Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2017-12-30

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

All times shown according to UTC.

Time Nick Message
02:58 ilbot3 joined #darcs
02:58 Topic for #darcs is now http://darcs.net/ | logs: http://irclog.perlgeek.de/darcs/ | darcs 2.12.5 is out http://darcs.net/Releases/2.12
09:14 lelit hi there! Trying to connect to my account on hub.darcs.net (after a looong time...), I get something like
09:15 lelit ssh lelit@hub.darcs.net
09:15 lelit Unable to negotiate with 173.255.219.222 port 22: no matching cipher found. Their offer: aes256-cbc,aes192-cbc,aes128-cbc
09:15 lelit
09:16 lelit I had the "KexAlgorithms +diffie..." in my ~/.ssh/config and I removed it as the FAQ tells, but that didn't change the error
09:24 lelit running "ssh -Q cipher" on my local machine definitely lists all those aes algos
09:24 lelit any hint?
13:36 gh_ joined #darcs
13:41 leg joined #darcs
14:06 pointfree lelit: You could add "Ciphers +aes256-cbc,aes192-cbc,aes128-cbc" to ~/.ssh/config where "KexAlgorithms +diffie..." was previously.
14:07 leg joined #darcs
14:10 pointfree darcshub's ssh server now has support for a newer kex, but since then some additional ciphers got deprecated in openssh.
14:14 lelit pointfree: thanks, I will try that later
14:27 stef204 joined #darcs
15:51 lelit pointfree: great, that fixed it
15:52 pointfree :)
15:52 lelit should the faq mention that?
15:52 pointfree lelit: It should.
15:53 pointfree sm: ^
16:29 sm hi lelit, pointfree. What should change in the FAQ ?
16:30 lelit mention that [on some systems] an explicit "Ciphers +aes..." may be needed
16:30 lelit not sure why on my Debian sid I had to do that, even if "ssh -Q cipher" lists them
16:33 sm so I think the last "Prior to" paragraph at https://hub.darcs.net/#how-do-i-configure-ssh-access needs update
16:33 sm what's the ... in aes... ?
16:34 lelit yes, seems appropriate
16:34 pointfree "Ciphers +aes256-cbc,aes192-cbc,aes128-cbc"
16:35 pointfree sm: ^
16:35 lelit given a patch, how can I determine the closer succeding tag that includes it?
16:36 Heffalump lelit: there's no guarantee that's unique
16:36 lelit more specifically, I'm trying to lookup in which version the "darcs query" subcommand has been dropped as an alias for "darcs show"
16:37 lelit I found the patch
16:37 lelit patch 3f1c73e0ca880ee2ec4098ae33ece7f0ccbd5af8
16:37 lelit Author: Guillaume Hoffmann <guillaumh@gmail.com>
16:37 lelit Date:   Mon Apr 18 20:36:59 CEST 2016
16:37 lelit * remove all hidden aliases of show, deprecated for years
16:37 lelit
16:37 sm https://hub.darcs.net/#how-do-i-configure-ssh-access text updated
16:37 lelit but there's no mention of it in the changelog, it seems
16:38 pointfree Thanks sm!
16:38 sm thank you
16:47 pointfree darcs test --linear "grep -qir 'darcs query'" | grep -B4 '^Success!$' # the patch that removed "darcs query"
16:52 lelit :)
16:56 lelit I used "darcs log src/Darcs/UI/Commands/Show.hs", then a manual browse of "darcs log" searching for that hash, then going backward (in the log output, forward in time) looking for "^\s+tagged" revealed that the alias has been dropped in 2.12
17:00 leg joined #darcs
17:05 pointfree There are a number of haskell implementions of the newer ciphers. Maybe I can look into when I have more time..
17:07 pointfree A lot of darcs users (and git users) are switching to pijul as it continues to improve.
17:08 pointfree Some darcs users (such as Ben) still prefer darcs over pijul.
17:10 lelit are there specific reasons, or is it just because there are lots of missing functionalities?
17:10 pointfree Is there a document explaining what darcs' role should be once pijul is a complete alternative to git or darcs?
17:12 pointfree lelit: I think there are reasons why some users prefer darcs patch theory over pijul patch theory.
17:13 pointfree (so it's not just about current feature parity)
17:15 sm pointfree: is that true ? I haven't seen any noise about pijul
17:19 pointfree sm: http://irclog.perlgeek.de/darcs/search/?nick=bfrk&amp;q=pijul
17:21 pointfree Ben Franksen also posted some critique to the darcs mailing list but I'm having trouble finding it right now.
17:22 sm I don't see "a lot of darcs users (and git users).. switching to pijul" in that..
17:23 sm I see their IRC channel has 48 vs #darcs 30 right now though
17:24 pointfree sm: The #pijul channel has more people and activity than #darcs and many of those in #pijul have never used darcs.
19:41 Heffalump pointfree: I think it was basically that the pijul implementation is inpenetrable
20:22 pointfree Heffalump: yeah. I'll often propose an idea on pijul channel or forum but right now I'm all talk and no code :( because I want to fully understand the pijul patch theory implementation first.
20:22 pointfree Pijul is only based on that paper although there is that very nice series of high level blog posts on the pijul theory. I mostly blame myself.
20:50 pointfree Many people wouldn't mind seeing all these language specific package managers drawn-and-quartered in Times Square.
20:51 pointfree ...Of course I started writing one for my (interactively-compiled) concatenative dataflow language for embedded arm microcontrollers.
20:51 pointfree Then I thought: this dichotomy between package managers and dvcs is an unnecessary one, possibly even historical or political baggage.
20:51 pointfree So now I'll pivot this into a language specific dvcs. A dvcs integrated into the language by extension understands that language and doesn't need to be so general purpose as to work with opaque data.
20:52 pointfree Such a constraint could simplify the problem and the experience enough for it to fit handily into a 256KiB microcontroller and maybe even my head.
22:16 Heffalump Pijul has a much better theory than darcs, at least for handling conflicts in text files
22:17 Heffalump darcs is perhaps simpler in the way it defines patches, giving more possibilities in general (but realistically is badly hamstrung by its conflict handling)

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