Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2016-04-01

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

All times shown according to UTC.

Time Nick Message
01:08 Big_G joined #darcs
01:23 darkfox joined #darcs
01:25 darkfox joined #darcs
02:16 mizu_no_oto joined #darcs
02:55 mizu_no_oto joined #darcs
05:18 Heffalump fr33domlover: I've pushed all your patches except the modular exponentiation one (which I'm rejecting given the timing attack issue) and the DBRG one which depends on it. Is the DBRG patch worthwhile independently of the exponentiation one? If so it'll need rebasing again to break the dependency - I can do that again (though this time it should be pretty easy!)
05:20 Heffalump pointfree: unfortunately as your patch depends on both you'll need to update it again, though from what you said I guess you might be changing it anyway?
05:20 Heffalump sorry for the all the noise, you wait ages for contributions and then two come along at once :-)
05:23 pointfree Does darcsden ssh-server build against the new ssh changes?
05:27 pointfree Okay it looks like ssh is constrained to 0.3.1 in darcsden.cabal so that's okay for now.
05:29 Heffalump I also won't upload to hackage until we've got all these changes integraeto a good stpoppopping point and got an pudate for darcsden
05:31 pointfree I'm not using the modexp function from ssh. Keys are generated by the curve25519 package. I haven't put much thought into what random number generator I use. The curve25519 cryptonite api has no mention of a rng: https://hackage.haskell.org/package/cryptonite-0.13/docs/Crypto-PubKey-Curve25519.html
05:33 Heffalump pointfree: I suspect your patch just depends on them textually
05:33 Heffalump i.e. changes the same or adjacent lines
05:34 pointfree Oh right
05:37 Heffalump I don't know what fr33domlover wants to do about those two patches, but you could just develop against a baseline that doesn't have them (the latest state of my ssh repo)
05:37 Heffalump it's easier to resolve conflicts later than to cut out dependencies
05:37 pointfree Yes
05:39 pointfree Also the libssh2 based tests need to be replaced with openssh shell commands. Currently my only way to test the kex is darcsden-ssh.
06:01 Heffalump sorry, just catching up on the discussion about libssh2. I think I actually looked at openssh for testing before I wrote the tests against libssh2 - the biggest problem is that the openssh server can't be configured to get users from a source other than the OS
06:01 Heffalump what's the actual problem with libssh2 here - it just doesn't support the new KEX either?
06:03 Heffalump fr33domlover: btw, if you could also update the CHANGES file I think that'd be quite useful for anyone else picking up your updates
06:04 pointfree yeah, libssh2 doesn't support the new kex (openssh uses libssh)
06:14 pointfree I'm taking the opportunity to add the new kex with cryptonite this time. Should be much cleaner this time. Anyone know how I might read an integer into a ByteArray?
06:18 pointfree Ah looks like it's writeByteArray
06:51 pointfree Heffalump, why do you need to test with the openssh server? Wouldn't it just test the openssh client against the haskell ssh server?
08:17 fr33domlover Heffalump, sure - but what's the next version number to use in the title?
08:17 fr33domlover pointfree, the 'ssh' code has bidi conversion between ByteString and Integer
08:18 fr33domlover you can use that for related string<->number conversions
08:19 inhortte joined #darcs
10:19 sm ssh-1.0 ?
11:43 Heffalump pointfree: oh yes, sorry. I must have been thinking of when we considered replacing ssh completely.
11:43 Heffalump fr33domlover: 0.4.0 is the current version planned for the next release in the cabal file
11:51 Big_G joined #darcs
12:06 BitPuffin joined #darcs
12:26 mizu_no_oto joined #darcs
12:39 burp joined #darcs
13:06 peb` joined #darcs
13:57 Riastradh joined #darcs
15:10 alexei_ joined #darcs
16:58 pbgc joined #darcs
17:12 pbgc Hi! I apologize if this question was already asked .. or if it's a newbie prob .. but I'm having problems with Darcs and OSX El Capitan. darcs diff allways gives me: "darcs: /usr/bin/diff: fileAccess: permission denied (Operation not permitted)". I had no problem with Maverics
17:12 Heffalump does /usr/bin/diff work from outside darcs?
17:12 Heffalump I haven't heard of it before
17:13 alexei_ joined #darcs
17:34 pointfree pbgc, I think that has to do with El Capitan's rootless "feature" http://www.infoworld.com/article/2988096/mac-os-x/sorry-unix-fans-os-x-el-capitan-kills-root.html
17:34 pointfree https://github.com/elm-lang/elm-package/issues/109
17:35 pbgc Heffalump: yes /usr/bin/diff works perfectly from outside darcs
17:37 pbgc pointfree: thanks. I will take a look. But I guess something will have to change in the darcs build for osx ... maybe darcs is trying to write to invoke diff inside /usr/bin .. I don't know
17:43 pointfree Looks like we may need to bump the unix package to at least unix-2.7.1.0 which is where this issue was fixed: https://github.com/haskell/cabal/issues/2653#issuecomment-174143161
17:46 pbgc Looks like the issue is related with GHC <  7.10 and it was fixed in 7.10. It's a matter of recompile darcs .. I'm using a pre-built binary
17:47 pbgc pointfree: sorry. wrote without reading you line ...
17:52 gh_ joined #darcs
18:00 pointfree ...Updating the mac that was in my closet to El Capitan so I can create a new prebuilt executable for Mac OSX with the bumped dependency.
18:09 alexei_ joined #darcs
18:31 pbgc pointfree: just finished building darcs 2.10.3 with stack and GHC 7.10.3 and can confirm the issued was fixed. It's working now
18:40 pointfree pbgc, good to know!
19:28 Riastradh joined #darcs
20:33 alexei_ joined #darcs

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