17:00:08 <geppetto> #startmeeting fpc
17:00:08 <zodbot> Meeting started Thu Dec  8 17:00:08 2016 UTC.  The chair is geppetto. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:08 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:00:08 <zodbot> The meeting name has been set to 'fpc'
17:00:08 <geppetto> #meetingname fpc
17:00:09 <geppetto> #topic Roll Call
17:00:09 <zodbot> The meeting name has been set to 'fpc'
17:00:18 <orionp> hello
17:00:21 <ignatenkobrain> hi
17:00:25 <geppetto> #chair orionp
17:00:25 <zodbot> Current chairs: geppetto orionp
17:00:29 <tibbs> Hey.
17:00:39 <geppetto> #chair tibbs
17:00:40 <zodbot> Current chairs: geppetto orionp tibbs
17:00:43 <geppetto> Hey
17:02:23 <mbooth> Hi
17:02:27 <geppetto> #chair mbooth
17:02:27 <zodbot> Current chairs: geppetto mbooth orionp tibbs
17:03:41 <tibbs> I actually found a little bit of time to take care of a couple of items from my backlog.
17:03:56 <orionp> Thank you tibbs
17:04:59 <geppetto> Cool
17:05:18 <tibbs> I'll get there.  I think my master todo list is actually shrinking now, although now I have a pile of F25 updating to do.
17:05:28 * geppetto nods
17:05:58 <geppetto> It's mostly been painless, for me, but then I don't use gnome … so no wayland
17:06:19 <Rathann> hi
17:06:29 <geppetto> And the big thing that broke recently was within F24 … so I can't blame F25 for that ;)
17:06:33 <geppetto> #chair Rathann
17:06:33 <zodbot> Current chairs: Rathann geppetto mbooth orionp tibbs
17:06:41 <geppetto> Hey! … and then there were 5!
17:07:00 <tibbs> Yeah, F25 has been OK for me.  Still the same selinux issues as F24 but I haven't seen any _additional_ pain.
17:07:12 <geppetto> #topic Schedule
17:07:15 <geppetto> https://lists.fedoraproject.org/archives/list/packaging@lists.fedoraproject.org/message/62VTMZA373AEJC4P5KVERSKLGWDSLNWX/
17:07:51 <geppetto> #topic #660  Review Process Exception: python2 (renamed from python)
17:07:56 <geppetto> .fpc 660
17:07:57 <zodbot> geppetto: #660 (Review Process Exception: python2 (renamed from python)) – fpc - https://fedorahosted.org/fpc/ticket/660
17:07:59 <geppetto> #chair racor
17:07:59 <zodbot> Current chairs: Rathann geppetto mbooth orionp racor tibbs
17:08:28 <geppetto> This seems like an obvious +1
17:08:30 <tibbs> I have no problem with this.  +1
17:09:04 <tibbs> Of course if there is cruft in the python package then it would be nice to get it fixed, but I see no reason to gate an obviously needed rename on a complete spec overhaul.
17:09:10 <racor> +1
17:09:25 <mbooth> +1
17:09:35 <tibbs> Part of the re-review process is making sure that they actually get the Obsoletes:/Provides: lines right, but.... I'm sure the world will let them know if they don't.
17:09:53 <mbooth> tibbs: A complete spec overhaul might introduce bugs anyway....
17:10:13 <orionp> +1
17:10:14 <tibbs> Yes, potentially worse ones.
17:10:40 <geppetto> Rathann: vote?
17:11:04 <Rathann> *sigh*
17:11:17 <Rathann> +1
17:11:26 <geppetto> #info Please double check the new obsoletes/provides, as there won't be a review
17:11:34 <geppetto> #action Review Process Exception: python2 (+1:6, 0:0, -1:0)
17:11:36 <Rathann> the least they could do is use https for source url
17:11:52 <geppetto> #info Please change source URL to use https
17:11:53 <geppetto> :)
17:11:55 <Rathann> I'm assuming python.org provides one
17:12:05 * geppetto ¯\_(ツ)_/¯
17:12:20 <geppetto> #topic #661  Formalize/standardize %check-optional packaging
17:12:23 <geppetto> .fpc 661
17:12:24 <zodbot> geppetto: #661 (Formalize/standardize %check-optional packaging) – fpc - https://fedorahosted.org/fpc/ticket/661
17:12:28 <Rathann> yup they do
17:13:43 <tibbs> This probably needs more discussion elsewhere.
17:13:56 <tibbs> It's also possible I was misunderstanding some of the RPM functionality.
17:14:25 <geppetto> I'm happy with #1 #2 and #4 … not sure #3 is worth it, or a good idea anyway
17:14:40 <tibbs> https://bugzilla.redhat.com/show_bug.cgi?id=1134397 has the real details.
17:15:01 <tibbs> This is one of those things where either RPM upstream needs to just make something happen, or we need to drive some macro changes.
17:15:47 <tibbs> I can live without 3, of course.  I keep playing around with versions of that type of convenience macro but have never managed to actually put anything forward.
17:16:18 <tibbs> But I do think I'm confused about what --nocheck actually does, if it does anything, and how that differs from --without check.
17:16:39 <tibbs> Was kind of hoping someone else actually knew so I wasn't just bumbling around.
17:17:20 <geppetto> As panu says in the BZ --nocheck has to be identical to --without check
17:18:07 <tibbs> I think the implication is that it isn't, but that he promises to make sure that whatever he does doesn't break --without check.
17:18:19 <tibbs> And any spec mechanisms we put in place to deal with it.
17:18:48 <geppetto> The 3 thing just seems less readable to me, but I just like all my Requires/BuildRequires/etc. aligned to the left
17:19:31 <ignatenkobrain> ++ from me
17:19:32 <geppetto> Where do you get that impression?
17:20:22 <tibbs> geppetto: Talking to me?
17:20:26 <geppetto> yeh
17:20:35 <tibbs> I was just reading comment 12 from that bugzilla ticket.
17:21:14 <tibbs> He says that in the future he would have to make --nocheck an alias for --without check.
17:21:38 <tibbs> I had assumed that was the case already, faked in using popt, but I seem to have been mistaken.
17:21:59 <geppetto> Yeh, but --nocheck doesn't exist now right? … he's talking about how they'd do it, as they have to behave the same?
17:22:13 <tibbs> For some reason I thought it existed.
17:22:28 <orionp> It does exist
17:22:38 <orionp> it prevents running %check
17:22:41 <geppetto> Yeh, as does --noclean
17:22:49 <geppetto> I'm getting more confused
17:22:59 <orionp> but I don't think it does anything else
17:23:21 <tibbs> I think the point is that --nocheck just doesn't run %check.  But it sets no visible macro, so you can't use it to make any difference to anything.
17:23:23 <geppetto> Ok, so --nocheck doesn't currently set the bcond for {with check} tests?
17:23:36 * geppetto nods
17:23:45 <geppetto> Ok, less confused, more gurmpy
17:23:46 <tibbs> Right, and in macros there's no way at all to know if it was passed.
17:24:26 * geppetto nods
17:24:40 <tibbs> So basically, you can pretend it doesn't exist and then hack up all of your specs to make --without check act like --nocheck with the added bonus that you can conditionalize things based on it.
17:25:07 <tibbs> I would really, really dislike seeing every package with a %check section have to add an %if bit.
17:25:17 <geppetto> Yeh
17:25:41 <geppetto> At least for the %check part itself
17:25:44 <tibbs> You can get it down to one line with %{?without_check: exit 0}
17:26:03 <geppetto> magic!
17:26:23 <geppetto> put that in a random %global goaway
17:26:33 <tibbs> But if we mess with the one macro, we can make it no lines by adding that exit 0 at the beginning automatically.
17:26:47 <tibbs> Which was my #2.
17:26:58 <tibbs> In https://fedorahosted.org/fpc/ticket/661
17:27:09 * geppetto nods
17:27:17 <geppetto> I don't mind that, due to the echo
17:27:42 <tibbs> With #1 and #2 you need very little.
17:28:07 * geppetto nods
17:28:18 <orionp> I like it
17:28:21 <tibbs> So I can pretend I never wrote #3, assume that #1 and #2 are OK and actually carry out #4.
17:28:32 <geppetto> Sounds good
17:28:55 <geppetto> I am curious why you think that modularity will make this a problem in the near future though?
17:28:56 <tibbs> I will do some experimentation inside of redhat-rpm-config and some test builds to make sure that what I said will work will actually work.
17:29:06 * geppetto nods
17:29:13 <geppetto> +1 for all of that anyway
17:29:28 * Rathann is +1 too
17:29:48 <tibbs> Given Panu's promise in comment 12, we can do that safely without worrying about whether RPM will break us.
17:29:59 * geppetto nods
17:30:01 <orionp> +1 for #1, #2, #4
17:30:06 <tibbs> I'll +1 and try to scrape together a draft.
17:30:47 <geppetto> #action tibbs to do a draft on #1 and #2 in redhat-rpm-macros, ignoring #3
17:31:04 <geppetto> #topic #662  Migration of FPC trac to pagure
17:31:09 <geppetto> .fpc 662
17:31:13 <zodbot> geppetto: #662 (Migration of FPC trac to pagure) – fpc - https://fedorahosted.org/fpc/ticket/662
17:31:45 <tibbs> That was just an informational thing.
17:32:11 * geppetto nods … we should all look at it though, I think :)
17:32:21 <tibbs> More so I have a way to track and notify if anything changes.  There's a new pagure version up so I need to go try to do another import.
17:32:37 * geppetto nods
17:32:48 <geppetto> All the timestamps/etc. seem to be there
17:33:08 <tibbs> I think all of the committee folks have access to that test repo, so feel free to play with settings and such.  The initial import was rather unsatisfactory, of course, but I'll just have it not try to import the tags next time.
17:33:14 <geppetto> So assuming we can get the meeting schedule reports, we are ding pretty good
17:33:34 <tibbs> If anyone knows how to do ticket templates in pagure, that would be nice to try as well.
17:33:48 <tibbs> We need to make sure that ticket submission works well, too.
17:34:02 <geppetto> Maybe ;)
17:35:06 <geppetto> #topic #663  New python egg macros
17:35:11 <geppetto> .fpc 663
17:35:13 <zodbot> geppetto: #663 (New python egg macros) – fpc - https://fedorahosted.org/fpc/ticket/663
17:35:54 <geppetto> orionp: Your turn
17:35:56 <tibbs> I did a bit of cleanup to the Packaging:Python_Eggs page, which sadly screws up the diff, but....
17:36:08 <geppetto> Although FWIW this looks like a trivial +1 to me
17:36:17 <tibbs> Yeah, I'm +1.
17:36:36 <geppetto> tibbs: The diff url posted in the ticket seems fine
17:36:46 <tibbs> Yeah, it just doesn't diff against the current page.
17:36:50 * geppetto nods
17:36:58 <orionp> I also ended up adding %py2_build/install_wheel as well - though only python-pip seems to do that currently
17:36:58 <tibbs> Not a big deal; I can fix that up when I do the writeup.
17:37:36 <tibbs> orionp: Where did you add those?  The macro package, or the draft?
17:37:38 <geppetto> wheel?
17:37:52 <orionp> to the macros, haven't documented yet
17:38:03 <tibbs> geppetto: It's yet another packaging format thing that the python people invented.
17:38:11 <geppetto> nice
17:38:16 <tibbs> They really like packaging, for some reason.
17:38:23 <geppetto> The great thing about stds. etc. etc.
17:38:45 <Rathann> huh, I don't know enough about python - I'd love to see a real-life example spec making use of this
17:38:46 <tibbs> orionp: Can we pretend those macros don't exist for now so we can just do the egg thing?  Or are they interrelated?
17:39:00 <orionp> we can pretend they don't exist
17:39:48 <tibbs> So, really, the proposal boils down to replacing one call to setup.py with calls to %py*_build_egg, and...
17:40:04 <tibbs> the mkdir and easy_install calls with %py*_install_egg calls.
17:40:33 <orionp> right
17:40:41 <tibbs> There was also something about removing a section?
17:40:42 <geppetto> +1
17:40:56 <tibbs> Or did I remove that already when I did the cleanup?
17:41:41 <tibbs> I'm +1 to the %py*_*_egg macros, certainly.
17:42:13 <orionp> this is the caveat --
17:42:14 <orionp> __requires__='MODULE[VERSIONINFO] ': This works but the setuptools author feels that it is only a workaround and will not support it. It works presently but could stop in a future version of setuptools. Some upstreams use this method and may need to be fixed if the setuptools author ever changes the interface.
17:42:16 <orionp> [edit]
17:42:27 <orionp> I think it is supported
17:43:02 <orionp> as it is in the setuptools docs currently
17:44:16 <Rathann> I'm +1 to the macros
17:45:17 <tibbs> orionp: Let's try to get the vote on the two macros done, but once that's done I'd still like to talk about the document.
17:45:50 <mbooth> +1
17:46:04 <mbooth> TIL about python "wheels" though... interesting
17:47:02 <geppetto> racor: vote?
17:47:03 <tibbs> orionp: Can we assume you're +1?
17:47:08 <geppetto> orionp: I'm assuming you are +1 :)
17:47:13 <orionp> +1 :)
17:47:14 <geppetto> jinx
17:47:30 <racor> +1
17:47:40 <geppetto> #action New python egg macros (+1:6, 0:0, -1:0)
17:47:51 <geppetto> #topic #665  SSLCertificateHandling policy update
17:47:56 <geppetto> .fpc 665
17:47:57 <zodbot> geppetto: #665 (SSLCertificateHandling policy update) – fpc - https://fedorahosted.org/fpc/ticket/665
17:49:36 <geppetto> I meant to get dwmw2 to comment, but nevermind
17:51:15 <tibbs> I'll bring up the eggs guiedline again for open floor.
17:51:26 <tibbs> I have so little idea about 665, though.
17:51:40 <tibbs> I mean, to me it doesn't look like it has anything to do with SSL certificates.
17:52:06 <tibbs> It would be nice to have guidelines for SSL certificates.  Where you put them, what perms they should have, etc.
17:52:15 <geppetto> It looks fine, to me, but I'm not sure I'd spot anything that got broken
17:52:36 <Rathann> I'm not sure how this is related to packaging
17:52:41 <tibbs> But when looking for guidance I end up at this guideline and I'm left with more questions than when I started.
17:52:58 <Rathann> surely we won't be packaging SSL certs on HSMs or smart cards...
17:53:27 <Rathann> I mean, that's also a way of packaging, but surely not RPM packaging ;)
17:54:50 <Rathann> I the proposal mandates that programs supporting certificates located on such devices understand the  RFC7512 standard
17:54:58 <Rathann> *I think the...
17:54:59 <orionp> I think the concern is that software in fedora hooks into the proper framework
17:55:07 <tibbs> I really just can't say anything here.  Though... good point about questioning why this is a packaging guideline.
17:55:10 <Rathann> which is fine by me
17:55:37 <orionp> But yeah, not sure that this is really packaging guidelines
17:55:48 <Rathann> and also supporting system-wide config gets +1 from me
17:56:08 <Rathann> though I'm not sure why this deserves a separate page
17:56:11 <mbooth> Well, it does tell you to install provider modules in  %{_libdir}/pkcs11/ etc
17:56:39 <mbooth> But I agree the rest is of questionable use to packagers directly
17:56:56 <racor> mbooth: A guideline would have to tell "how" to install there
17:57:58 <Rathann> current working examples are also useful, so I wouldn't move them out to a separate page
17:58:33 <racor> they are referring to a "p11-kit" - Are we supposed to invoke some install/uninstall scripts? No idea.
17:59:27 <Rathann> so, I'd ask for a diff between current and proposed guideline and also ask to answer the concerns we voiced
17:59:53 <Rathann> because as it is, it's really not easy to tell what's changed
18:02:37 <racor> also, I suspect, there need to be some mandatory "Requires:" and/or "BuildRequires" somewhere.
18:05:37 <Rathann> racor: I guess if an application supports p11-kit, it already depends on its libraries, but sure, that should be clarified
18:07:24 <racor> Rathann: May-be - I am just trying to find some sense in this - well - "proposal" ;)
18:10:44 <geppetto> #info We'd like a diff. from current guidlines, just so we can easily know what changed
18:10:47 <tibbs> I certainly have nothing to add.
18:12:07 <geppetto> #info See/answer random confused questions asked in meeting, we aren't pk11 experts
18:12:23 <geppetto> #topic Open Floor
18:12:28 <geppetto> Ok, that's it for new tickets
18:12:43 <tibbs> Can we go back to the Eggs guideline page for a few?
18:12:50 <geppetto> Sure
18:13:08 <geppetto> #topic #topic #663  New python egg macros
18:13:13 <geppetto> #topic #663  New python egg macros
18:13:21 <tibbs> Basically, I don't understand the purpose of everything on that page from "There are several other" and below.
18:13:32 <tibbs> https://fedoraproject.org/wiki/Packaging:Python_Eggs
18:14:02 <tibbs> Maybe the Egg "Features" to avoid section is OK; I'm not sure.
18:14:31 <tibbs> The links section seems pointless.
18:14:35 <orionp> I'm pretty green of all of this too
18:14:40 <geppetto> orionp: ^?
18:14:47 <tibbs> But I don't know enough about this, and sadly that page doesn't really help me know more.
18:15:11 <tibbs> Which is OK, it's a guideline.
18:15:26 <tibbs> I seem to recall Toshio wrote most of it originally.
18:16:04 <orionp> But I've come to think that the __requires__='MODULE[VERSIONINFO]'  method should actually be recommended
18:16:18 <tibbs> Well, recommended how?
18:16:41 <tibbs> This is just for doing multiple versions of one module, right?
18:17:04 <orionp> right - there are two issues:
18:17:06 <tibbs> Which packages do that currently?  Are they consistent?
18:17:17 <orionp> - How to package an alternate version of a module
18:17:32 <orionp> - How to have a package use an alternate version of a module
18:18:00 <tibbs> I will note that if I had those questions, a page named "Python Eggs" isn't where I'd look.
18:18:34 <geppetto> indeed
18:18:35 <tibbs> Which at least calls for a better reference from the main python guideline page.
18:18:39 <orionp> it's linked from a section about multiple versions
18:18:46 <tibbs> Is it?  Crap.
18:19:02 <tibbs> I thought it was only linked from a section that's already talking about eggs.
18:19:35 <orionp> oh, wait
18:20:00 <orionp> you're right - eggs talks about multiple versions
18:20:31 <orionp> a lot of this stuff is historical and probably should be removed
18:20:46 <orionp> I mean, eggs is just what we do now
18:21:54 <tibbs> I guess we need to involve the python sig and see what can just go.
18:22:30 <orionp> yeah
18:23:00 <tibbs> Outdated guidelines pages are worse than incomplete or missing ones, I think.
18:24:03 <tibbs> Anyway, I guess that's all we can do on that.
18:24:12 <tibbs> So if anyone has anything else, feel free.
18:24:30 <orionp> these seem to be the python packages providing versioned eggs - http://paste.fedoraproject.org/501919/81221460/  so not too many
18:24:33 <geppetto> #info Python sig needs to look at the page
18:24:57 <orionp> I'll send an email
18:25:17 <geppetto> yeh, pretty small … only 9 packages
18:25:27 <geppetto> #topic Open Floor
18:25:51 <geppetto> tibbs: Any of the older tickets we should look at?
18:26:17 <tibbs> Hmm.
18:26:30 <tibbs> We should probably make a pass over the needinfos to make sure they're really needinfo.
18:27:04 <geppetto> Ok
18:27:11 <tibbs> I did "complete" that draft rewrite of the versioning guidelines.
18:27:13 <geppetto> was just talking about the ones on the schedule :-o
18:27:22 <geppetto> The tilda one?
18:27:39 <tibbs> I didn't touch the tilde-using draft.
18:27:57 <geppetto> ok
18:28:00 <tibbs> I just did a draft of a rewrite of the versioning guidelines with a goal of making them more comprehensible.
18:28:06 * geppetto nods
18:28:08 <tibbs> I don't know if I achieved that goal, but....
18:28:18 <tibbs> Now where did I put that draft.
18:29:05 <tibbs> https://fedoraproject.org/wiki/User:Tibbs/VersioningCleanup
18:30:01 <tibbs> So as before, I'd love comments on the draft, and I'd live to have proper examples for the separate examples page.
18:30:43 <tibbs> https://fedoraproject.org/wiki/User:Tibbs/VersioningCleanupExamples
18:30:57 <tibbs> And ideas on how to format those examples, because I am terrible at wiki.
18:33:41 <geppetto> They seem fine to me
18:33:44 <tibbs> Once the examples are done I guess I should open a ticket for replacing the current versioning page.  But if people like what's there, then I promised a redo of the tilde versioning draft in the same format.
18:33:54 <tibbs> I do believe there are functional changes in there.
18:34:37 <geppetto> Release tag is a bit confusing
18:34:46 <tibbs> One is the thing we talked about but never voted on, I think, regarding snapshot versioning and using a period instead of "git" or "snap" or whatever.
18:35:13 <tibbs> geppetto: No kidding.  That's basically where all of the pain is.
18:36:02 <tibbs> You need like four pieces of information, some of which might not be there.  Then you join some of them with periods as a separator, except for the dist tag which you just append.
18:36:31 <Rathann> the main page is quite good although I can't find a place where all the described components of the Release: field come together
18:37:09 <geppetto> tibbs: *nods*
18:38:01 <tibbs> Rathann: In the "More complex versioning" section, directly after the three pieces of information you may need are defined.
18:38:56 <Rathann> tibbs: yes, they're defined but never used together
18:39:21 <Rathann> but I guess that's what the Examples page is for
18:39:27 <tibbs> " Those that are present are joined together, with periods as separators. To that, the Dist tag MUST be appended to make the final Release: tag. "
18:39:53 <Rathann> ah
18:39:54 <tibbs> The problem is that you can't just say "pkgrel.extraver.snapinfo%{?dist}" because you might not have all three.
18:39:58 <Rathann> missed that
18:40:27 <tibbs> I could list a table of every possible combination, but... there are eight if you count the minor bump thing.
18:40:33 <Rathann> right
18:40:37 <tibbs> And pkgrel might have its own periid.
18:41:38 * Rathann remembers BNF suddenly ;)
18:41:40 <tibbs> But I can see a reason to embolden or otherwise emphasize the actual construction of the Release: tag.
18:41:52 * geppetto nods
18:41:53 <tibbs> Yeah, BNF.... I almost went there.
18:42:14 <geppetto> "pkgrel[.extraver[.snapinfo]]%{?dist}" ?
18:42:22 <tibbs> But really, if you think it would be clearer stated some other way, I'm all there.
18:42:31 <Rathann> yeah ;)
18:42:42 <tibbs> Well, you might have snapinfo without extraver, so nesting them isn't right either.
18:43:48 <Rathann> anyway, https://fedoraproject.org/wiki/User:Tibbs/VersioningCleanup is quite good as it is
18:44:01 <Rathann> thanks for working on it
18:44:27 <tibbs> pkgrel[.extraver][.snapinfo]%{?dist}[.minorbump]
18:44:47 * geppetto nods
18:44:54 <geppetto> And, yeh, thanks for the work on this.
18:44:57 <tibbs> I'll see how that works.
18:46:59 <geppetto> Ok, anything else in the last 15 minutes?
18:49:12 <geppetto> Ok, thanks for coming.
18:49:16 <geppetto> #endmeeting