16:00:07 <geppetto> #startmeeting fpc
16:00:07 <zodbot> Meeting started Thu Apr 17 16:00:07 2014 UTC.  The chair is geppetto. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:07 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:07 <geppetto> #meetingname fpc
16:00:08 <geppetto> #topic Roll Call
16:00:08 <zodbot> The meeting name has been set to 'fpc'
16:00:15 * RemiFedora here
16:00:20 * pmackinn here
16:00:33 <tibbs|w> Hello.
16:00:34 <geppetto> #chair geppetto RemiFedora
16:00:34 <zodbot> Current chairs: RemiFedora geppetto
16:00:40 <geppetto> #chair tibbs|w
16:00:40 <zodbot> Current chairs: RemiFedora geppetto tibbs|w
16:01:16 <geppetto> SmootherFrOgZ: you around?
16:02:15 <SmootherFrOgZ> yep
16:02:16 <SmootherFrOgZ> here
16:03:00 <geppetto> #chair SmootherFrOgZ
16:03:00 <zodbot> Current chairs: RemiFedora SmootherFrOgZ geppetto tibbs|w
16:03:09 <geppetto> one more
16:03:36 <geppetto> spot: abadger1999: Either of you around?
16:04:00 <geppetto> pretty sure both are traveling or afk.
16:04:06 <tibbs|w> abadger1999 joined about 90 minutes ago from pycon.
16:04:13 * geppetto nods
16:04:52 <geppetto> the other three aren't int he channel … which doesn't bode well.
16:05:22 <geppetto> pretty soon I'm going to have a complex about running the meeting ;)
16:06:24 <SmootherFrOgZ> eh!
16:06:37 <tibbs|w> Yeah, this time of year is always tough.  We're going to have to do business in the tickets if we want to make any progress at all.
16:09:31 <geppetto> pmackinn: you are here for 415?
16:09:35 * pmackinn is here for https://fedorahosted.org/fpc/ticket/415 and https://fedorahosted.org/fpc/ticket/422 as time permits
16:09:53 * geppetto nods … 415 seems fairly simple
16:10:05 <geppetto> #topic #415     Temporary Javascript bundling exception for Ambari dependencies
16:10:10 <sagitter> please, before postpone again your meeting, could you go on in some way with https://fedorahosted.org/fpc/ticket/391#comment:15 ?
16:10:15 <geppetto> https://fedorahosted.org/fpc/ticket/415
16:11:03 <geppetto> sagitter: the page change?
16:11:08 <RemiFedora> as web-asset still unusable, and as we have already various exception for jquery, I'm +1 for one more
16:11:16 <geppetto> sagitter: That's not done during the meetings
16:11:30 <pmackinn> what, in general, is the state of javascript bundling? seeing other reviews go through with bundled js (not just jquery)
16:11:40 <geppetto> RemiFedora: Yeh, seemed pretty simple +1 for a release or two for web-assets
16:12:43 <pmackinn> e.g., https://bugzilla.redhat.com/show_bug.cgi?id=1061056
16:13:18 <geppetto> I think we've generally allowed JS bundling anyway
16:13:43 <geppetto> but we'd still prefer you use web-assets eventually, if you can.
16:14:02 <geppetto> tibbs: SmootherFrOgZ Want to vote?
16:14:05 <RemiFedora> pmackinn, there is a Guidelines which imply unbundling for JS, but which is not usable for now
16:14:28 <RemiFedora> and there is so much issue with JS... especially the minify stuff...
16:14:33 <tibbs|w> It's going to take a while before enough dependencies get in the distribution before we can really think of usefully unbundling javascript everywhere.
16:14:51 <geppetto> Rathann: You around?
16:14:53 <Rathann> hi, sorry for being late
16:14:57 <geppetto> #chair Rathann
16:14:57 <zodbot> Current chairs: Rathann RemiFedora SmootherFrOgZ geppetto tibbs|w
16:15:01 <RemiFedora> and there is nothing to ensure soem API stanbility with JS
16:15:04 <geppetto> woo, 5 ppls
16:15:17 <RemiFedora> quorum \o/
16:15:27 <geppetto> Rathann:  https://fedorahosted.org/fpc/ticket/415
16:15:29 <tibbs|w> I'm generally +1 to javascript exception requests now except when dependencies are actually in the distro.
16:15:31 <SmootherFrOgZ> +1 from me
16:15:37 <tibbs|w> So +1 on 415 from me.
16:15:38 <pmackinn> great, so what can i tell my pkg reviewer today? does #415 get closed with the omnibus exception approved?
16:15:53 <geppetto> pmackinn: if Rathann votes +1, then yes
16:16:22 * pmackinn slips Rathann an envelope stuffed with cash
16:16:43 * RemiFedora wants one to not change his vote
16:17:03 * pmackinn looks for his wife's purse
16:17:07 <tibbs|w> I can trivially be bribed with Dr. Pepper.
16:17:20 <SmootherFrOgZ> tibbs|w: :)
16:17:26 <Rathann> +1 from me for temp exception
16:17:33 <geppetto> #accepted
16:17:54 <pmackinn> fantastic, thanks all
16:18:51 <tibbs|w> So it's true; we can get something done.
16:19:04 <geppetto> #info Temporary Javascript bundling exception for Ambari dependencies PASSED (+1:5, 0:0, -1:0)
16:20:09 <geppetto> #topic #412     Please change the packaging guidelines to include packaging policy regarding systemd timer units
16:20:29 <geppetto> notting: ping
16:20:36 <notting> i'm here
16:20:55 <geppetto> So … you want us to change the wording back to MUST, can you explain the reasoning?
16:21:28 <notting> sure. (although, to be fair, i missed the original discussion)
16:22:21 <notting> a lot of it is about avoiding two-way dependency creep - if the app already depends on systemd, it should be using the system facility for this. if it *doesn't* depend on systemd, it shouldn't grow a systemd dependency just for a timer unit without using systemd for the rest of its needs
16:23:17 <tibbs|w> Is it simpler to just state that as "don't depend on both systemd and cron"?
16:23:18 <RemiFedora> notting, you use "should" ;)
16:23:22 <geppetto> notting: One thing that was brought up is that systemd timer == default off, and cron timer == default on.
16:23:41 <notting> RemiFedora: fine, "mustn't"
16:24:10 * RemiFedora agrees with tibbs|w proposal
16:24:11 <notting> geppetto: that can be controlled with presets, which can be changed
16:24:30 <geppetto> notting: but not within the package
16:24:34 <tibbs|w> Actually I like the idea that timers default to off.
16:24:50 <geppetto> notting: What do you think of tibbs|w proposal?
16:24:56 <notting> geppetto: and it's not really 'default to off', it's 'default to whatever the preset is'
16:25:12 <tibbs|w> It wasn't really a proposal; I was just trying to make sure that was the actual reasoning.
16:25:45 <tibbs|w> If so, I'm behind it, because it seems kind of dumb to require systemd around but then still use cron.
16:26:41 <notting> tibbs|w: yes
16:27:49 <geppetto> I can kind of see the first one … if you are already depending on systemd, you have to suck it up and drink all the kool aid.
16:28:16 <geppetto> But what's the rationale for not allowing people to move to systemd.timer files even if they don't dep. on systemd in other ways?
16:28:21 <notting> tibbs|w: that's the intent - to consider it part of the systemd conversion (even though it's coming a bit later than the initial service units); doesn't make sense to do it partially
16:29:00 <geppetto> So … I'm happier with adding something like tibbs|w's wording … don't explicitly dep. on both
16:29:25 <geppetto> As that allows someone to migrate if they want … or is the intention that people shouldn't do that for some reason?
16:30:02 <tibbs|w> I think the general idea is that we want to move away from cron simply to remove the dependency from base installs.
16:31:21 <tibbs|w> Not that it's much of a dependency.  But as an admin I welcome not only having a single way to deal with all of the ways things can get started on a machine, but actually being to control "cron jobs" without having to edit the files.
16:31:31 <notting> geppetto: because cron  & timers are not a complete 1-for-1 swap, the idea is not "replace cron with systemd everywhere", it's "replace it where systemd is in use and it makes sense"
16:32:12 <geppetto> Ok, so how about changing the first to a MUST and changing the second from "Packages which do not already depend on systemd" to "Packages which already depend on cron", and using MUST NOT?
16:32:15 <tibbs|w> I need to read further to see how this interacts with /etc/cron.daily and the like.
16:32:42 <geppetto> notting: right … that's another thing … it doesn't say anything about what to do if what you want can't be done using systemd
16:32:57 <notting> geppetto: i believe the issue with that second bit is that a bunch of packages with cron jobs are missing the cron dependency ATM
16:33:03 <geppetto> notting: but for the cases they overlap, the above is the intention, right?
16:33:08 <tibbs|w> Heh, yes, there is that.
16:33:17 <geppetto> facepalm
16:33:28 <notting> (it's part of the change proposal to fix that as well)
16:33:53 <geppetto> yeh, I think we should ignore that and assume packages have the right deps. … at least for policy :)
16:35:28 <RemiFedora> I still don't see any reason to not allow someone to switch from crontab to timer
16:35:56 <geppetto> RemiFedora: that is allowed
16:35:58 <notting> RemiFedora: i don't see it as disallowing that
16:36:18 <RemiFedora> "Packages which do not already depend or require systemd MUST NOT use timer units"
16:36:23 <geppetto> Ok, I altered the page … is that fine notting ?
16:36:34 <geppetto> RemiFedora: fixed that wording
16:37:12 <notting> geppetto: wfm.
16:37:19 <geppetto> Ok, cool.
16:37:27 <geppetto> I'm +1 on the new wording
16:37:35 <SmootherFrOgZ> +1
16:37:49 <RemiFedora> where is the new wording ?
16:38:11 <tibbs|w> https://fedoraproject.org/wiki/User:Notting/timer ?
16:38:28 <RemiFedora> I still see " Packages which depend on cron MUST NOT use timer units and instead use cron as specified here, to avoid introducing unnecessary new dependencies on systemd directly. "
16:38:33 <RemiFedora> so I'm still -1
16:38:54 <geppetto> RemiFedora: yeh, you want packages to depend on both, why?
16:39:00 <geppetto> tibbs|w: yeh
16:39:04 <RemiFedora> no, the third line is clear
16:39:29 <RemiFedora> but, as I understand the second line, it deny the switch from cron to timer
16:39:31 <geppetto> RemiFedora: So what's wrong with the second line?
16:39:50 <geppetto> RemiFedora: If they switch, they won't depend on cron anymore … so using timer is fine
16:40:05 <RemiFedora> and I don't understand the "avoid introducing unnecessary new dependencies on systemd directly. "
16:40:13 <Rathann> what about subpackages?
16:40:29 <Rathann> can't we have foo, foo-cron and foo-systemd?
16:40:38 <geppetto> Rathann: I don't see why not
16:40:51 <geppetto> At least from the wording on this page … we might not want that though.
16:41:33 <tibbs|w> Maybe add "(But of course packages can switch from cron to timer units.)" after the second sentence.  I don't know; seems a bit verbose to me.
16:42:07 <RemiFedora> If I create a new package, with a timer, I of course introduce a new dependency on systemd... why should I avoid this ?
16:42:40 <Rathann> I'm +1 in general
16:42:56 <geppetto> RemiFedora: Maybe change "new dependencies" to "additional dependcies" ?
16:42:59 <Rathann> and I have to go away for about half an hour
16:43:22 <RemiFedora> drop the second line, the third is enough
16:44:33 <geppetto> The 2nd seems a lot more forceful than the 3rd, to me
16:44:58 <geppetto> So dropping it gets us back to SHOULD territory, I think
16:45:05 <notting> i'm not seeing how the second sentence applies to or restricts creating a new package with a timer unit
16:45:22 <geppetto> me either … or even to moving a package from cron to timers
16:45:29 <RemiFedora> notting, I just say it is unclear and ambiguous
16:45:55 <geppetto> RemiFedora: Maybe "Packages which currently depend on cron"?
16:46:54 <tibbs|w> I'm +1 on this as is, but if someone wants to tweak the wording of the admon box I'm fine with that as well.
16:46:56 <RemiFedora> "Packages MUST  not depend on both cron and systemd." => should be enough
16:47:19 <geppetto> Ok, let me change it.
16:48:20 <geppetto> Ok, can everyone vote again?
16:48:21 <geppetto> +1
16:48:33 <geppetto> going to assume +1 from abadger1999 and Rathann
16:48:40 <geppetto> notting: I assume you are ok?
16:48:52 <notting> s/ong/on/, at least
16:48:53 <RemiFedora> +1
16:48:54 <tibbs|w> +1
16:49:11 <SmootherFrOgZ> +1
16:49:16 <geppetto> SmootherFrOgZ: want to vote for the record?
16:49:18 <geppetto> cool
16:49:23 <notting> geppetto: better than the should version, yes
16:49:46 <geppetto> #info systemd timer guidlines wording change to use MUST (+1:6, 0:0, -1:0)
16:50:35 <geppetto> pmackinn: For 422, I'm going to just close it based on abadger1999's comment, that fine with you?
16:51:04 <pmackinn> geppetto, ack, new change request i take it
16:51:10 * geppetto nods
16:51:43 <geppetto> Ok, try and at least get +4's ona few easy ones in the next few minutes
16:51:47 <geppetto> #topic #416 	Temporary bundling exception for ipython
16:51:52 <geppetto> https://fedorahosted.org/fpc/ticket/416
16:51:56 <RemiFedora> geppetto, if we have time for 420 (i'd like to work on start changing packages soon)
16:52:24 <RemiFedora> JS again, so again +1 for temp exception
16:52:26 <geppetto> RemiFedora: sure, we can start it … but I don't think we'll have 5 here before we end
16:52:45 <geppetto> I'm +1 for 416
16:55:19 <tibbs|w> Is 416 all about javascript?
16:55:37 <tibbs|w> I think so, and if so, +1.
16:57:13 <geppetto> yeh
16:57:55 <tibbs|w> Our ipython is embarrassingly  old and it's a shame to hold it up for a bunch of random javascript.
16:59:52 <geppetto> SmootherFrOgZ: ping
17:00:11 <SmootherFrOgZ> sorry, got side-tracked.
17:00:15 * SmootherFrOgZ reads
17:01:58 <SmootherFrOgZ> +1 from me
17:02:11 <geppetto> #info Temporary bundling exception for ipython (+1:4, 0:0, -1:0)
17:02:31 <geppetto> Need one more from Rathann or abadger1999 in ticket
17:02:45 <geppetto> #topic #420 	PHP Guidelines change - numeric prefix
17:02:49 <geppetto> https://fedorahosted.org/fpc/ticket/420
17:03:04 <geppetto> RemiFedora: You want to talk about it?
17:03:21 <RemiFedora> just something we should have do for years ;)
17:03:44 <RemiFedora> relying on alpha order is just a mess...
17:03:52 * geppetto nods … it seems sane to me
17:04:42 <geppetto> Maybe have type in the prefix, so like: 10-zend-opcache.ini … 40-other-zip.ini
17:04:47 <RemiFedora> and important: n ot fixed packages will continue to work
17:04:54 <geppetto> yeh, saw that
17:05:56 <geppetto> I'm +1 with the type of extension there (zend, core, other)
17:05:58 <RemiFedora> well... we only have 2 types "zend" and others... and only 2 zend_extension in the repo... so...
17:06:15 <geppetto> no core extensions?
17:06:20 <RemiFedora> I say "core" for extension coming from main php package
17:06:44 <RemiFedora> but this is not really a different type than other C or PECL extension
17:07:15 <geppetto> hmmm, ok … I'm +1 on wording as is then
17:07:17 <RemiFedora> and extension often move from pecl to core
17:07:25 <RemiFedora> or other side
17:07:30 <geppetto> hopefully people will read enough to understand they have to use 40+ as the number
17:07:49 <geppetto> Actually … you don't say what people should name them if they are indifferent
17:08:04 <geppetto> 64?
17:08:19 <RemiFedora> sorry ?
17:08:42 <geppetto> If I do a php extension … and I don't have any deps. I know of etc. … what number should I use as the prefix?
17:08:49 <RemiFedora> Here is a ls /etc/php.d on my computer => http://paste.fedoraproject.org/95013/13977544/
17:08:55 <RemiFedora> 40
17:09:17 <RemiFedora> "other"
17:09:30 <geppetto> ok, you don't want to make it a bit bigger just in case someone else knows they need to load before me?
17:09:51 <geppetto> Can you put the default value in the draft?
17:10:19 <geppetto> RemiFedora: cool so all the packages comply with this already?
17:10:41 <RemiFedora> in my temp PHP 5.6 repo
17:10:46 <geppetto> ahh
17:11:06 <geppetto> fine, I'm +1 anyway given all the work is done.
17:11:41 <geppetto> tibbs: SmootherFrOgZ want to vote before I push it out to the ticket?
17:11:56 <SmootherFrOgZ> +1
17:12:12 <RemiFedora> +1 (of course, just for the record)
17:12:21 <geppetto> :)
17:12:45 <tibbs|w> +1
17:12:49 <geppetto> #info PHP Guidelines change - numeric prefix (+1:4, 0:0, -1:0)
17:12:59 <geppetto> Need one more from Rathann or abadger1999 in ticket
17:13:17 <geppetto> Ok, going to close in a mo.
17:13:19 <geppetto> #topic Open Floor
17:13:23 <pmackinn> geppetto, for 415 do i need to do anything else to the ticket? or an fpc member/bot closes it with the vote results?
17:13:40 <geppetto> pmackinn: you don't need to do anything else
17:13:56 <pmackinn> geppetto, ty
17:14:50 <geppetto> Ok, going to close at :15 past.
17:15:15 <tibbs|w> geppetto: Thanks for running things.
17:15:33 <tibbs|w> Wish I could be less distracted.
17:15:39 <geppetto> No problem, I'll probably do it until abadger1999 gets back.
17:16:01 <geppetto> #endmeeting