15:58:43 <spot> #startmeeting Fedora Packaging Committee
15:58:43 <zodbot> Meeting started Wed May  9 15:58:43 2012 UTC.  The chair is spot. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:58:43 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:58:48 <spot> #meetingname fpc
15:58:48 <zodbot> The meeting name has been set to 'fpc'
15:58:52 <spot> #topic Roll Call
15:59:17 * limburgher here
15:59:17 <tibbs|h> Howdy.
16:01:54 <spot> abadger1999, rdieter_work, Smoother1rOgZ: ping
16:02:15 <rdieter> yo
16:02:39 * geppetto is here
16:03:30 <spot> well, thats quorum. i'll give abadger1999 and Smoother1rOgZ a minute or so to join
16:04:56 * abadger1999 here
16:06:14 <spot> #topic Better elaborate placement of RubyGems binary extensions - https://fedorahosted.org/fpc/ticket/168
16:06:40 <spot> This one seems like a clarification, as opposed to a change in practice, so i'm inclined to say "easyfix, +1"
16:07:09 <tibbs|h> I agree.
16:07:22 <geppetto> +1
16:08:00 <spot> So, i'm officially +1. :)
16:08:14 <rdieter> +1
16:08:24 <tibbs|h> We should always feel free to clarify things when the guidelines are confusingg.
16:08:52 <spot> counting tibbs agreement as a +1, we're at +4
16:09:28 <abadger1999> +1
16:09:58 <spot> limburgher: if you'd like to go on the record, now would be an excellent time to do so. :)
16:10:34 <spot> #action Clarification approved (+1:5, 0:0, -1:0)
16:10:47 <limburgher> +1
16:10:52 <limburgher> Sorry, distracted. :(
16:11:09 <spot> #topic Need Guidelines for packages with restrictive trademark licensing - https://fedorahosted.org/fpc/ticket/170
16:11:26 <spot> Yes, we do, but I'm pretty sure this falls into Fedora Legal's domain and not the FPC.
16:11:55 <spot> if anyone disagrees that this should be Fedora Legal's responsibility, please feel free to speak up. :)
16:11:59 <tibbs|h> I certainly don't want to wade into legal issues.
16:12:04 <limburgher> Yeah, shouldn't the BZ just block FE-LEGAL?
16:12:16 <spot> well, that much of a workflow is clear
16:12:27 <tibbs|h> We can add a link from the packaging guidelines to the legal guidelines if folks are missing it, I guess.
16:12:36 <spot> but it would be nice to have some sort of legal guidelines clarifying what is acceptable wrt a trademark license
16:13:10 <rdieter> legal++
16:13:12 <abadger1999> No problems there.
16:13:20 <abadger1999> Just send it to you then? ;-)
16:13:28 <spot> abadger1999: i will point you to the legal trac. :)
16:13:34 <abadger1999> hehe
16:14:03 <tibbs|h> I certainly don't envy the person who has to draft reasonable guidelines in that area while still allowing us to ship firefox.
16:14:22 * spot opens another bottle of beer
16:14:26 <geppetto> IceWeasel … :-o
16:14:36 <abadger1999> tibbs|h: We can always rename to iceweasel if it's impossible to do that ;-)
16:14:47 <spot> #topic Applying for bundling exception: python-trml2pdf in openerp-server - https://fedorahosted.org/fpc/ticket/171
16:15:01 <rdieter> luke-warm ferret
16:15:07 <tibbs|h> I think this is something we haven't come across before.
16:15:09 <abadger1999> spot: Do you want that in fedora-legal trac or on the legal mailing list?
16:15:10 <spot> i've been thinking about this one for a while now
16:15:18 <spot> abadger1999: the fedora-legal trac, please.
16:15:26 <abadger1999> will od
16:15:28 <abadger1999> *do
16:15:42 <limburgher> epinephrine, stat.
16:15:53 <abadger1999> :-)
16:16:01 <spot> my summarized feelings here are this: we should strive to ensure that the packages we offer in Fedora are as useful and up to date as possible
16:16:53 <spot> so the situation we have now where the packaged python-trml2pdf is old and outofdate (and used only by one other package) seems... poor
16:17:22 <spot> i'd much rather either see that package renamed to python-trml2pdf12 or give satchmo (the user package) a bundling exception for it
16:17:31 <limburgher> So in effect we have two projects that want to consume two diverging versions of a dead upstream?
16:17:52 <rdieter> repoquery --whatrequires python-trml2pdf   => empty  ?
16:17:56 <spot> and allow the python-trml2pdf that is maintained (by OpenERP) to be in python-trml2pdf
16:18:16 <spot> rdieter: i think satchmo uses it, possibly only as a BR?
16:18:16 <tibbs|h> rdieter: The requiring package isn't even in the distro.  (Maybe in the future, I think.)
16:18:23 <rdieter> spot: oh, ok
16:18:37 <tibbs|h> Or maybe I'm wrong.
16:18:39 <spot> oh,nm, tibbs is right
16:18:39 <abadger1999> limburgher: mostly correct -- from the trac ticket, it looks like upstream for the library is the same as the second applicatoin (openerp)
16:18:44 <spot> satchmo never made it into Fedora
16:19:01 * rdieter prefers the 'allow bundling' option here
16:19:26 <abadger1999> limburgher: so viewed in one way, they (openerp) aren't bundling... they've just chosen to fold their code from a library layer back into the application.
16:19:39 <limburgher> abadger1999: Nod.
16:19:42 <tibbs|h> If we allow bundling we should consider giving guidance on what should happen to the existing package.
16:20:03 <abadger1999> yeah -- the existing package and satchmo if it gets in.
16:20:29 * spot almost prefers to see existing python-trml2pdf renamed to python-trml2pdf12 and the maintained OpenERP version moved into python-trml2pdf
16:20:46 <racor> abadger1999: This essentially the same as statically linking against a bundled library in other languages
16:20:47 <limburgher> spot: I like that.
16:21:07 <spot> just because it seems silly to have a python-trml2pdf that is an old fork, in case someone tried to use it
16:21:25 <tibbs|h> Another consideration is that python-trml2pdf includes an executable.
16:21:34 <tibbs|h> So it's actually kind of misnamed.
16:21:42 <limburgher> spot: Right, nothing about the name says "Old and crufty", but if I saw two I'd know which one I wanted.
16:22:00 <tibbs|h> Does the replacement include an executable?  Might someone be relying on having that executable around?
16:22:24 <spot> tibbs|h: i thought that was something different?
16:22:47 <tibbs|h> Not sure what you mean.
16:22:52 <spot> nm.
16:22:56 <abadger1999> racor: not really -- the code isn't being bundled at build time.  it's being bundled when the tarball is released by upstream.
16:23:15 <rdieter> dealing with co-installability isn't really within the scope of what we're discussing here, is it? (ie, lets not get lost in the weeds)
16:23:19 <abadger1999> racor: oh sorry --- I see
16:23:25 <spot> well, clearly, no one Requires python-trml2pdf now, so if they need the executable, they can Requires:python-trml2pdf12 just as easily.
16:23:51 <abadger1999> racor: you said "statically linked against a bundled library" -- I misread and missed the bundled library part.
16:24:17 <spot> rdieter: given that this is python, it shouldn't be terribly difficult to adapt the older package to avoid conflicts
16:24:40 <spot> especially given that nothing in Fedora uses it currently (and it would be a very simple patch for future satchmo to adjust to it)
16:24:45 <geppetto> Do we even need to provide the old one?
16:24:56 <abadger1999> If python-trml2pdf was moved to the openerp version, it would need to use the openerp tarball as it's Source.
16:25:03 <rdieter> spot: true, I'm just arguing that should not concern us (at the moment, at least)
16:25:13 <abadger1999> and then copy the code from there and install it manually to the correct location.
16:25:17 <spot> abadger1999: alternately, the openerp tarball could generate python-trml2pdf
16:25:23 <spot> s/tarball/srpm
16:25:25 <abadger1999> yeah
16:25:31 <abadger1999> that's an option.
16:25:33 <spot> which i think i like better, tbh. :)
16:25:40 <abadger1999> still have to manyally copy it to the correct location
16:25:46 * spot nods
16:25:52 <spot> but there is less risk of source divergence
16:25:55 <limburgher> spot: +1
16:25:59 <abadger1999> <nod>
16:27:08 <spot> So, i propose the following: The existing python-trml2pdf package must either be renamed to python-trml2pdf12 or retired. OpenERP must generate a python-trml2pdf subpackage (placed in the proper python site-packages path).
16:27:10 <abadger1999> I'm +1 tto allowing openerp to contain this library and also +1 to the present python-trml2pdf moving to python-trml2pdf12... not sure about making the openerp maintainer create a python-trml2pdf subpackage.
16:27:12 <rdieter> openerp isn't concerned with the public api here I thought?  generating a python-trml2pdf from it seems unwise to me
16:27:29 <spot> rdieter: perhaps, but they are the upstream
16:27:45 * rdieter would rather simply allow the bundling and be done with it then
16:28:46 <rdieter> but, if agree'ing to making a python-trml2pdf pkg out of it is what it takes to get consensus here, +1 either way
16:29:23 <geppetto> Pretty sure we don't have any API requirements on anything, esp. so on python stuff.
16:29:24 * spot would not lose much sleep over the openerp bundling with no subpackage, as long as the existing package either renames or retires
16:29:46 <geppetto> I think it's better to make it a sub-package.
16:29:55 <spot> geppetto: i lean in that direction as well
16:30:08 <abadger1999> +1 without the need for the openerp maintainer to agree to the subpackage.  Also okay with requesting that he consider making a subpackage but not make it a requirement.
16:30:10 <limburgher> +1 subpackage, rename old.
16:30:18 <geppetto> +1
16:30:18 <rdieter> seems like a waste to make subpkg no one will use, and that upstream doesn't want, but meh
16:30:32 <spot> okay, lets do this
16:30:57 * rdieter thinks that should be the pkg maintainers ultimate call, not ours
16:31:01 <spot> So, i propose the following: The existing python-trml2pdf package must either be renamed to python-trml2pdf12 or retired. FPC ask the openERP maintainer to consider making a python-trml2pdf subpackage (placed in the proper python site-packages path), but is not required to do so.
16:31:26 * rdieter likes that much better, thanks, +1
16:31:29 <tibbs|h> Works for me.  +1
16:31:40 * spot is +1
16:31:45 <abadger1999> +1
16:32:33 <geppetto> +1
16:33:36 * spot is waiting on racor and limburgher to vote on this proposal
16:33:53 <racor> 0, I am hesitant. Doing the opposite (force openerp to rename their subpackage) would be less intrusive and more respectful (spot's proposal can be interpreted as sanctioning a hostile take-over)
16:34:33 <tibbs|h> There's no takeover; the upstream of both sources is the same as far as I understand things.
16:34:52 <rdieter> I think he means the downstream maintainers
16:35:25 <racor> rdieter: I mean those individuals, who are providing the source tar balls.
16:35:37 <tibbs|h> It's the same people in both cases.
16:35:38 <rdieter> ok, nvm
16:36:11 <racor> tibbs|h: Then, there'd be another option: Ask them to clarify the situation.
16:36:59 <limburgher> +1
16:37:05 <spot> #action The existing python-trml2pdf package must either be renamed to python-trml2pdf12 or retired. FPC asks the openERP maintainer to consider making a python-trml2pdf subpackage (placed in the proper python site-packages path), but is not required to do so. (+1:6, 0:1, -1:0)
16:37:45 <abadger1999> Yeah -- if the situation isn't as represented in the ticket, I'm willing to vote differently.
16:38:06 <abadger1999> I see that our tarball for python-trml2pdf is from the satchmo upstream website
16:38:38 <spot> if someone wants to investigate this and reopen the ticket, feel free.
16:38:56 <abadger1999> but it looks like they took a snapshot of the original upstream to get that tarball (I assume from our ticket, that those people would be openerp)
16:39:01 <spot> it seems like the reporter has done due diligence on the situation, nor am i inclined to distrust the ticket info
16:40:48 <spot> #topic Open Floor
16:41:54 <tibbs|h> Nothing from me.
16:42:09 <limburgher> Nope.
16:43:31 <abadger1999> Nothing here
16:44:03 <abadger1999> spot: Is the legal trac setup so that people on the CC can view tickets?
16:44:16 <spot> abadger1999: i think so, yes.
16:44:38 <abadger1999> k /me adds the current CC list and invites others who find the FPC ticket to request to be put on the  legal CC.
16:48:17 <abadger1999> Oh, just a celebreatory FYI: It looks like zlib will finally be unbundled from rsync :-)
16:48:36 <tibbs|h> Wow, finally.
16:48:53 <spot> nice
16:49:08 <spot> and with that, we're done, thanks everyone. :)
16:49:10 <spot> #endmeeting