Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2016-04-12

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

All times shown according to UTC.

Time Nick Message
00:43 Big_G joined #darcs
02:29 Riastradh joined #darcs
03:09 mizu_no_oto joined #darcs
03:22 castlelore joined #darcs
04:43 alexei___ joined #darcs
05:48 castlelore joined #darcs
06:42 Heffalump does darcs revert clear them out?
07:02 inhortte joined #darcs
08:48 castlelore joined #darcs
09:11 peb` joined #darcs
09:17 ntnn joined #darcs
09:18 fr33domlover Heffalump, I'll try next time they appear :P the last darcs record made them disappear again
09:50 ntnn oy lads - how is darcs being pronunciated? just 'dark-s' or is there something special?
10:03 fr33domlover ntnn, afaik just Darcs? :P
10:04 fr33domlover Heffalump, another weird thing: One of my files appears as 'a' after editing it, even though I already recorded it several times and it should have been 'A'
10:04 fr33domlover repo went crazy :/
10:06 fr33domlover or maybe it's me being crazy? :P
10:21 fr33domlover sm: On Hackage there's a package highligher2 which claims to be faster than highlighter. It's a fork by Alex Suraci. Why does Darcsden depend on the older highlighter?
10:22 fr33domlover (I expected it would use highlighter2 since Alex is Darcsden's author :P)
11:46 BitPuffin joined #darcs
12:14 mizu_no_oto joined #darcs
13:07 mizu_no_oto joined #darcs
13:08 gour joined #darcs
13:09 gour pijul's FAW says: "Yes, we solved the exponential merge problem..." how does darcs behave in regard these days?
13:09 gour *FAQ
13:16 Heffalump gour: darcs-2 semantics are better than darcs-1, but they can still fail or get exponentially slow with deeply nested merges
13:17 gour it looks that pijul is quickly developing considering they even change language implementations...is it all due to darcs' lessons learnt from the past?
13:19 * gour is reading http://darcs.net/Using/Model preparing to pissibly return to darcs
13:19 gour *possibly
14:00 sm fr33domlover: darcs hub has been using the better-performing highlighting-kate for a long time. I hadn't seen highlighter2
14:10 Big_G joined #darcs
14:18 fr33domlover sm, I'm aware of highlighting-kate, I just want to use a Haskell highligher as long as it works reasonably :P or at least make it possible with a flag
14:33 sm fr33domlover: ok.. I think highlighter is still supported, with an off-by-default flag.. you could replace that with highlighter2
14:33 sm if it turns out to be better than h-k, we could drop the latter (but I doubt it will)
14:34 sm if it isn't better than h-k, is there a good reason to offer the option ?
14:34 sm easier to build on windows, I suppose ?
14:34 sm bbl o/
14:35 fr33domlover sm, there's also a package illuminate from 2011 which claims to be faster than h-k, but even its author now uses h-k by default, it seems (in Gitit; same author)
14:36 fr33domlover but i'm asking because i want to choose what to use for another web app
14:36 fr33domlover currently using highlighter2, if it turns out to be too slow i'll either offer h-k as an option, or look into trying to make a new faster highlighter
14:37 fr33domlover highlighter/2 uses pcre-light and operates on strict ByteString, maybe there's a smarter way
14:38 fr33domlover e.g. merely using lazy BS instead could mean constant-memory and it may make things faster (no need for huge allocation?) or maybe use something faster than pcre (maybe alex?)
14:40 mizu_no_oto joined #darcs
15:12 gour left #darcs
15:41 Guest98483 joined #darcs
15:44 ronny joined #darcs
16:04 Heffalump fr33domlover/pointfree: what's the outcome of the ssh discussions? Are you planning on updating the outstanding submissions (DBRG from fr33domlover and the new KEX from pointfree) or migrating to ssh-hans?
16:07 fr33domlover Heffalump, DRBG is ok, the problem with my code was the modexp thing (integer-gmp)
16:08 fr33domlover I'm waiting for feedback from pointfree
16:08 fr33domlover I want to try ssh-hans too, but it will take a while (hopefully few weeks, not more)
16:09 Heffalump so ssh 0.4 should just stay on hold for now? (that's fine, just want to know when to ping again)
16:15 alexei___ joined #darcs
17:21 carter What is ssh Hans?
17:22 carter Heffalump: what's the network engineering context ?
17:33 Heffalump https://github.com/glguy/ssh-hans
17:33 Heffalump what do you mean by network engineering context?
18:16 mizu_no_oto joined #darcs
18:26 Riastradh joined #darcs
19:00 carter Oh, is this for darcs or darcs den Heffalump ?
19:04 sm carter: fr33domlover and pointfree are updating the ssh package, which darcsden uses (and/or replacing it by ssh-hans)
19:04 sm I think darcs uses command-line ssh
19:08 Heffalump carter: darcsden
19:08 carter Cool
19:45 fr33domlover Heffalump, aside of the integer-gmp patch, which you said is an issue due to timing attacks, a step i think is important is to make a patch for darcsden to work with the new version, and release 4.0 only after testing with darcsden. For example, when I wanted to use it last week, I noticed that due to Channel and Session being newtypes and not aliases, they aren't instances of MonadBaseControl anymore
19:46 fr33domlover (I can add these instances, the point it that I'd never notice without testing with real world usage :P)
19:47 fr33domlover I'm working on implementing the Git protocol in Haskell, so I'll be testing with the 'ssh' package very soon. For now it's on hold unless pointfree says "I patched darcsden to work with ssh-hans" :)
19:58 pointfree fr33domlover, I haven't finished patching darcsden-ssh to use ssh-hans. It still needs to retrieve pubkeys from couchdb instead of ~/.ssh/authorized_keys and then run darcs commands when prompted.
19:58 fr33domlover pointfree, ok no rush :)
19:58 fr33domlover but i'd love to see the cod when you're done
19:59 pointfree :)
20:01 pointfree I'd probably be more efficient at this if I sit down to read a haskell tutorial. Haskell is a difficult language to muddle through.
20:10 fr33domlover pointfree, I'll get to that myself eventually so you can also leave it to me, if you don't mind waiting for a few weeks :P
20:29 alexei___ joined #darcs
21:01 inhortte joined #darcs
22:41 mizu_no_oto joined #darcs

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