:20 amgarchIn9 joined #darcs
22:32 amgarchIn9 joined #darcs
22:34 amgarchIn9 joined #darcs
23:12 mizu_no_oto joined #darcs

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