Camelia, the Perl 6 bug

IRC log for #darcs, 2013-10-25

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

All times shown according to UTC.

Time Nick Message
01:25 intripoon_ joined #darcs
02:18 mizu_no_oto joined #darcs
02:42 arpunk joined #darcs
02:47 mtp joined #darcs
02:48 mtp joined #darcs
03:04 mtp joined #darcs
03:35 preflex_ joined #darcs
03:44 mtp joined #darcs
03:44 mtp joined #darcs
04:01 mizu_no_oto joined #darcs
04:09 mtp joined #darcs
04:09 mtp joined #darcs
06:28 amgarchIn9 joined #darcs
06:31 lelit joined #darcs
07:11 raichoo joined #darcs
07:17 kowey joined #darcs
07:23 raichoo joined #darcs
07:42 arpunk_ joined #darcs
07:54 kowey good morning!
08:33 notdan Hello
08:39 schlaftier joined #darcs
08:52 donri joined #darcs
09:29 owst joined #darcs
09:52 kowey it's an owst!
09:57 owst it's a kowey
10:06 kowey (nothing in particular, just waving)
10:06 owst Cool :-)
10:07 owst How're things?
10:07 owst I've not had any darcs time for so long :-(
10:08 kowey things are alright, thanks (day off today).  how about you?
10:08 * kowey vanished from the darcs world for at least half a year
10:09 kowey not sure if I'm back yet, takes a bit of pump priming
10:11 owst Busy busy, but otherwise good
10:11 owst Hmm, same. I'm not back, I'm sure of that :-)
10:12 kowey :-)
10:13 kowey Do you have teaching responsibilities on top of research?
10:13 kowey I didn't do much teaching, but can't imagine how people who did managed to juggle it all
10:20 owst Yeah, though that's not too bad
10:21 owst Though research isn't for me, my plan is to finish as close to the end of Sep '14 as I can, and move to London and get a job in The Real World
10:22 kowey oh! good job on finding out early
10:22 kowey I bet the Real World will find good things for you to work on
10:23 owst I'm sure it will!
10:23 owst Now if I can find a job writing Haskell rather than Python, I'll be happy ;-)
10:23 kowey :-) oh god Python
10:23 kowey (speaking as a more regular although still somewhat inept user)
10:24 kowey why do I have to run this to find out it's broken? what do you mean variables still exist outside of for loops?
10:24 owst But but but, rapid development ;-)
10:25 kowey heh heh heh, drives a programmer to drink
10:25 kowey rapidly
10:27 owst I think I'd settle for C# if I had to. I can't see myself ever caring enough about ridiculous details to program in C++ well
10:29 kowey well we have a local F# shop that came from C#
10:29 kowey it'd at least make a good “we'll still be compatible with old code” case
10:30 owst Mmm, F# would do too
10:50 mornfall Real World Sucks
10:51 mornfall owst: C++11 is a *much* better language than C# or Java
10:52 mornfall lately, I suspect that I maybe like it more than Haskell
10:54 notdan Really? I think C++11 is still much more verbose than C#
10:54 owst In what sense?
10:54 notdan and the errors are still way more compilcated than C#'s or Java's
10:54 owst Maybe it gets (some?) things "more right" but that's surely obscured by its complexity?
10:54 mornfall It's very expressive. And unlike Java it has a static type system.
10:55 mornfall It's not like GHC error messages were a pinnacle of clarity either.
10:55 owst I'm not sure I've ever bought the "more expressive" argument. To what end? To whose benefit?
10:55 owst I suppose all error messages are an aquired taste of the particular language/compiler :-)
10:55 mornfall owst: Do you enjoy typing in assembler much? :-P
10:56 owst Oh yeah, all the time ;-)
10:56 notdan sm: do you think this look OK? https://files.app.net/z953PPZA.png
10:57 mornfall More expressive translates to fewer lines translates to less fluff to keep in your head when reading code.
10:57 mornfall Programs are meant to be read. They are executable by accident. :-)
10:57 owst But isn't the more expressive argument saying "hey, if you want to care about bit patterns, memory allocation patterns and word alignment, you can!" rather than abstracting such things away from the "surface" of the language
10:57 owst Sure. Maybe I don't understand the use of the word expressive :-)
10:58 mornfall RG uses the word succinct.
10:58 owst In C++ world it seems to mean "yeah you can do all that crazy stuff if you want", i.e. being able to express those ideas in the language, whereas (to me) it should mean "the intent of my code can be easily expressed"
10:58 notdan Admitedly I haven't played with C++11 too much, but std errors are horrible, come on
10:58 notdan even Control.Lens errors are more bearable
10:59 owst notdan: haha!
10:59 mornfall notdan: You surely don't use g++, do you?
11:00 mornfall I guess VC++ might be a worse compiler, but come on. :-)
11:01 mornfall (If you didn't yet, get clang or something with the EDG frontend in it.)
11:02 mornfall I agree that it's a shame that concepts didn't make it, but there've been arguably good reasons for that.
11:02 notdan OK, I'll keep that in mind :)
11:02 mornfall Also, auto parameters in C++14. Yay.
11:02 notdan yeah I use g++, when I need to write C or C++
11:03 mornfall There's exactly one version of g++ that can compile our code without segfaulting or falling into an infinite loop. And yes, the error messages are straight from hell.
11:04 mornfall (Hm, it's not actually infinite, I think. It just takes a dozen hours or so to finish.)
11:06 mornfall I suspect a major reason why gcc can't keep pace is that it's ancient C codebase.
11:07 mornfall I.e. g++ is not a self-hosting compiler. (Yes, I also think that being self-hosting is a major advantage when writing compilers.)
12:42 arpunk joined #darcs
13:12 gh_ joined #darcs
13:21 mizu_no_oto joined #darcs
13:54 sm notdan: I think that's progress, but not ideal. Maybe the sidebar is not the best place to show a long url
13:55 sm also remember there's two urls (http and ssh)
13:57 sm though I see github does just what you do
14:02 sm unlike github, our ssh urls have several variants, I wonder which to show: [YOU@]hub.darcs.net:[OWNER/]REPO
14:04 gh_ joined #darcs
14:05 sm I suppose YOU@hub.darcs.net:OWNER/REPO which always works, and let them figure out how to abbreviate
14:05 sm as you've done
14:06 sm s/always works/always works if you have ssh access and owner or member rights to REPO/
14:08 sm again, it's simpler over at github - their ssh url always works (at least for read access) if you have ssh access
14:09 notdan Well the code I made shows the ssh url when the repository is private or the user browsing the repository is an owner/member
14:09 sm hi.. aha
14:09 notdan Yeah I don't like it in sidebar either, but I don't know where to put it :(
14:10 sm well it seems a step forward, let's deploy it ?
14:10 sm can you add a little whitespace to separate it from the buttons above
14:11 sm like between delete and fork
14:11 notdan Yeah that's a good idea
14:11 notdan should I use the <div> styled as a packs <div> or a .well?
14:12 sm uh.. not sure, I trust your judgement :)
14:12 sm I don't have a firm grasp of UI styles atm
14:13 notdan OK, I'll test it some more, as I haven't actually tested it with regard to showing the correct url when you are a member of the repository but not an owner
14:13 sm cool
14:15 sm this has made me realise darcs get hub.darcs.net:OTHERUSER/PUBLICREPO should probably always be allowed - right ?
14:15 sm currently it's not
14:15 notdan via ssh? yeah probably
14:15 sm ie, ssh read access to other peoples' repos even if you're not a member
14:16 sm public repos
14:29 notdan oh, the 'User' passed to repoBase and similar functions are not the 'current' user?
14:32 sm ...right, it's the user in the url
14:33 sm current user is (sUser session)
14:33 notdan btw I was wondering, how is it to work with Redis from Haskell?
14:33 sm I dunno, it just works :)
14:34 sm zero problems
14:34 notdan ha, ok. I myself am looking for a in-memory database with time-to-live keys
14:34 sm I haven't had to change anything there, so don't really know
14:34 sm but it sounds like a good option
14:35 sm back in a bit
14:46 qinglingquan joined #darcs
14:47 qinglingquan left #darcs
15:08 notdan okay I think I pushed all the patches
15:16 arpunk hi notdan :), is paste.hskll.org down?
15:16 arpunk joined #darcs
15:18 notdan arpunk: hi
15:18 notdan not for me
15:21 arpunk notdan: http://www.downforeveryoneo​rjustme.com/paste.hskll.org
15:21 arpunk it timeouts from here
15:22 notdan hm, that's weird
15:23 notdan Can you access http://hdiff.luite.com?
15:23 mizu_no_oto joined #darcs
15:26 arpunk is not loading for me
15:27 notdan Hm this is very strange, maybe there are some networking problems in the datacenter
15:28 notdan I can access the server from my computer and from my remote shell
15:29 arpunk :/, anyways, just discovered interactive-diagrams, awesome piece of service
15:34 notdan thanks! i am still working on it though
15:34 notdan it will be faster with the release of ghcjs
15:36 sm notdan: thanks! patches deployed on darcs hub, welcome to the project
15:37 sm I wonder how it would be to drop the outer box and grey background and shrink the font. Also the packs message needs to move to the repo settings page
15:38 amgarchIn9 joined #darcs
15:38 notdan Thanks :)
15:39 sm so you did that GSOC on interactive diagrams eh.. nifty
15:48 notdan yea
15:51 whaletechno joined #darcs
15:56 notdan For some reason I can't get the text in the input to become smaller https://files.app.net/z9z2sFXz.png
15:56 notdan bleh :(
15:57 sm ha, hours of fun :)
15:58 sm I'd love to have the UI overhauled and a clear style guide
16:08 lelit joined #darcs
17:25 dleverton joined #darcs
17:40 dleverton joined #darcs
17:41 arpunk_ joined #darcs
18:01 kmels joined #darcs
18:07 dleverton joined #darcs
18:09 mizu_no_oto joined #darcs
18:35 raichoo joined #darcs
18:37 gh_ Heffalump, n-dolio, I've sent a patch that attempts fix the cannot-cancel-record-in-spite-of-text-editor and failing-text-editor-produces-empty-patch-name issues you have with the current UI: http://bugs.darcs.net/patch1088
18:38 gh_ althought the way to cancel the record is a little hidden (first leave patch name empty in editor, then CTRL-C at prompt)
18:39 gh_ I haven't screened it yet
18:46 lelix joined #darcs
19:05 nomeata joined #darcs
19:08 Heffalump oh, thanks
19:09 dleverton joined #darcs
20:11 dleverton joined #darcs
20:25 kowey joined #darcs
21:11 amgarchIn9 joined #darcs
21:36 jyyou joined #darcs
23:44 schlaftier joined #darcs
23:55 javier_rooster joined #darcs
23:58 javier_rooster1 joined #darcs

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