From Fedora Project Wiki

(Redirected from Packaging/Minutes20071204)

Fedora Packaging Committee Meeting of {2007-12-04}

Present

  • DavidLutterkort (lutter)
  • JasonTibbitts (tibbs)
  • JesseKeating (f13)
  • RalfCorsepius (racor)
  • RexDieter (rdieter)
  • ToshioKuratomi (abadger1999)
  • VilleSkyttä (scop)

Writeups

The following draft has been accepted by FESCO and is to be written into the guidelines:

Votes

No votes this week.

Other Discussions

No real meeting this week; those present discussed some of the issues surrounding the lack of proper Java packaging guidelines.

IRC Logs

[11:14]  <tibbs> So, FPC meeting?
[11:14]  * lutter is here
[11:16]  --> bhanupatial has joined this channel (n=bhanu@dialpool-210-214-123-127.maa.sify.net).
[11:16]  <racor> i am here
[11:16]  <tibbs> I set the font proposals to writeup.
[11:18]  * scop is here, but only for 10 minutes
[11:18]  <tibbs> The only question I had is regarding the groups portion of the font guidelines which we decided weren't our business.
[11:18]  <tibbs> Do those get written in as normal?
[11:19]  --> smooge has joined this channel (n=smooge@canopus.unm.edu).
[11:19]  <rdieter> imo, yes, it's just something outside of our approval/oversight. :)
[11:20]  <tibbs> I can't think of where else it would live if not below Packaging.
[11:21]  <rdieter> yep
[11:23]  <tibbs> The only packaging-related stuff I had to talk about regards java, and that's just random musing.
[11:24]  <tibbs> I started getting some useful answers to a couple of the java questions I asked.
[11:24]  <tibbs> Most annoying is the fact that some of the jpackage stuff is actively broken and we really shouldn't have it in Fedora.
[11:24]  <scop> +1
[11:26]  <tibbs> The only real point is due to that there's no hope for a wholesale incorporation of jpackage into Fedora, which is what some folks seem to have wanted.
[11:26]  <tibbs> Personally I'm happy about that, but now if we have to go modifying every package to remove broken jpackage-isms, we might as well have our own set of clean guidelines.
[11:26]  <tibbs> Unfortunately nobody from the java team wants to step forward to help write them.
[11:28]  <tibbs> As I review more java packages, though, I might be able to cobble together enough random knowledge to at least propose something.
[11:28]  <scop> I'm also interested in helping out with that
[11:28]  <tibbs> Maybe having a concrete proposal on the table could at least generate some feedback.
[11:29]  <tibbs> The big problem is that everything is going to change when icedtea becomes fully functional.  As I understand things, it's supposed to support PPC now.
[11:30]  <tibbs> Maybe we can ditch the "gcj_support" stuff at some point.
[11:31]  <tibbs> Because this isn't pleasant: %define gcj_support %{?_with_gcj_support:1}%{!?_with_gcj_support:%{?_without_gcj_support:0}%{!?_without_gcj_support:%{?_gcj_support:%{_gcj_support}}%{!?_gcj_support:0

[11:31] <rdieter> I'd rather get a papercut and pour lemon juice on it [11:32] <scop> well, that could be simplified a lot [11:32] <scop> %bcond_with gcj_support [11:33] <lutter> it seems they can't make up their mind how to indicate gcj_support [11:33] <f13> "backwards support" [11:33] <lutter> tibbs: is thaty in every rpm or in some macro file somewhere ? If the latter, it doesn't seem like that big a deal [11:34] <f13> lutter: every rpm at the top [11:34] <lutter> yuck [11:34] <tibbs> And completely meaningless to us because you can't pass flags to the buildsys. [11:34] <scop> the example tibbs posted also looks broken, btw [11:34] <f13> yes. [11:35] <f13> its pure junk from jpackage. [11:35] <scop> s/posted/pasted/ [11:36] <tibbs> I think a basic java template would be nice and simple. I really don't understand why all of the complexity is in there. [11:36] <f13> jpackage. [11:36] <scop> back when I was an active JPackage member, I tried to resist quite a bit of junk and complexity, but unfortunately with pretty thin results [11:36] <f13> and the plethora of jvms [11:37] <tibbs> But for Fedora, we just have to consider icedtea. [11:37] <tibbs> (Assuming icedtea now has sufficiently complete support for our platforms.) [11:37] <scop> I don't think it's quite _that_ simple [11:37] <scop> as long as the alternatives system is in place [11:38] <tibbs> I don't see why; that can't alter how the packages are built. [11:38] <tibbs> We just need to make sure jars are in the proper place. [11:39] <scop> and that the stuff is compiled with appropriate -source flags [11:39] <scop> and that it loads a sufficiently new java runtime when run [11:39] <tibbs> All of that is meaningless to me. [11:39] <tibbs> And the example packages I'm looking at make no mention of that kind of thing. [11:39] <scop> ditto to most people out there, which is why I actively resisted the alternatives mess in the first place [11:40] <tibbs> I still contend that icedtea is the only thing we need to consider in guidelines. [11:40] *** dwmw2 is now known as dwmw2_gone. [11:40] <scop> I disagree strongly unless we rip out the alternatives stuff [11:41] <tibbs> If someone wants to write up a document on how to deal with "your package doesn't work with Sun JVM" bugs, then good. [11:41] <tibbs> But we simply can't consider the implications of every random JVM that someone might want to plug into their system. [11:42] <scop> but by having the alternatives stuff in place, we are giving a "permission" to not only do exactly that, but to make it the system default [11:42] <tibbs> That then falls on the end user to deal with the mess. [11:42] <tibbs> People can try to build a 2.4 kernel, too. We don't care if it works. [11:43] <scop> not the same thing IMO [11:44] <tibbs> In the end all this is going to result in is a continued lack of any java guideline at all. [11:45] <scop> I disagree, but we'll see [11:45] <tibbs> I do not think it appropriate to expect any community maintainer of a java-using package to have to worry about what random jvm end users might try to plug in. [11:45] <abadger1999> scop: +1 to some extent. Using alternatives does seem to imply that we support the idea of using alternate jvms. [11:45] <abadger1999> OTOH, I don't know if we need to extend it to jvm's that Fedora doesn't ship.... [11:45] <scop> tibbs, you're overreacting [11:45] <tibbs> Erm, no. [11:46] * scop shrugs [11:47] <scop> anyway, I should have left already 15 minutes ago, can't hang around any longer tonight [11:47] <abadger1999> later scop [11:47] <scop> later [11:47] <-- scop has left this channel ("Leaving"). [11:49] <tibbs> Well, too bad, I guess. [11:49] <-- bhanupatial has left this channel. [11:51] <tibbs> I'm going to wander off unless someone has anything else. [11:53] <abadger1999> not I.