Wiki

Clone wiki

SCons / BugParty / IrcLog2008-08-18

17:27:40 * garyo-home (n=chatzill@209-6-158-38.c3-0.smr-ubr3.sbo-smr.ma.cable.rcn.com) has joined #scons 18:58:18 <garyo-home> Hi folks; who's here so far? 19:00:12 * stevenknight (n=stevenkn@c-98-234-62-147.hsd1.ca.comcast.net) has joined #scons 19:00:27 <stevenknight> hello again 19:00:39 <garyo-home> Hi Steven. 19:02:16 <stevenknight> do we have a quorum? 19:02:25 <garyo-home> Just me and you so far I think. 19:02:31 <stevenknight> fair enough 19:02:52 <garyo-home> We can look at the ParseConfig one. Good patience with that one! 19:03:01 <garyo-home> (1500) 19:04:56 <stevenknight> thnx 19:05:26 <garyo-home> Cygwin actually does the same thing. 19:05:27 <stevenknight> greg's got a point re: translating paths that come from anywhere outside, not just ParseConfig 19:05:29 * Greg_Noel is no longer marked as being away 19:05:31 <Greg_Noel> Hey, it's after 19h00; I'm late 19:05:41 <garyo-home> Hi Greg. 19:05:44 <stevenknight> so it is... :-) 19:06:01 <garyo-home> We're just looking at 1500 (ParseConfig and msys) 19:06:30 <garyo-home> If the mapping can always be done backwards, we could just have a .msys_path like .win32_path. 19:06:43 <Greg_Noel> OK, I'm there. 19:06:47 <stevenknight> no, that's the reverse of what he wants 19:06:55 <stevenknight> this is on input, not output 19:07:24 <stevenknight> i agree that a .msys path might be useful, but that's orthogonal 19:07:45 <stevenknight> right now i'm leaning towards 2.0 19:07:45 <garyo-home> Maybe not though. On msys, shouldn't gcc be called with all msys paths? 19:07:49 <Greg_Noel> If I understand what Steven wrote, it should be done when the name is looked up (i.e., in Entry()) 19:08:16 <stevenknight> i think Greg's right in larger sense 19:08:26 <stevenknight> but i'm more nervous about unintended side effects there 19:08:42 <stevenknight> for now, i think we mark it 2.0 p3 19:08:46 <garyo-home> Yes, you need to translate msys -> win32, but maybe also win32 -> msys. 19:08:52 <Greg_Noel> I'll go with that. 19:08:54 <stevenknight> and the implementation can be hashed out by whoever gets stuck with it 19:08:57 <garyo-home> But anyway I agree w/ 2.0 p3 unknown person. 19:09:06 <stevenknight> garyo, you might be right re: needing to pass msys paths to some tools 19:09:19 <stevenknight> done 19:09:26 <stevenknight> 1500: 2.0 p3 draft_choice_to_be_named_later 19:09:42 <garyo-home> 2176 consensus 1.0.x p2 ludwig? 19:09:42 <Greg_Noel> Ludwig 19:09:43 <stevenknight> cool 19:09:48 <Greg_Noel> He's said he plans to hang around 19:10:12 <stevenknight> 2176 done 19:10:17 <Greg_Noel> 2177: Don't get me wrong; I think it should be applied ASAP. It's just that niggling little thing about the policy. Steve's point about the policy evolving is a good one, and I have to admit that the rebuild would be for a pretty rare corner case, so I'm not going to argue for 2.0 very hard (i.e., not really at all). And let's not forget that it gives the user a very cool way of seeing what SCons thinks is in the directory. I do worry about it not being sorted; I think the test harness has sorting turned on and I don't know if it's on normally. 19:11:02 <garyo-home> Sorting is vital. Exceptions are OK as long as they're properly documented, IMHO. 19:11:17 <garyo-home> (Exceptions... to the no-rebuild policy) 19:11:26 <stevenknight> agree re: sorting 19:11:52 <stevenknight> and re: exceptions 19:12:04 <stevenknight> my only hesitation is that it does start down the slippery slope 19:12:07 <Greg_Noel> I think we need to apply the patch and then try it out under some not-test conditions to see if the names are sorted. 19:12:09 <garyo-home> If no-rebuild were really a showstopper for someone we could show them how to turn it off. 19:12:22 <garyo-home> turn this feature off. 19:12:27 <stevenknight> right 19:12:48 <stevenknight> but turn it off by making the sorting configurable? or by showing them how to modify that spot in the code? 19:12:55 <stevenknight> making it configurable complicates applying the patch a little... 19:13:15 <garyo-home> IMHO: modify the code. I think nobody will complain and we won't have to do it. Configurability would be a waste. 19:13:20 <Greg_Noel> No, only sort when calculating the contents, not when using the children. 19:14:09 <stevenknight> hmm, i think we've had this discussion re: directory sorting before 19:14:10 <Greg_Noel> I agree with Gary FOR THIS CASE. 19:14:42 <Greg_Noel> It's only significant for directories 19:15:07 <Greg_Noel> the order of the children doesn't matter 19:15:05 <garyo-home> I'm not worried about slippery slope; as long as we try hard not to cause rebuilds, then most releases won't (because of the policy). If we end up with too many of those, we batch them up. 19:15:25 <stevenknight> garyo: agreed 19:15:39 <garyo-home> greg: right, like the order of children of an Alias doesn't matter. Just the sig. I agree. 19:15:49 <stevenknight> okay, i'm with you guys 19:15:58 <garyo-home> OK, sounds like we're all on board for 1.0.x p2(or p1) 19:16:06 <stevenknight> 2177: 1.0.x p[12] anyone 19:16:17 <Greg_Noel> p2, p1 is for emergencies 19:16:24 <stevenknight> suppose we ought to make "anyone" more specific.... 19:16:25 <Greg_Noel> Ludwig 19:16:40 <garyo-home> either will get done (p1 or p2), I say p2. 19:16:41 <stevenknight> Ludwig++ 19:16:46 <garyo-home> fine. 19:16:49 <stevenknight> 1.0.x p2 Ludwig 19:16:49 <Greg_Noel> done 19:16:50 <stevenknight> done 19:16:54 <stevenknight> 2178: 19:17:05 <Greg_Noel> consensus 19:17:17 <garyo-home> sure. p3, split the difference? 19:17:24 <stevenknight> 2.0 p3 Ludwig 19:17:25 <stevenknight> done 19:17:31 <stevenknight> 2179: 19:17:35 <Greg_Noel> Will the 2.3 floor be in 1.0.1? 19:17:43 <garyo-home> whole rest of the page is consensus. 19:17:50 <stevenknight> cool 19:17:51 <stevenknight> go us 19:18:02 <garyo-home> Greg: I didn't think so... 19:18:08 <stevenknight> no, updating floor is what 2.0 is about 19:18:19 <Greg_Noel> warning 19:18:26 <garyo-home> But we could start warning maybe? 19:18:35 <stevenknight> we're already warning 19:18:37 <Greg_Noel> er, will the 2.3 warning be in 1.0.1? 19:18:48 <stevenknight> if we've settled on 2.3 19:18:49 <Greg_Noel> but for 2.2 19:19:12 <stevenknight> i haven't yet been compelled that 2.3 gives us that much more than 2.2 19:19:14 <garyo-home> I'm staying out of this one this time. 19:19:30 <stevenknight> so far it's the tarfile module and a couple of Ludwig's patches 19:19:36 <stevenknight> am i overlooking anything? 19:19:48 <garyo-home> tarfile is not insignificant though. 19:19:59 <stevenknight> true 19:20:07 <Greg_Noel> Well, it seems to be the floor for a lot of other Python-based projects; go with the flow 19:20:25 <stevenknight> why start now? :-) 19:20:32 <Greg_Noel> point 19:20:48 <bdbaddog> I vote for 2.3 19:20:51 <bdbaddog> :) 19:20:57 <Greg_Noel> Hi, Bill 19:21:00 <stevenknight> i'm not balking at tarfile per se 19:21:01 <garyo-home> Google says 19000 hits for "oldest supported python version 2.2" but 96,000 for "... 2.3". :-) 19:21:16 <stevenknight> it's the idea that we're going to let what other modules do/don't use decide for us 19:21:35 <stevenknight> instead of making the decision based on the actual underlying features supported 19:21:40 <stevenknight> (or by real user data, which we don't have) 19:22:22 <stevenknight> and for "oldest supported python version 1.5." one hit? us? 19:22:26 <bdbaddog> I still haven't gotten any responses in the couple of emails I floated to the user mailing list indicating that they were unable to move forward to any particular version of python for use by the build tools. 19:22:30 <stevenknight> :-) 19:22:37 <Greg_Noel> The problem is that it's a decision based on positioning from next year; all we can do is guess. 19:22:58 <Greg_Noel> I just think 2.3 will be a better floor by then. 19:23:37 <stevenknight> well, shall we open up the "2.0 time frame" can of worms too, then? 19:23:52 <bdbaddog> even if it was today. I think 2.3 is a good floor. what distros do you exclude? Also take into account that any projects (opensource) which will use scons to build themselves will only be released with newer versions of current distros. 19:23:56 <stevenknight> greg, I know you have it in mind for ~6 months from now, right? 19:23:58 <Greg_Noel> No, let's drop it. I was just curious. 19:24:22 <Greg_Noel> 2.0 in six months? About right. 19:24:30 <garyo-home> bdbaddog: RHEL3 I think has 2.2. But that was 2003 or earlier. 19:25:10 <bdbaddog> yup, and as I said, no projects which are on rhel3 will be updated, so requiring newer python won't be barier to distro builders. 19:25:24 <Greg_Noel> Figure one month for 1.0.1 and another for 1.0.2; two months for 1.1; two more for 2.0 19:25:39 <garyo-home> at a bare minimum! 19:26:03 <Greg_Noel> yeah, that's pushing it, all right, but it's what you've said you want to do. 19:25:56 <bdbaddog> you see 2.0 as a linear progression from 1.1 ? 19:26:18 <garyo-home> me? 19:26:20 <Greg_Noel> How couldn't it be? 19:26:33 <Greg_Noel> you == you guys 19:26:46 <bdbaddog> (you = greg noel) 19:26:49 <bdbaddog> sorry bout that. 19:27:21 <garyo-home> greg: I think it's aggressive, but let's see what our newly expanded team can get done. 19:27:37 <Greg_Noel> No, I originally figured a year for 1.x cycles; you guys have wanted it faster. 19:27:22 <bdbaddog> I see 2.0 branched from 1.0, work removing 1.5.2 limitations started, in parallel 1.0.1 and 1.1 getting merged to 2.0. 2.0 in a few months. 19:28:10 <Greg_Noel> bdbaddog, ain't gonna happen; not enough resources to work two branches. 19:28:16 <garyo-home> bdbaddog: +1 on branching for 2.0 soonish. But not too soon because of possible merge headaches. 19:28:19 <stevenknight> but a year based on...? a specific set of features that merit 2.0? fixing "all" of the 1.x issues? 19:29:00 <Greg_Noel> There are too many issues in 1.x now; fixing all of them would take two years. 19:29:10 <garyo-home> let's focus on some bugs now. I don't think we can decide 2.0 release date tonight. 19:29:11 <stevenknight> agreed 19:29:16 <stevenknight> but 19:29:20 <stevenknight> okay, garyo's right 19:29:24 <stevenknight> take this to the ML 19:29:28 <Greg_Noel> works 19:29:29 <bdbaddog> :) 19:29:45 <stevenknight> where were we? 19:29:52 <garyo-home> The 2006h1 sheet 19:30:01 <bdbaddog> should I float another can you move to 2.2,2.3,2.4,2.5 python for a new release mail? 19:30:06 <Greg_Noel> 1377 19:30:12 <Greg_Noel> oops, 1388 19:30:19 <stevenknight> bdbaddog: hold off, don't think we'll get much better info than last time 19:30:21 <garyo-home> bdbaddog: can only help to get us some real user info. 19:30:38 <garyo-home> I defer to steven. 19:31:06 <garyo-home> Greg: is there a Poll module for moinmoin? 19:31:22 <Greg_Noel> Um, not sure; I think there's something 19:31:48 <stevenknight> doesn't it RSS feed? I think i've seen Google Reader offer to let me subscribe 19:32:00 <garyo-home> anyway, 1388 should be 2.0 unless it's a special case that can be addressed earlier. 19:32:03 <stevenknight> or maybe that's GR polling... 19:32:16 <garyo-home> stevenknight: sorry, I meant a way to take a user poll with questions and a graph of results. 19:32:17 <stevenknight> yes, 2.0 19:32:18 <stevenknight> p2 19:32:25 <stevenknight> ah 19:32:38 <stevenknight> i don't know what I was thinking, Unicode getting "solved" in 1.x... 19:32:42 <Greg_Noel> 1388 needs to know what the system encoding is; that's a function in 2.3 19:32:47 <stevenknight> 1388: 2.0 p2 hero 19:32:53 <garyo-home> ok, 1388 2.0 p2 someone. 19:32:56 <Greg_Noel> done {;-} 19:33:13 <garyo-home> 1392: Greg, look again? 19:33:25 <stevenknight> 1392: 1.x p3 me 19:33:40 <Greg_Noel> garyo-home, look at what? 19:33:43 <garyo-home> (oh, you already did.) 19:34:00 <Greg_Noel> done 19:34:01 <garyo-home> never mind. I think you had an out-of-place line in the ssheet or something. Or maybe it was me. 19:34:19 <Greg_Noel> I moved it just before we started. 19:34:34 <garyo-home> 1399 closed, 1409? 19:34:38 <garyo-home> Sorry, 1402 19:35:05 <garyo-home> Steven: I'd really like the vstudio stuff. There's some work already done. Should it wait for 1.x? 19:35:15 <garyo-home> (vsvars.bat etc.) 19:35:46 <stevenknight> garyo-home: pretty sure it won't be ready in the 1.0.x time frame 19:35:56 <stevenknight> unless you're referring to specific issues with patches that could integrated early 19:36:01 <garyo-home> ok, 1.x then. It makes more sense there anyway. 19:36:19 <stevenknight> 1402: 1.x p2 me? 19:36:22 <garyo-home> No, I mean ripping out all the registry junk and just parsing the output of vs/vcvars.bat. 19:36:31 <Greg_Noel> all the VisualStudio stuff is 'anytime' 19:36:32 <garyo-home> 1402: 1.x p2 steven. 19:36:36 <stevenknight> that would be a god intermediate step 19:36:55 <stevenknight> good 19:36:59 <stevenknight> 1406: 19:37:22 <garyo-home> I don't like the submitted patch. 19:37:28 <garyo-home> Too specialized. 19:37:48 <stevenknight> agreed 19:37:54 <garyo-home> Has anyone retried it recently to see if it's still broken? 19:37:58 <stevenknight> research jim 19:38:03 <garyo-home> OK. 19:38:05 <Greg_Noel> done 19:38:10 <stevenknight> 1417: 19:38:31 <stevenknight> research me VisualStudio keyword 19:38:37 <Greg_Noel> anytime, stevenknight, VisualStudio 19:38:45 <Greg_Noel> all the VisualStudio stuff is 'anytime' 19:39:04 <Greg_Noel> meaning not tied to a release schedule 19:39:00 <stevenknight> done 19:39:06 <stevenknight> 1418: research david 19:39:15 <stevenknight> Greg_Noel: right 19:39:19 <garyo-home> That's not 100% right, the actual problem is you can't use $SOURCES.(any-attr) if your source list may be empty. 19:39:20 <Greg_Noel> done 19:39:28 <garyo-home> Sorry I'm still on 1417 here. 19:39:55 <garyo-home> I'm OK w/ 1418 research david though. 19:40:36 <Greg_Noel> 1417: garyo-home, is that true? 19:41:17 <garyo-home> Sure, because it expands to an empty list (None) which has no .windows attribute. 19:41:35 <Greg_Noel> Not empty list []? 19:41:52 <stevenknight> ouch, that fact eluded me. that's bad 19:41:53 <garyo-home> [].windows wouldn't work either. 19:42:01 <Greg_Noel> That could be an easy fix, if that's all that's wrong. 19:42:17 <garyo-home> Expand to something other than None or emptylist? 19:42:27 <Greg_Noel> Doesn't it iterate over the list, applying the attribute? 19:42:53 <garyo-home> Don't know. Steven? 19:42:56 <stevenknight> it should 19:43:08 <stevenknight> they're supposed to be a list like object that does that 19:43:15 <stevenknight> but i could see that the expansion order might be wrong 19:43:28 <Greg_Noel> research, stevenknight? 19:43:28 <stevenknight> and it might "efficiently" return an actual list, not the object 19:43:46 <stevenknight> at some point where it then tries to apply the attribute too late 19:43:49 <stevenknight> yes, research stevenknight 19:43:53 <garyo-home> ok 19:43:55 <Greg_Noel> done 19:44:03 <Greg_Noel> That endeth this list... 19:44:10 <garyo-home> nice! 19:44:11 <stevenknight> garyo: good catch 19:44:14 <garyo-home> thx 19:44:21 <stevenknight> all right, done with 2006 19:44:36 <Greg_Noel> garyo++ 19:44:50 <stevenknight> everyone have time to wade into the 1.{0.x,x} re-triaging? 19:44:54 <bdbaddog> gotta run. 19:44:58 * bdbaddog (n=bdeegan@adsl-71-131-3-114.dsl.sntc01.pacbell.net) has left #scons 19:44:59 <garyo-home> Sure, a bit. 19:45:10 <Greg_Noel> 'nite, Bill 19:45:31 <Greg_Noel> I have all night, until my wife wants to watch Olympics again... 19:47:48 <Greg_Noel> Gary, Steven and I seem to be in agreement for most of the 1.0 retriage issues; if you concur, we're done. 19:47:54 <stevenknight> greg+gary: looks like a lot of consensus in the 1.0.x retriage 19:47:57 <stevenknight> done 19:48:04 <garyo-home> What do we need for 2071: do we make up our own or use an existing one? I have an old emacs release I signed for FSF we could use. 19:48:17 <garyo-home> From my quick look, I don't disagree w/ any of them. 19:48:44 <stevenknight> 2071: there's an existing SCons release that assigns ownership to SCons Foundation 19:48:51 <stevenknight> two versions, one for employer, one for employee 19:48:59 <stevenknight> because it assigns ownership it scares corporate lawyers 19:49:20 <garyo-home> great. So it's just getting people signed up? 19:49:31 <stevenknight> if we go with our current one 19:49:36 <garyo-home> I think I already signed one, now that I think about it. ??? 19:49:47 <stevenknight> we can make it less scary by finding a version that lets them just license to us under our same MIT trms 19:49:52 <stevenknight> garyo-home: yes, youdid 19:50:13 <stevenknight> needs some research time to google for examples of other projects who have done that 19:50:26 <stevenknight> and maybe run it by our SFLC lawyer 19:50:34 <Greg_Noel> I feel left out; nobody asked me ;-( 19:50:45 <garyo-home> it was a long long time ago. 19:51:08 <stevenknight> yes, it was enough of a bother relative to the benefit we were getting that i got lazy and dropped it 19:51:24 <Greg_Noel> There are several samples on the net; I found one for Python and Mozilla pretty easily. 19:51:25 <stevenknight> i mean, it involves talking to lawyers... :-/ 19:51:49 <stevenknight> i thought those were ownership agreements 19:51:53 <stevenknight> bt that's from a long time ago 19:52:02 <stevenknight> so i could be out of date 19:52:43 <garyo-home> ok, well glad we have something at least. 19:52:51 <Greg_Noel> IANAL, but I think all it has to do is agree to license under MIT terms 19:53:00 <stevenknight> yes, that's right 19:53:23 <garyo-home> Wow, there's quite a long retriage list for 1.0.x. 19:54:00 <stevenknight> yeah 19:55:28 <stevenknight> i'm thinking we should take this part off line 19:55:33 <garyo-home> too much for me tonight. 19:55:34 <stevenknight> rather than get started on it right now 19:55:37 <stevenknight> exactly 19:55:50 <stevenknight> greg, i see you've started on the spreadsheet 19:55:54 <stevenknight> if you can do your update 19:55:56 <Greg_Noel> Maybe meet again tomorrow? 19:56:05 <garyo-home> I'll work on that for next week though, as well as actually getting some contributions in :-/ 19:56:11 <stevenknight> and gary, if you have time to at least skim it in the next day or so for any places where you disagree with greg+me? 19:56:17 <Greg_Noel> Actually, I've been through it once, but I only annotated those I thought should change 19:56:19 <garyo-home> yes, definitely. 19:56:21 <stevenknight> me too re: contributions 19:56:42 <garyo-home> good. Next week, same time, same place? 19:56:59 <stevenknight> works for me, so far as i know 19:57:03 <Greg_Noel> If there's a consensus of three, whoever's the third should just do it. 19:57:05 <stevenknight> we're going to be in the middle of moving again, unfortunately 19:57:13 <Greg_Noel> again? 19:57:15 <stevenknight> Greg_Noel: just do it ++ 19:57:19 <stevenknight> yes 19:57:24 <stevenknight> our landlord is in foreclosure 19:57:25 <garyo-home> makes sense to me too. 19:57:27 <stevenknight> welcome to california 19:57:39 <Greg_Noel> But don't californicate... 19:57:40 <garyo-home> Steven: right, you were wrestling with that last week. 19:57:44 <stevenknight> LOL 19:57:49 <garyo-home> Too bad! 19:58:07 <garyo-home> Maybe your new place will be better anyway? 19:58:07 <stevenknight> yeah, we ended up finding a better place, but still have the hassle now of actually moving 19:58:12 <garyo-home> :-) 19:58:38 <garyo-home> well good luck with it. 19:58:46 <stevenknight> thnx 19:58:51 <Greg_Noel> When's the deadline to move? Should we consider delaying 1.0.1 until the end of the week? 19:58:51 <garyo-home> I'm going to get some sleep now, g' night. 19:59:06 <stevenknight> no, keep it on schedule 19:59:21 <Greg_Noel> ok, that makes triaging harder, but we'll do what we can. 19:59:34 <Greg_Noel> Gary, you still there? 19:59:40 <garyo-home> yes. 20:00:12 <Greg_Noel> Oh, I already said that; old age: I'll just repeat that if you form the consensus, just mark the issue. 20:00:21 <garyo-home> Will do. 20:00:29 <Greg_Noel> then g'night all 20:00:35 <garyo-home> ok, c u later. 20:00:44 <stevenknight> okay, i'm going to go see if i can find the right combination of ubuntu hardy packges to build the doc... 20:00:47 * Greg_Noel has been marked as being away 20:00:54 <stevenknight> and then maybe get a checkpoint release out 20:01:00 <stevenknight> 'night 20:01:09 * stevenknight has quit ("Leaving") 20:01:15 <garyo-home> Steven: what about my pkg list on the wiki? 20:01:18 <garyo-home> Oh well, he's gone. 20:01:21 <Greg_Noel> too late 20:01:30 <garyo-home> ok, bye now. 20:01:35 <Greg_Noel> 'nite 20:01:40 * garyo-home has quit ("ChatZilla 0.9.83 [Firefox 3.0.1/2008070208]")

Updated