Posted by jpetso on May 31, 2009 at 8:31pm
Ok, so gathering #2 was largely technical and a bit unfocused at times, I won't post that one because it probably won't make such a good read if you haven't been in #drupal-vcs in the first place. Today's gathering #3 seemed more interesting to me, and focused largely on workflow issues. The core outcome (imho) is that the students will post short daily "micro-blogging" updates so that mentors stay up to date on the status even if no progress is made for any reason. Using a dedicated d.o issue and/or Twitter respectively identi.ca for these purposes. We also follow each other on Github now.
If all goes well, the students should be pushing out their first weekly updates by tomorrow evening, which will also be tracked in a d.o issue (as discussed earlier). Complete log of today's IRC meeting follows below.
[20:33] <jpetso> oh, hey, it's that time of week again
[20:34] <jpetso> chrono325, skiquel, sdboyer: hi all!
[20:34] <sdboyer> hola
[20:34] <Druplicon> hello
(...some discussion between jpetso, sdboyer and Druplicon that is not directly related to SoC...)
[20:41] <jpetso> chrono325, skiquel: are you there?
[20:41] <jpetso> marvil07 is also missing yet
[20:41] <sdboyer> damn interfaces are strict
[20:41] <sdboyer> LOVE IT
[20:42] <jpetso> strict, in how far?
[20:42] <skiquel> hey jpetso
[20:42] <jpetso> skiquel: heya
[20:42] <skiquel> now i want you to put ya hands in the air
[20:42] * chrono325 is here now, fails at using away.
[20:42] <skiquel> wave dem like you just dont care
[20:42] * jpetso puts both hands in the air
[20:43] <jpetso> i did it, really
[20:43] <jpetso> you can't check, but in the air they were
[20:43] <jpetso> hah.
[20:43] <chrono325> wow, I missed quite the party
[20:43] <jpetso> chrono325: that's what'll always happen when sdboyer is there and you aren't :P
[20:44] <jpetso> just kidding, we didn't really start yet
[20:44] <skiquel> o fortuna velut luna
[20:44] <chrono325> damn, I have to lurk on IRC more then.
[20:44] <sdboyer> jpetso: attachInputHandler(ProcInputHandler &$io, $descriptor) != attachInputHandler(ProcInputHandler &$io, $descriptor = 0);
[20:44] <skiquel> chrono325: look into irssi + bitlbee + screen
[20:44] <jpetso> sdboyer: heh
[20:45] <jpetso> sdboyer: on the other hand, it really only makes sense that way
[20:45] <jpetso> chrono325: or Quassel... which i need to try out too ;)
[20:45] <sdboyer> jpetso: f'sho. php's implementation of interfaces in general is kinda squirrely, though, i see some weird stuff happening, they're sometimes not as strict as i'd like
[20:45] <sdboyer> jpetso: so cases like this where it's MORE strict than i anticipated make me happy :)
[20:45] <jpetso> :)
[20:46] <jpetso> ok, getting on topic now that we're (nearly) complete
[20:46] <jpetso> chrono325: soooo....
[20:46] <chrono325> :)
[20:46] <jpetso> chrono325: how are you doing with your project?
[20:46] <jpetso> chrono325: i didn't really hear anything from you this week
[20:46] <jpetso> skiquel: or did you, and i missed the party?
[20:47] <chrono325> jpetso: good. I've finally gotten caught up on the apis and stuff and got coding for realz
[20:47] <skiquel> chrono325 was asking me about hooks in svn lib
[20:47] <chrono325> jpetso: I finished (but haven't really tested) xsvn-pre-commit.php
[20:47] <skiquel> and if he should be using prop_ / lock/ pre/ post
[20:48] <chrono325> and I didn't think we really needed to track revprop or lock changes.
[20:48] <jpetso> agreed
[20:49] --> marvil07 has joined this channel (n=marvil07@64.76.110.180).
[20:49] <jpetso> chrono325: although revprop might make sense because otherwise commits might be completely empty for Version Control API
[20:49] <jpetso> hi marvil07!
[20:49] <marvil07> hi all
[20:51] <chrono325> jpetso: I don't follow. revprop changes are disabled by default (except for svnadmin), so how could commits be empty if we don't implement revprop hooks?
[20:51] <skiquel> haha
[20:51] <skiquel> hey marvil
[20:51] <skiquel> I feel lightheaded
[20:52] <jpetso> chrono325: oh. then i must be mixing up revprop with those other props that you can assign ("svn:executable" and the likes)
[20:53] <chrono325> jpetso: yeah, it's a little weird, but subversion considers those kinds of properties just another part of a commit. There is a separate facility to i.e. change a commit message after a commit has been made.
[20:54] <jpetso> chrono325: yeah. sorry, i always mix those up
[20:54] <chrono325> jpetso: those are "unversioned properties", since they don't enter the repository through a new revision
[20:54] <jpetso> chrono325: which is weird because they're named "revprop", that confused me. but yeah, we don't need to track those
[20:55] <chrono325> jpetso: yeah, they are weird.
[20:56] <jpetso> chrono325: so, i guess the schedule doesn't totally match anymore then, as i estimated that start of SoC == start of coding
[20:57] <skiquel> yeah
[20:57] <skiquel> is marvil coding?
[20:57] <chrono325> jpetso: yeah, I misjudged how long it would take me to gain competence with the drupal APIs and stuff.
[20:57] <jpetso> chrono325: well yeah, that's hard to estimate i guess
[20:58] <jpetso> chrono325: any suggestion on how to go about the schedule? should we drop something, or just compact some parts and hope it'll work out anyways?
[20:58] <chrono325> jpetso: Well, there was a week at the end which was left open as "flex time".
[20:59] <chrono325> jpetso: And if necessary, we could drop the hg hooks or bzr backend
[20:59] <marvil07> skiquel: just starting... this week was really bad for me
[20:59] <jpetso> chrono325: ok, so we'll make use of the "flex time" first then, right?
[21:00] <chrono325> jpetso: sounds good. Sorry to falter out of the gate, but it's always the "getting started" part of things that's hard for me. Once I get in the groove, I can cruise.
[21:01] <jpetso> chrono325: estimates are hard, that's ok imho. looking forward to seeing you cruising :D
[21:02] <chrono325> jpetso: me too :) much more fun than staring at a giant wall of undeciphered code.
[21:02] <jpetso> chrono325: could you please add the midterm evaluation and pencils-down dates from the official SoC timeline to the wiki schedule? i can't remember those dates really, and those are important ones
[21:02] <jpetso> ok, so.
[21:02] <jpetso> i feel like i missed out on stuff this week.
[21:03] <marvil07> jpetso: not much about me :(
[21:03] <chrono325> jpetso: July 6: Mentors and students can begin submitting mid-term evaluations.
[21:04] <chrono325> jpetso: July 13: Mid-term evaluations deadline;
[21:04] <jpetso> chrono325: thanks :)
[21:04] <jpetso> apart from marvil07's issue updates at the start of the week and chrono325's new g.d.o discussion, this week i was pretty much totally in the dark about the status of your projects
[21:04] <jpetso> which is worse than just knowing that nothing happens, or non-coding stuff happens
[21:04] <jpetso> so, ...
[21:05] <jpetso> would it be ok for you to establish some "micro-blogging" daily status updates, where you state what happened or whether nothing happened?
[21:06] <jpetso> no idea if the right platform for that is d.o, or g.d.o, or maybe identi.ca or Twitter
[21:06] <jpetso> just so that i know what is or is not going on.
[21:07] <jpetso> marvil07, chrono325: what do you think?
[21:07] <chrono325> August 17th: "Firm pecils down"
[21:07] <marvil07> excellent idea jpetso
[21:08] <marvil07> twitter :D
[21:08] <skiquel> chrono325: yeah
[21:08] <skiquel> you probably should do the same :)
[21:08] <chrono325> jpetso: Makes sense
[21:08] <skiquel> you can hook your twitterfeed/blog into planet-soc
[21:09] <marvil07> marvil07 in all places :D, so twitter too
[21:09] <jpetso> skiquel: :D
[21:09] <jpetso> cool, then let's do that
[21:10] <chrono325> I haven't been quite sure about how active I can be in i.e. filing bugs for the projects and asking about random little things.
[21:11] <chrono325> I have some notes I've taken about things I saw or thought should be fixed, but I didn't know if it was "following protocol" to report every last little thing.
[21:11] <jpetso> chrono325: yeah, good point, i thought about that too
[21:11] <sdboyer> twitter or blogs. either way's good for me
[21:12] <jpetso> chrono325: essentially, that's the price we pay not opening up separate SoC-related projects and reusing versioncontrol for such stuff
[21:12] <chrono325> I suppose it would be better to err to the side of too much reporting, and then learn what the boundaries and expectations are.
[21:13] <jpetso> chrono325: right, better more communication than too little
[21:13] <jpetso> i'm unsure about how far we should take the versioncontrol issue queue for stuff like that
[21:13] <jpetso> some possibilities:
[21:13] <chrono325> jpetso: Yeah. I just don't have a good sense of what the expected channels of communication are, other than IRC.
[21:14] <jpetso> a) we open up one more meta-issue for each of you two, and use it for all the minor stuff (also the "micro-blogging" updates)
[21:14] <jpetso> ^ my favorite
[21:15] <jpetso> b) we use separate issues for all of those little things, and close them with "closed" instead of "fixed"
[21:15] <sdboyer> chrono325: err on the side of too much communication, for sure. worst case scenario is the reporting gets redirected to some other place, which really just means "our communication gets clearer"
[21:15] <sdboyer> (which in no way contravenes what jpetso is saying about approaches here :P)
[21:16] <jpetso> c) like (a), but with a g.d.o discussion instead - pro: threading, contra: spamming other people
[21:16] <chrono325> sdboyer: Okay. So I should add the things I have to the versioncontrol_svn issue queue?
[21:16] * sdboyer agrees with jpetso on a) being optimal
[21:16] <sdboyer> chrono325: ja.
[21:17] <jpetso> chrono325: if it concerns a single feature, the issue queue is always the right place for something
[21:18] <jpetso> i guess you could consider "hook scripts" a single feature.
[21:18] * sdboyer flips proc_open() the bird
[21:18] <sdboyer> additionally, if we find that there are enough clustered issues, we can always add a new component to the project
[21:18] <jpetso> right
[21:19] <marvil07> a +1
[21:19] <jpetso> nice, then let me open two more issues
[21:20] <jpetso> ...which reminds me, there's another meta-issue per student which should get updated weekly ;)
[21:20] <jpetso> let's get the first weekly update in there until tomorrow evening, k?
[21:21] <chrono325> Hmmm, it appears that I cannot add issues to versioncontrol_svn
[21:22] <jpetso> chrono325: wah?
[21:22] <marvil07> ok jpetso
[21:22] <sdboyer> chrono325: http://drupal.org/node/add/project-issue/versioncontrol_svn doesn't work?
[21:22] <jpetso> marvil07: thanks!
[21:23] <chrono325> ahh, well, I didn't see a link to it on the project page...
[21:23] <jpetso> chrono325: the link is on the issue queue page
[21:23] <chrono325> jpetso: so it is.
[21:25] <sdboyer> they took the link ot add a new issue off the project page because they figured it was resulting in WAY too many duplicate issues
[21:25] <sdboyer> (since people would just add a new one instead of actually checking existing ones)
[21:25] * sdboyer dunnos about the real logic of that, but
[21:25] <sdboyer> not my call
[21:26] * jpetso agrees on that logic
[21:26] <jpetso> people should always have a look at the issue queue before posting new issues
[21:26] <jpetso> anyways.
[21:26] <chrono325> I see
[21:27] <jpetso> marvil07: so, why was it a bad week for you? caused by problems in personal life, or just something related to the current issue that didn't work out yet?
[21:28] <marvil07> jpetso: on the issue http://drupal.org/node/471772 I see instruction to post to gdo ? so.. I thought (a) option does not spam gsoc group
[21:28] <Druplicon> http://drupal.org/node/471772 => Meta-issue: Weekly updates of Marco Villegas (marvil07) => Version Control API, Documentation, normal, active, 1 IRC mention
[21:28] <jpetso> marvil07: that's for weekly updates, the SoC administration is interested in those
[21:29] <jpetso> daily is probably too much, though
[21:29] <jpetso> for them. not for us.
[21:29] <marvil07> ok
[21:29] <marvil07> so I need to follow your accounts on twitter guys
[21:30] * jpetso doesn't yet have a twitter account :-o
[21:30] * jpetso remembers his employer urging us workers to open such accounts
[21:30] * marvil07 knows that, I searched you so many times :D
[21:31] <sdboyer> sdboyer for me
[21:32] <jpetso> marvil07: i'll be there soon, but i first need to equip my website with OpenID delegation so that I can use jakob.petsovits.at as OpenID domain (even if it actually uses myopenid.com)
[21:32] <jpetso> marvil07: then, i'll create an identi.ca account, and then I'll forward it to twitter
[21:33] <jpetso> step by step. but hey, i've already created that OpenID account on myopenid.com at least :P
[21:33] <marvil07> :-D, I see why you does not have one now
[21:33] <marvil07> no prob, if you follow me I inmeddiatly gong to add you
[21:33] <marvil07> and you chrono325 ?
[21:33] <jpetso> ok, i'll remember to do that
[21:34] <chrono325> I think I registered on identi.ca a while back
[21:34] <chrono325> I haven't used either twitter or anything similar, but now's a good a time as any to start
[21:34] <jpetso> skiquel: on Github, if someone commits code to a watched repository, do i get notified on that?
[21:35] <marvil07> what a geek image sdboyer :D
[21:35] <sdboyer> marvil07: :P
[21:35] <skiquel> jpetso: unknown.
[21:35] <sdboyer> woah. unit tests pass.
[21:35] <skiquel> :)
[21:35] <skiquel> because i never wathced
[21:35] <sdboyer> jpetso: well if you watch svnlib, you can find out right now :)
[21:35] <skiquel> However, we should add each other, I am www.github.com/skiquel
[21:35] <skiquel> please paste your github URL's
[21:35] <skiquel> so i can view
[21:36] <sdboyer> http://www.github.com/sdboyer , though you've got mine
[21:36] <skiquel> please add me back, nothing worse than a recipricol deviant that ignores you
[21:36] <jpetso> that reminds me, i'd need to read all conversations in there, even the ones that don't affect me at all (and it seems there are some of those)
[21:36] <skiquel> sociology 101
[21:37] <jpetso> how did the thing with hashtags work again?
[21:37] * marvil07 thinks you guessed mine: http://github.com/marvil07
[21:37] <sdboyer> skiquel: :P
[21:37] <sdboyer> dun did
[21:38] <chrono325> http://github.com/chrono325
[21:39] <chrono325> sdboyer: is the "svnlib" on github the same as "new_svnlib" in the versioncontrol_svn?
[21:39] <sdboyer> chrono325: yeah, although that version is outdated
[21:39] <chrono325> sdboyer: the github one or the drupal one?
[21:39] <sdboyer> chrono325: now that i'm actually nearing another release point, i'll do that
[21:40] <sdboyer> chrono325: the drupal one. all actual svnlib development is done on github
[21:40] <skiquel> jpetso start following us please
[21:40] <sdboyer> chrono325: i've just been syncing to drupal periodically
[21:40] <chrono325> sdboyer: got it. So I should report svnlib issues directly to GitHub?
[21:40] <jpetso> skiquel: mmkay
[21:40] <sdboyer> chrono325: yup
[21:41] <marvil07> well guys, I gtg now, and at today's night I goona work on code
[21:41] <sdboyer> chrono325: with the next sync, i'll also probably move it out of the "new_svnlib" spot into just "svnlib"
[21:41] <jpetso> marvil07: yay :D
[21:41] <sdboyer> marvil07: g'luck :)
[21:42] <marvil07> :D
[21:42] <-- marvil07 has left this server ("out").
[21:42] * sdboyer grumbles
[21:42] * sdboyer is gonna have to write a LONG README entry on how to use proc handlers without either causing a PHP hang or getting no data back
[21:43] <jpetso> sdboyer: when development is done on Github primarily, how does syncing work? i guess you don't use your git scripts for that one and just copy files manually, don't you?
[21:43] <chrono325> btw, my identica account is http://identi.ca/hax
[21:43] <sdboyer> jpetso: yeah, esp. because there are often things i need to change a bit in order to make it work with versioncontrol_svn itself
[21:43] * sdboyer is actually probably going to start maintaining a separate branch just for the sync
[21:47] <jpetso> sdboyer: oh, btw, didn't you intend to push the new svnlib to a different branch on d.o CVS?
[21:47] <jpetso> oh wait, need to check...
[21:48] <jpetso> sdboyer: k, sorry, i forgot that cvs.module doesn't tell me when branches are created :P
[21:48] <jpetso> ts.
[21:48] <jpetso> chrono325: i'll follow you there as soon as i've got my identi.ca account set up
[21:49] <jpetso> skiquel: what does it mean on Github to follow other people? i can't seem to find information on that at first glance... how do i get notified?
[21:55] <jpetso> whatever, i'll just subscribe to the feed and let go of email notifications if there might be any
[21:55] <chrono325> jpetso: Where should my status update go? As a wiki page on g.d.o or as an update to my project wiki page?
[21:56] <jpetso> chrono325: the weekly update?
[21:56] <chrono325> jpetso: yeah, sorry should have clarified
[21:56] <jpetso> chrono325: that one should be a g.d.o discussion
[21:56] <jpetso> chrono325: (discussion rather than wiki, so that people can comment
[21:56] <jpetso> )
[21:57] <chrono325> jpetso: a new discussion for each week's updates, or one discussion for my whole project?
[21:57] <jpetso> chrono325: one for each week, please
[21:57] <jpetso> chrono325: as long as it's more than just one or two paragraphs, which i hope it will be :)
[21:58] <chrono325> jpetso: alright. I'll get on that. I've got to get some food before my fingers fall off :)
[21:58] <jpetso> chrono325: r0xx0rz
[21:58] <jpetso> chrono325: i'll create the d.o short-notice issues in the meantime
[21:59] <chrono325> jpetso: sounds good
[22:13] <jpetso> ok, i'll declare this week's meeting as finished. push out code and mini-updates to your heart's content!