17:00:33 <geppetto> #startmeeting fpc
17:00:33 <zodbot> Meeting started Thu Nov 12 17:00:33 2015 UTC.  The chair is geppetto. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:33 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:00:34 <geppetto> #meetingname fpc
17:00:34 <zodbot> The meeting name has been set to 'fpc'
17:00:34 <geppetto> #topic Roll Call
17:00:42 <Rathann> geppetto sent the announcement with two different times: 17:00 in the subject and 16:00 and 17:00 in the body
17:00:43 <gbcox> Good Morning... I'm back... ;-)  taking my back row seat!
17:00:45 <geppetto> Yeh, DST damage should be over by now :)
17:00:55 <Rathann> unfortunately I can't stay
17:00:56 <geppetto> Rathann: Bonus :)
17:01:04 <tomspur> Hi
17:01:08 <geppetto> Rathann: The one in the rktime output should be authoritive
17:01:13 <Rathann> I added a comment in one of the tickets
17:01:20 <Rathann> #540 that is
17:01:27 <geppetto> #chair tomspur
17:01:27 <zodbot> Current chairs: geppetto tomspur
17:01:28 <geppetto> #chair mbooth
17:01:28 <zodbot> Current chairs: geppetto mbooth tomspur
17:01:45 * racor is here
17:02:04 <geppetto> #chair racor
17:02:04 <zodbot> Current chairs: geppetto mbooth racor tomspur
17:02:16 <Rathann> I didn't see anything else to comment on
17:02:33 <geppetto> Yeh, there probably isn't
17:02:33 <Rathann> if there's any voting, I'll add my vote in the ticket later if necessary
17:02:39 * geppetto nods
17:02:40 <Rathann> take care and bye
17:02:50 * Rathann goes away
17:04:19 <geppetto> #chair tibbs
17:04:19 <zodbot> Current chairs: geppetto mbooth racor tibbs tomspur
17:05:14 <geppetto> Ok, we have 5
17:05:22 <tibbs|w> Sorry folks, I'm going to have to reboot again.
17:05:29 <geppetto> Ok, no problem
17:05:35 <geppetto> orionp: FPC ping
17:06:34 <orionp> hello
17:07:12 <geppetto> #chair orionp
17:07:12 <zodbot> Current chairs: geppetto mbooth orionp racor tibbs tomspur
17:08:07 <tibbs|w> Hopefully that's done.
17:09:06 <geppetto> ok
17:09:14 <geppetto> #topic Schedule
17:09:19 <geppetto> https://lists.fedoraproject.org/pipermail/packaging/2015-November/011086.html
17:09:38 <geppetto> #topic #540 	Define gcc and gcc-c++ as providing minimal C and C++
17:09:40 <geppetto> compilation environments.
17:09:40 <geppetto> .fpc 540
17:09:40 <geppetto> https://fedorahosted.org/fpc/ticket/540
17:09:42 <zodbot> geppetto: #540 (Define gcc and gcc-c++ as providing minimal C and C++ compilation environments.) – fpc - https://fedorahosted.org/fpc/ticket/540
17:09:56 <geppetto> #topic #540 	Define gcc and gcc-c++ as providing minimal C and C++ compilation environments.
17:10:08 <geppetto> .fpc 540
17:10:09 <geppetto> https://fedorahosted.org/fpc/ticket/540
17:10:10 <zodbot> geppetto: #540 (Define gcc and gcc-c++ as providing minimal C and C++ compilation environments.) – fpc - https://fedorahosted.org/fpc/ticket/540
17:10:54 <tibbs|w> I'm +1 on doing something like this.  I think the text is OK, but for something linked off of the main guidelines I'd like it to be moved into the Packaging: hierarchy.
17:11:37 <tibbs|w> From there we can link out to a tips and tricks document or something if they want to keep the ability to edit something.
17:11:46 <tibbs|w> We do that kind of thing already.
17:11:49 * geppetto nods
17:11:57 <mbooth> tibbs|w: Agree
17:11:59 <tibbs|w> Crap, someone at the door.  Back in a couple of minutes.
17:14:42 <tomspur> It will be just difficult for the c/c++ teams to maintain the packaging:c/c++ page. Somehow it doesn't sound like that draft is in the final state right now
17:15:31 <mbooth> What's missing? Carlos's latest draft in comment 7 reads fine to me
17:16:06 <tomspur> I just said it _sounds_ not in the final state. But looks good so far
17:16:40 <orionp> So we're just deciding on making a link to it?
17:17:02 <tibbs|w> Back.
17:17:37 <tomspur> orionp: Moving to packaging:c/c++ and a link to that, I think
17:17:44 <tibbs|w> orionp: Well, adding a link on the main guideline page and moving under packaging.
17:17:50 <tibbs|w> Well, Packaging:
17:17:54 <geppetto> Yeh, I'm pretty sure it's: https://fedoraproject.org/wiki/C_and_C%2B%2B
17:18:29 <geppetto> which is the formal version of comment #7 linked in comment #17
17:19:11 <geppetto> and, yeh, it looks fine to me
17:19:54 <mbooth> I'm +1 to this plan (of promoting the C_and_C%2B%2B page and linking to it from main guidelines)
17:20:11 <geppetto> +1
17:20:14 <orionp> +1
17:20:43 <tibbs|w> +1
17:20:44 <tomspur> +1, with the MUSTs etc in bold like https://fedoraproject.org/wiki/Packaging:Java#BuildRequires_and_Requires to have some similar layout?
17:21:16 <tibbs|w> Personally I don't like all of the bolded MUST stuff but if people want that....
17:21:40 <geppetto> it's probably a bit easier to read
17:21:52 <geppetto> I'm happy to have stuff bolded
17:21:58 <orionp> It's easier to skim :)
17:22:10 <orionp> but yeah, I like the bold
17:22:14 <racor> +1, with a grim gut feeling. I guess this will require a lot of more work
17:23:46 <geppetto> #action
17:23:59 <tibbs|w> I don't disagree that there's a lot that folks could do with such a guideline.
17:24:17 <tibbs|w> But we've lived without it forever, so I'm still not sure what the big deal is.
17:24:18 <geppetto> #action Promote C and C++ page and link to it from main guidelines (+1:6, 0:0, -1:0)
17:24:45 <geppetto> tibbs: Didn't the default buildroot change recently?
17:24:55 <tibbs|w> Not that I know of.
17:25:06 <tibbs|w> But it _could_, I guess.
17:25:35 <geppetto> I thought gcc was removed from it … but I'm pretty sure all my packages required it anyway, so who knows.
17:25:38 <tibbs|w> That would still only a require a guideline that says you need to depend upon a C/C++ compiler if you need to compile C/C++ code.
17:26:00 <geppetto> Isn't that what this is?
17:26:20 <tibbs|w> Seems a bit long if all it's saying is that.
17:27:41 <tibbs|w> I've always been on the fence about saying that you "MUST" do something when the package will fail to build anyway.
17:27:43 <racor> geppetto: No it isn't. It has the semantics of "R: c-compiler" and "R: gcc" mixed up.
17:28:14 <tibbs|w> That's true, saying "a C compiler" and saying "gcc" are not the same thing given LLVM.
17:28:31 <racor> It also has the semantics of "R: gcc" (the package) and /usr/bin/gcc (the binary) mixed up.
17:28:35 <tibbs|w> But right now we do have a declaration that "the C compiler in Fedora is gcc".
17:28:54 <geppetto> yeh, I think the assumption is that you only use llvm if you must
17:29:06 <geppetto> or clang, or whatever
17:29:08 <tibbs|w> racor: If you have suggestions for what should be changed, I'd be happy to incorporate them into a diff.
17:30:24 <tibbs|w> We can always change the guideline after I write it up.
17:30:29 * geppetto nods
17:30:37 <racor> tibbs|w: As you know, I have always be in favor of file requirements instead of packages, but I had to learn this will be fighting wind mills ;)
17:31:44 <geppetto> the way yum/dnf/repodata works atm. … file requirements are much worse than package ones
17:32:05 <racor> geppetto: dnf doesn't function
17:32:07 <geppetto> although I understand the desire
17:32:20 <geppetto> We still ship yum :)
17:32:32 <geppetto> just under and annoying name atm.
17:32:50 <racor> geppetto: No, I mean this literally.
17:32:50 <tibbs|w> We live with the package manager we have, I guess.
17:33:03 <racor> it's just a bunch of bugs
17:33:30 <geppetto> racor: I also meant it literally … you can still use yum-deprecated
17:33:32 <tibbs|w> That's kind of always been our cross to bear.  Otherwise this packaging stuff would be much simpler.
17:34:48 <geppetto> racor: We even added support for weak deps. into yum for F23
17:34:56 * geppetto nods
17:35:16 <geppetto> Anyway … any diffs. for this?
17:35:26 <tibbs|w> Anyway, that is a completely different topic, and I'm sure we could complain about dnf forever.
17:35:31 <tibbs|w> Ninja'd.
17:36:05 <geppetto> #topic Open Floor
17:36:18 <geppetto> I had a quick look at the other tickets, and there doesn't seem to be much to discuss
17:36:24 <geppetto> Well anything
17:36:44 <tibbs|w> I am moving the spectool thing upstream.  Or, maybe, just becoming the upstream for spectool.
17:36:52 <geppetto> Very cool
17:36:59 <tomspur> tibbs++
17:36:59 <zodbot> tomspur: Karma for tibbs changed to 1 (for the f23 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
17:37:00 <tibbs|w> But that's become sort of orthogonal to the python macros
17:37:03 <geppetto> Is it in rawhide?
17:37:22 <tibbs|w> Not yet; still working with scop to figure out the best way to do things.
17:37:26 * nirik has 2 things to note/mention real quick... ;)
17:37:29 <tibbs|w> Will probably spit it out into a separate package.
17:37:35 <geppetto> nirik: Sure, we
17:37:44 <geppetto> nirik: Sure, we'll probably let you speak ;)
17:38:16 <nirik> so I have been moving builders to f23... so things now should be able to use new rpm stuff like file triggers (well, I still have a few builders to do, but after that we should be ready)
17:38:33 <tibbs|w> That is great news.
17:38:36 <geppetto> tibbs: Is it on your fedora people page or something, for anyone else to play with? Or you prefer to wait until it's in rawhide?
17:38:55 <tibbs|w> geppetto: It's in a pagure repo: https://pagure.io/python-macros
17:39:07 <nirik> and fesco in the last meeting wanted to note on bundling that FPC keeps the list of provide names and handles any disputes with them... pending you all being ok with doing that. ;)
17:39:15 <tibbs|w> Just one script right now; I'll split it up soon.
17:39:45 <tibbs|w> nirik: I'd think we can do that; just have to add some process for it.
17:40:06 <geppetto> Yeh, how would that work with no gatekeeping on bundling?
17:40:10 <tibbs|w> I don't think it would need committee involvement much; just a way to request a P: bundled(XXX).
17:40:23 <tibbs|w> And one of us would simply stick it on a page.
17:40:38 <tibbs|w> Either that or just move the page out of Packaging: and let folks have a go.
17:40:44 <nirik> yeah, unless it's something where there's already that thing by another name or something.
17:41:43 <nirik> I also have some dnf RFE's to file around the f23 move... it's not very nice on it's builddep output...
17:42:03 <tibbs|w> Good luck with that one.
17:42:11 <mbooth> tibbs|w: Actually I like the idea of moving the page out of Packaging: and let it be self-policing.
17:42:28 <tibbs|w> Also, someone made mention that Boolean dependencies are supported in the buildsystem.
17:42:36 <tomspur> Same for me. Opening a ticket when there is a naming confusion should be enough.
17:42:49 <tibbs|w> I didn't think that was the case, but if they are, then let us know and we can drop the prohibition from the guidelines.
17:43:13 <tibbs|w> mbooth: I can easily do that; the page isn't in Packaging: now anyway since I deleted it.
17:43:37 <mbooth> Cool :-)
17:43:45 <tibbs|w> So I'll just resurrect it somewhere and add a link from the now-abbreviated bundling section.
17:43:48 <nirik> There's a fesco policy page on it, could be added there I guess?
17:43:54 <nirik> or just linked to from there.
17:45:56 <tibbs|w> Sure, I'll let fesco know after I resurrect the page into the public namespace.
17:46:11 <tibbs|w> Unless I can just edit that policy page.
17:47:10 <nirik> sure, go for it.
17:47:14 <nirik> it's not protected.
17:48:03 <tibbs|w> OK.
17:48:53 <tibbs|w> On 558, I'm trying to remember what I needed to do (if anything).
17:49:29 <tibbs|w> I had produced an overlong draft, I guess I was supposed to keep working on it.
17:50:20 <geppetto> Yeh, everyone was mostly in favour of it IIRC
17:51:15 <tibbs|w> That's what I recall, but I thought it was overly verbose and I think some folks had issue with the examples I used.
17:51:33 <tibbs|w> I'll dig out the meeting log and see if I can improve things.  Maybe drop it back to meeting for next week.
17:51:45 * geppetto nods
17:52:13 <tibbs|w> Just the other day I saw someone say in a review that something which is obviously an application needed to take a python-foo name just because it contained a module.
17:52:28 <geppetto> :(
17:53:58 <tibbs|w> I didn't have a chance to add a correction, and now I can't find the review ticket.
17:54:06 <tibbs|w> Anyway, another bit on my list.
17:54:55 <tibbs|w> On file triggers, I wonder what we need to do to get glibc to add one to call ldconfig.
17:55:09 <geppetto> Yeh
17:55:16 <geppetto> that's one of the missing ones
17:55:28 <tibbs|w> I would really like to see all of the big ones have them before f24 branches.
17:56:07 <tibbs|w> All of the major packages, that is.  Basically everything on the scriptletsnippets page, plus R and texlive.
17:56:07 * geppetto nods … it should be real soon … was going to be this week, but then I went out into the world and people gave me weird germs.
17:56:14 * geppetto nods
17:56:32 <tibbs|w> That's when I get I get most of my Fedora work done.  Otherwise I have to do work work.
17:56:40 <geppetto> :)
17:57:20 <geppetto> it was at the point that I was just drinking hot tea *quil and feeling miserable.
17:57:37 <tibbs|w> nirik: Where would I go to keep on top of what's happening with file triggers and our build system.
17:57:44 <tibbs|w> ?
17:58:28 <geppetto> The only thing I've found is a full checkout of pkgs-git
17:58:38 <geppetto> Then look through the specfiles
17:59:08 <tibbs|w> Oh, I did file a ticket with infra for getting us a current specfile tarball.
17:59:22 <tibbs|w> But I meant the whole deal with actually getting them working in the buildsys.
17:59:55 <nirik> tibbs|w: dunno... what info are you looking for?
17:59:59 <tibbs|w> Once the builders are all on F23, I'm sure there will still be more testing and bugfixing before we can tell people to start dropping scriptlets from packages.
18:00:32 <tibbs|w> I can hang out in #fedora-admin as usual but I'm not sure how we'll know that things are "ready".
18:00:49 <tibbs|w> Is there a ticket I can watch or something?
18:01:04 <nirik> no. not that I know of.
18:01:10 <nirik> we have just been fixing things as we hit them
18:02:49 <tibbs|w> OK, well, I'll ping occasionally.  Obviously for some of the less "core" things it doesn't make much difference, but if packages drop some of the more important scriptlets I'd hate to see the whole distro break.
18:04:00 * geppetto nods
18:04:08 <geppetto> Ok, anything else before I close?
18:04:31 <nirik> sure, there's already some that have...
18:04:35 <tibbs|w> I think that's all I had.   We have only five tickets open now, and one will close after this meeting.
18:05:08 * geppetto nods
18:05:16 <geppetto> #endmeeting