Wiki

Clone wiki

SCons / BugParty / IrcLog2008-09-22

18:48:49 * garyo-home (n=chatzill@209-6-158-38.c3-0.smr-ubr3.sbo-smr.ma.cable.rcn.com) has joined #scons 19:04:12 * stevenknight (n=stevenkn@c-67-169-176-171.hsd1.ca.comcast.net) has joined #scons 19:04:21 <stevenknight> hey garyo-home 19:04:28 * GregNoel is no longer marked as being away 19:04:37 <stevenknight> anyone else here? 19:04:48 <GregNoel> Hey, I just got here; give me a minute. 19:04:54 <garyo-home> Hi Steven. I'm trying to get some spreadsheet bugs marked up, but am too slow :-( 19:05:01 <garyo-home> Hi, Greg. 19:05:23 <stevenknight> i just got home 19:05:31 <stevenknight> have to walk the dog first... 19:05:38 <stevenknight> should be ~10 minutes 19:05:50 <stevenknight> do what you can to start and i'll catch up 19:05:49 <garyo-home> OK, I'll work on my spreadsheet comments 19:05:56 <stevenknight> thanks 19:14:07 <garyo-home> OK, I'm more or less ready; Greg are you there? 19:17:25 <garyo-home> Anyone? 19:17:56 <GregNoel> Hi, I went to get some tea... 19:18:28 <garyo-home> Hi. Shall we start? 19:19:03 <GregNoel> sure 19:19:13 <GregNoel> 2198 19:19:37 <GregNoel> I don't understand your comment 19:19:53 <garyo-home> Can we get the OP to add the 1.5.2 changes too? 19:20:13 <stevenknight> back 19:20:15 <GregNoel> If he has one. 1.5.2 is getting pretty rare. 19:21:06 <garyo-home> I just mean to modify all the files he mentioned, to get rid of .sources and .implicit -- that's what he's proposing, right? 19:21:06 <stevenknight> maybe if it doesn't apply cleanly to 1.5.2 we delay to 2.x 19:21:10 <stevenknight> more incentive to get it out 19:21:24 <GregNoel> hmmm... not a bad idea 19:21:35 <garyo-home> From looking at the bug, I think that's better 19:21:47 <garyo-home> 2.0 Ludwig p3? 19:22:17 <GregNoel> do you mean 2.0 or 2.x? 19:22:23 <garyo-home> 2.x is fine. 19:22:28 <GregNoel> works 19:22:36 <stevenknight> 2.x, with a note re: pulling it in if works w/1.5.2 19:22:45 <GregNoel> done 19:22:50 <garyo-home> I think he says it won't. 19:22:51 <GregNoel> 2199 19:23:12 <garyo-home> Your gmake file manip functions. I like it. 19:23:32 <GregNoel> I'm a bit reluctant to commit to 1.x, but I can give it a shot 19:23:50 <garyo-home> Good; if it slips, it slips. No big deal -- a nice-to-have feature. 19:24:09 <GregNoel> OK, 1.x p3 or p4? 19:24:22 <stevenknight> p4 if it's that slippable 19:24:27 <garyo-home> Seems like there's a lot in 1.x already... 19:24:34 <stevenknight> yep 19:24:37 <GregNoel> yeah 19:24:48 <garyo-home> ok, 1.x p4 greg. 19:24:54 <GregNoel> done 19:25:14 <garyo-home> 2200 you guys want in 1.x too, but I'm skeptical. 19:25:24 <GregNoel> I'm swayed by Steven's argument. 19:25:52 <garyo-home> Would any Execute clear all node caches? 19:26:06 <GregNoel> I agree that #2 would be more work, so I'll go for 2.x 19:26:13 <garyo-home> (Do we even really care about caching during reading SConscripts?) 19:26:27 <garyo-home> OK, 2.x p3 Ludwig then? 19:26:29 <GregNoel> Yes, any Execute() of an Action we don't provide. 19:26:45 <stevenknight> i think any Execute() should clear its targets 19:26:52 <GregNoel> (Yes, we care; it avoids disk hits, which cost time.) 19:27:00 <GregNoel> Execute() has no targets. 19:27:02 <garyo-home> But we won't know the targets w/ proposal #2. 19:27:12 <stevenknight> right, that's why i favor proposal #2 19:27:29 <stevenknight> wait, i see what you mean 19:27:33 <stevenknight> i read that as "w/out" 19:27:38 <stevenknight> hmm 19:28:30 <garyo-home> maybe we shouldn't design it here, just mark as "needs discussion" or something 19:28:42 <GregNoel> agree, but where? 19:29:02 <garyo-home> 2.x p3 (mark in the issue itself, in the text) 19:29:55 <GregNoel> OK, I'll make myself a CC 19:29:58 <stevenknight> (sorry, interrupt) 19:30:09 <GregNoel> Should I add you two as well? 19:30:14 <stevenknight> okay, 2.x p3 19:30:17 <stevenknight> yes, cc me 19:30:24 <garyo-home> ok. Or just paste in this section of the irc log :-) 19:30:30 <GregNoel> That, too 19:30:43 <stevenknight> okay, done? 19:30:49 <GregNoel> I'd like it a bit sooner; maybe 2.x p2? 19:30:57 <stevenknight> i can go with p2 19:31:00 <garyo-home> ok w/ me 19:31:03 <GregNoel> done 19:31:12 <GregNoel> 2201 19:31:13 <garyo-home> 2201: I just asked the OP for a testcase. 19:32:07 <GregNoel> It's not clear to me what he wants; Nodes have cached info, not Executors. 19:32:13 <garyo-home> Steven, if you think it's a real issue maybe 1.x/p3/Ludwig is the way to go. 19:32:25 <GregNoel> Let's see what he offers and review it next time. 19:32:36 <garyo-home> ok w/ me. 19:32:40 <garyo-home> defer. 19:33:05 <GregNoel> stevenknight, OK? 19:33:18 <stevenknight> (sorry, more interrupts) 19:33:37 <garyo-home> I think so 19:33:38 <stevenknight> well, just because i think it's an issue doesn't mean it is 19:33:41 <stevenknight> i haven't done the triage 19:34:00 <GregNoel> Let's look at it again next time, then 19:34:01 <stevenknight> so getting a test case seems like it should come before committing the resources 19:34:14 <stevenknight> done, ask for test case and re-triage 19:34:18 <GregNoel> done 19:34:21 <GregNoel> 2204 19:34:23 <garyo-home> 2204: I like Steven's idea of just a better error with more context. 19:35:34 <GregNoel> I'm not sure I understand what Seven is suggesting 19:35:20 <garyo-home> We need more developers. 19:35:31 <stevenknight> garyo-home: agreed 19:35:36 <GregNoel> yes 19:35:53 <GregNoel> Is Mati still around? 19:36:14 <garyo-home> Good question. I'll look him up. 19:37:08 <GregNoel> Steven, if you know how to fix 2204, do you really think it should be 1.2? 19:37:19 <stevenknight> GregNoel: all those errors come through the env.arg2nodes() methods that translate strings to Nodes 19:37:39 <stevenknight> that provides a place for some code to catch the mismatch between factory and returned Node 19:38:02 <garyo-home> At least print some context to help the user find the bug. 19:38:21 <stevenknight> right now the underlying lookup that catches it return a normal Python TypeError because I thought I was being "pythonic" in throwing TypeError for those mismatches 19:38:25 <GregNoel> Hmmm... OK, I'll trust you. a2n is called a lot, so it can't be slowed down too much. 19:38:35 <stevenknight> fair point 19:38:44 <stevenknight> 1.2, p3, me? 19:38:51 <GregNoel> done 19:38:51 <garyo-home> Right, but he's just going to catch the TypeError and throw something else. No slowdown. 19:38:59 <stevenknight> done 19:39:01 <stevenknight> 2205: 19:39:30 <GregNoel> 2205: dup 1957? 19:39:36 <stevenknight> there's too much in 1.x, i'm going with you guys on 2.x 19:40:15 <garyo-home> I think it has to be 2.x unfortunately. Could be dup of 1957; there's a patch in 1957 too, not sure of its quality 19:40:23 <garyo-home> (I didn't really look at it) 19:40:30 <stevenknight> it's pretty extensive 19:40:38 <stevenknight> not clear to me it's a dup without more triage 19:40:44 <GregNoel> defer? 19:41:22 <garyo-home> I think it is a dup, at least 1957 tries to convert all exceptions into BuildFailures, and 2205 is an exception coming through. 19:41:33 <garyo-home> (TaskmasterException) 19:43:08 <GregNoel> Either defer for more triage or Gary for research? 19:43:20 <stevenknight> defer 19:43:40 <garyo-home> ok, I'll research and if it's a dup I'll mark it as such. 19:44:02 <stevenknight> thanks 19:44:06 <GregNoel> done 19:44:26 <GregNoel> 2207 19:44:31 <garyo-home> 2207: works as designed imho 19:45:14 <GregNoel> Yeah, but IOError is not cool 19:45:29 <stevenknight> agree w/Greg 19:45:44 <stevenknight> stack traces scare users 19:46:06 <garyo-home> ... but finding everywhere scons opens a file and seeing if the i/o err is due to dangling symlink could be a big mess. 19:46:09 <GregNoel> It probably should throw the same error that the Taskmaster throws for a missing source 19:46:43 <stevenknight> right, but it doesn't have to be absolutely everywhere 19:46:44 <garyo-home> Actually I claim a missing source err would be even more confusing to users. 19:46:51 <stevenknight> just take care of this one and make the world that much better 19:47:12 <garyo-home> mumble, ok I guess 19:47:16 <stevenknight> there aren't that many places where do a direct open like this 19:47:26 <stevenknight> usually it's under get_contents() or something 19:47:35 <garyo-home> good point 19:47:50 <stevenknight> okay, done 19:47:53 <GregNoel> done 19:48:01 <GregNoel> er, wait 19:48:08 <garyo-home> steven? 19:48:11 <stevenknight> yes? 19:48:19 <GregNoel> did we agree on milestone and priority? 19:48:22 <garyo-home> are you taking it? 19:48:24 <stevenknight> oh 19:48:32 <stevenknight> yeah, i'll take it 19:48:38 <GregNoel> ok, when? 19:48:50 <stevenknight> 1.x p4 19:48:53 <GregNoel> done 19:49:04 <GregNoel> 2208 19:49:36 <GregNoel> I'd like to know what the performance impact will be, but I like the idea of a warning that's always on. 19:49:45 <GregNoel> er, defaults to on. 19:50:32 <garyo-home> agree w/ greg 19:50:53 <GregNoel> is it a dup of 'ancient bug'? 19:51:13 <stevenknight> 'ancient bug'? 19:51:23 <garyo-home> let me see... 19:51:33 <GregNoel> Gary says he filed an 'ancient bug' on it. 19:52:30 <stevenknight> ah 19:52:29 <garyo-home> Can't find it anymore. 19:52:52 <stevenknight> i think 1.x, p3, me 19:52:59 <GregNoel> OK, if we find it later, we'll worry about it then. 19:53:03 <GregNoel> done 19:53:15 <stevenknight> note re: vaidating performance impact before submitting 19:53:21 <GregNoel> right 19:53:40 <GregNoel> 2209 19:54:10 <stevenknight> future 19:54:36 <GregNoel> it's in script support, so it may not even be relevant to other front-ends 19:54:40 <garyo-home> future p4 ok w/ me, unless I'm missing something 19:54:50 <GregNoel> future p4 it is 19:55:02 <GregNoel> 2210 19:55:11 <stevenknight> shoot, i'm going to have to stop soon 19:55:20 <garyo-home> me too, sorry. 19:55:29 <garyo-home> We can do a few more 19:55:50 <garyo-home> 2210: future p2? 19:55:53 <stevenknight> 2210: future p2 19:55:55 <stevenknight> done 19:55:55 <GregNoel> I'll take it as anytime, but not soon 19:56:01 <garyo-home> 2211 anytime p5 steven? 19:56:24 <stevenknight> done 19:56:28 <GregNoel> done 19:56:40 <garyo-home> 2212 is vs_revamp 19:56:50 <stevenknight> yes, vs_revamp 19:56:56 <garyo-home> soon I hope 19:56:58 <GregNoel> How should I mark it? 19:57:10 <GregNoel> dup of something? 19:57:28 <stevenknight> no, 1.x p3, add VisualStudio keyword 19:57:33 <stevenknight> put david's name on it 19:57:33 <GregNoel> done 19:57:51 <stevenknight> 2213: 2.x p4 19:58:11 <GregNoel> done 19:58:20 <stevenknight> 2215 i just closed 19:58:32 <GregNoel> no, open is a function; file is a type. 19:58:49 <GregNoel> It may not be obvious, but there's a distinction. 19:59:01 <stevenknight> yes, but i submit it's not crucial for 1.5.2 compatibility 19:59:09 <stevenknight> which only has old-style classes anyway 19:59:12 <GregNoel> good point. 19:59:17 <GregNoel> OK, done 19:59:31 <stevenknight> all right, gotta run 19:59:33 <GregNoel> last one 19:59:42 <stevenknight> right 19:59:48 <garyo-home> 2216? 19:59:47 <stevenknight> 2216: anytime p5 me 19:59:53 <GregNoel> done 20:00:06 <stevenknight> okay, catch you guys next week 20:00:21 <garyo-home> ok, sounds good. Will try to get my 1.1 issues done! 20:00:22 <GregNoel> Next week, or should we go for two weeks? 20:00:31 <stevenknight> i'll probably send email re: trying for a different time 20:00:38 <GregNoel> When's the RC coming out? 20:00:42 <stevenknight> my transportation situation has changed a bit so this time isn't working as well 20:00:56 <garyo-home> actually now that you mention it every week is beginning to be a family issue :-/ 20:00:56 <GregNoel> OK, we'll look for the message 20:01:01 <stevenknight> is that this week? (you'd think i'd read the roadmap... :-)) 20:01:18 <GregNoel> yes, next week is supposed to be 1.1 20:01:20 <stevenknight> yeah, i'm not popular at home on Monday nights... :-( 20:01:31 <GregNoel> Charger football!!! 20:01:33 <stevenknight> okay, i'll get an RC out in the next day or two 20:01:45 <garyo-home> we can discuss scheduling bug parties later... 20:01:47 <stevenknight> oh, man, Greg, sorry things are starting so rocky for you guys 20:02:04 <GregNoel> don't tell me; it's being TiVoed 20:02:18 <stevenknight> i meant the season, not tonight 20:02:22 <stevenknight> tough losses 20:02:27 <stevenknight> especially to ^AS#(*& Denver 20:02:43 <GregNoel> yeah, two losses by a total of three points. There ought to be a law... 20:03:01 <GregNoel> Or we should kill all the lawyers, as Shakespere says (the ref is a lawyer)... 20:02:57 <stevenknight> all right, i'm out of here 20:02:59 <stevenknight> later... 20:03:06 * stevenknight has quit ("Leaving") 20:03:08 <GregNoel> G'night 20:03:38 <GregNoel> I'm going to go watch the game; cu later. 20:03:46 <garyo-home> bye 20:03:55 * GregNoel has been marked as being away 20:04:10 * garyo-home has quit ("ChatZilla 0.9.83 [Firefox 3.0.1/2008070208]")

Updated