IRC Logs for #circuits Tuesday, 2015-05-05

*** RNeville has joined #circuits00:54
prologichttp://stackoverflow.com/questions/30042360/echoing-messages-received-through-udp-back-through-a-different-tcp-port00:56
prologicsomeone answer this question00:56
prologicbut with a circuits based solution :)00:56
prologichttps://gist.github.com/prologic/ae987fe4f6fbbd9ca41501:13
spaceonei am writing01:30
spaceonehm01:31
spaceonei already wrote also one :D01:31
spaceoneprologic: he don't want to broadcast01:35
spaceonehe just want to proxy01:35
spaceone4 am now01:38
prologichttp://stackoverflow.com/questions/30042360/echoing-messages-received-through-udp-back-through-a-different-tcp-port/30042906#3004290601:42
prologicthere01:42
prologicI answered it01:42
prologicand tested it01:42
prologicanyone with SO accounts pleaes upvote :)01:42
prologichaha01:42
prologicspaceone: pretty sure hw ants to broadcast01:43
prologicyou can't proxy the incoming UDP out01:43
prologicwho do you proxy it to? :)01:43
prologicwhich connected client? :)01:43
prologicin any case01:43
prologicthe exact details are up to the OP :)01:43
prologicit's not hard to change it from a broadcast01:43
prologicto specifically writing to a connected client01:43
prologicwoot :)01:49
prologic+1001:49
prologichaha01:49
prologicthis is such selfless self promoting :)01:49
prologicright? :)01:49
pdurbinnothing wrong with a little self promotion01:52
prologichaha01:53
prologicI edited the response to mention Twisted at least01:53
prologicit would be interesting to see a similar solution based on Twisted from someone well adverse in Twisted01:53
prologicand that's not really me :) sadly01:53
prologicI know enough about Twisted internals not to use it actively01:53
prologicbut enough to actually integrate Twisted protocols and apps on top of circuits01:54
prologicusing circuits as the driver (event loop)01:54
prologicor in Twisted terms -- circuits as the reactor01:54
*** RNeville has quit IRC03:59
*** qwebirc86127 has joined #circuits07:47
spaceoneprologic: btw. why do we always have the 'merge' commit when you accept a pull request? can't the commits just be cherry-picked?09:24
prologicgood question09:25
prologicI *thought* Github has fast-forward merges?09:25
prologici.e: merge-rebase09:25
prologicMaybe it's an option you have to select when "clicking the Merge button"?09:25
spaceonei don't know, let's have a closer look when doing the next PRmerge09:29
prologicagreed :)09:31
prologicit does make the commit tree a bit cleaner09:31
pdurbinspaceone: probably you would have had to update your PR (rebase the latest underneath it) to avoid a merge commit10:26
pdurbinalthough... come to think of it...10:26
pdurbinmaybe github always make a merge commit for PRs10:27
pdurbin"Pull requests are merged using the --no-ff option." https://help.github.com/articles/merging-a-pull-request/10:28
pdurbinso yeah, always a merge request if you use the github web interface10:28
spaceonepdurbin: i always do a rebase before PR10:36
spaceonePull requests are merged using the --no-ff option.10:37
spaceonehttps://help.github.com/articles/merging-a-pull-request/10:37
pdurbinspaceone: that's what I just said :)10:54
prologicdamn11:35
prologicI *thought* Github had an option to do --fast-forward merges?11:35
prologichttp://stackoverflow.com/questions/9994093/automatic-merge-of-pull-requests-on-github-without-the-merge-bubble11:37
prologicoh well11:37
spaceoneotherwise can we remove the merges afterwards?11:48
prologicI don't think so11:51
prologicwell in theory yes11:51
prologicbut it would require history rewriting and force push11:51
prologicuggh11:51
prologicbut otoh do we really care that much?11:51
prologicwe could just do the merges in future by hand via the cli (fast forward)11:51
prologicor not care that much really and just hit the button :)11:51
spaceone^^11:52
pdurbinI wonder why GitHub does this. Maybe they think it's best to easily see who performed the merge.12:28
*** qwebirc86127 has quit IRC17:37
prologicprobably20:43
spaceoneprobably20:44
pdurbinprobably20:59
spaceoneprologic: what do you think about circuits.net.socket as a new package containing socket improvements with API change?21:33
spaceonemultiplexing things reqires socket as event argument21:34
*** Zimsky-- has joined #circuits22:28
*** pdurbin1 has joined #circuits22:28
*** Zimsky has quit IRC22:30
prologicpoo22:38
prologicthat question got voted down22:38
prologicand someone more familar with Twisted answered :)22:39
prologicspaceone: Don't we already have circuits.net.socket?22:39
prologicor you mean socket as in singular?22:39
spaceoneprologic: i mean singular22:45
prologicyeah yeah go for it22:47
prologicsounds good22:47
prologicthen perhaps we can later replace TCPServer, TCPClient, etc22:47
prologicwith factory functions that use Socket()22:47
spaceonegood night ;)22:48
pdurbinprologic: https://github.com/openhatch/oh-mainline/issues/1515 was just closed. Were you going to add circuits to OpenHatch or was I?22:59
prologicI think/believe you were :)23:01
pdurbinok, opened an issue for this: https://github.com/circuits/circuits/issues/2923:04
spaceoneprologic: what is the merge policy?23:28
prologicJust that are we going to try to do fast forward merges from here on23:29
prologicor just hit the merge pull request button?23:29
prologicwhat we talked about last night23:29
spaceoneah, i remember23:31
prologicpersonally I kind of sit on the fence23:32
prologicon one hand it's good to cleanup your commits "sometimes"23:32
spaceonewell, badger/curl is cherrypicking everything by hand23:32
prologicbut I don't strongly believe in squashing to a single commit23:32
prologicbut in terms of keeping merge commits out of the tree - *meh* no opinion on that23:33
prologicyeah23:33
prologicit's just that it's more work if we do it from the CLI ofc23:33
spaceoneyes23:33
prologicdo the merge commits pose any down sides per se?23:33
prologicthey *do* act as "merge points" of when something was merged23:33
prologicso they are useful somehwat23:33
prologiceven if noisy23:33
spaceonethey are disturbing when i use git log --grep, git show -n $x etc.23:34
prologicis there no option to exclude merge commits?23:34
spaceonealso rebasing is harder with one merge commit23:34
spaceonei don't think so23:34
prologichmm23:34
prologicdamn23:34
spaceonenot via github23:34
prologicMercurial has such an option :)23:34
spaceone^^23:35
prologic -M --no-merges           do not show merges23:35
prologichaha23:35
prologichg log -M23:35
prologicoh well :)23:35
prologicmaybe in a later version of Git?23:35
prologicrebasing is harder with merge commits in the ttree?23:35
spaceonegit log --no-merges23:36
prologicok so here's the thing23:36
prologiclet's open an issue about this23:36
prologicand discuss it more openly23:36
prologichopefully see what others might think23:36
prologicwe've already seen a couple of new contributors in the past week or so23:36
prologicgood to see :)23:36
spaceonewell if i do an interactive rebase with a merge commit HEAD~2 will contain every commit of that merge commit23:37
prologicI don't mind either way ofc23:37
spaceoneinstead of 2 commits23:37
prologicbut if we do the whole manual merging of prs23:37
prologicwe should create a set of fabric tasks for this23:37
prologicto make it easier to manage23:37
prologicor find a tool that helps manage this23:37
prologicre merge commits and rebase -- ahh I see23:37
prologicyes that makes sense23:37
spaceonefor my commits it should be easy, you can add my repo as remote repo named "spaceone". and then git rebase spaceone; git push +master;23:38
prologicyeah yeah23:39
prologichttp://pythonhackers.com/p/whilp/git-pulls23:42
prologicsomething like this I guess23:43
prologicmaybe there's a Python one:)23:43
spaceonei'll go sleeping now ^^23:45
prologickk23:45

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!