13:05:55 * Azverkan (n=fakeuser@209.172.105.155) has joined #scons
16:48:00 * stevenknight (n=stevenkn@nat/google/x-46e723d28b1b3479) has joined #scons
16:55:25 stevenknight GregNoel: are you here yet?
16:55:53 * garyo-home (n=chatzill@209-6-158-38.c3-0.smr-ubr3.sbo-smr.ma.cable.rcn.com) has joined #scons
16:58:09 garyo-home hi folks
16:58:45 stevenknight hi gary
16:58:57 stevenknight Greg doesn't seem to be here yet
16:59:55 garyo-home ok, I'm just starting in on the spreadsheet.
16:59:55 GregNoel Hello, am I late?
17:00:01 garyo-home nope, just in time.
17:00:09 stevenknight hi Greg
17:00:26 GregNoel We have relatives in town; we were out with them and just got back.
17:00:41 stevenknight ah, special thanks for making it, then
17:00:56 GregNoel Give me a sec to get set up
17:00:47 stevenknight hope we're not taking you away too much
17:01:06 GregNoel No, we'll see them again tomorrow.
17:02:13 stevenknight BTW, on the ReadWrite page, the 2007 Q1 link takes you to the same sheet as the Q4 link
17:02:16 stevenknight or takes me there, anyway
17:02:46 stevenknight unlikely we'll get that far today, of course, but for next week
17:03:07 GregNoel Yeah, the spreadsheet isn't ready yet.
17:03:02 garyo-home Unfortunately I only have 45 min tonight :-(
17:03:14 GregNoel OK, I'm ready
17:03:22 stevenknight okay, 2061
17:03:22 GregNoel Why only 45 min?
17:03:35 garyo-home kid duty
17:03:49 * stevenknight nods knowingly...
17:04:14 * GregNoel gets to miss all that fun...
17:04:21 stevenknight 2061: 1.x p3?
17:04:24 garyo-home Before we talk about the actual issues, can I ask about 1.x vs. 2.x in general?
17:04:29 GregNoel sure
17:04:29 stevenknight sure
17:04:33 stevenknight jinx
17:04:33 GregNoel jinx
17:04:42 garyo-home How much should we put into 1.x vs. deferring to 2.x?
17:04:59 GregNoel Depends on how long you expect 1.x to last
17:05:02 garyo-home 2061 is easy, for instance. But so are *lots* of others.
17:05:07 stevenknight i think we'll need to rebalance 1.x once we get past 1.0 anyway
17:05:16 GregNoel stevenknight, true
17:05:16 garyo-home Greg: right. Is there a 2.0 schedule?
17:05:44 garyo-home Are we thinking 2.0 this year for instance?
17:05:46 GregNoel None, other than some people say one month and some say one year. Probably between the two.
17:06:00 stevenknight yeah
17:06:00 garyo-home Greg: OK, that's sort of my thought too.
17:06:13 stevenknight my best guess is some time in Q4
17:06:19 stevenknight enough time for 1.x to soak
17:06:22 garyo-home OK, then I say 2061 should be 1.x.
17:06:30 stevenknight i'm okay with gut feel
17:06:41 stevenknight if we'd "like" something in 1.x for any reason, mark it as such
17:06:55 stevenknight we'll end up with too much 1.x, but then we just re-prioritize those to make it manageable
17:06:53 GregNoel I assume that once 1.0 is out and we've killed a little more of the backlog, we'll look at 1.x (and maybe 2.x p1) and adjust.
17:07:05 * stevenknight agrees w/GregNoel
17:07:15 stevenknight so 2061: 1.x p3
17:07:43 garyo-home stevenknight: ok
17:07:48 GregNoel Isn't 2061 the one that's just been on the mailing list?
17:07:54 stevenknight 2062: moot, I went ahead and checked in my fix about half an hour ago... :-)
17:08:06 garyo-home 2062: good
17:08:19 stevenknight also already RESOLVED the issue
17:08:26 garyo-home 2064 is also easy, so by same logic should be 1.x
17:08:30 stevenknight 2064: consensus 1.x p3
17:08:31 stevenknight right
17:08:46 GregNoel ok
17:08:51 garyo-home 2064 ok
17:08:53 stevenknight 2065: consensus 1.x p4, Rob Managan
17:09:40 garyo-home 2066: VS8 is the current version, we should support it well if we can
17:09:45 garyo-home (though I don't use it yet)
17:09:55 stevenknight 2066: i'm conflicted
17:10:09 stevenknight the fix looks like a no-brainer, but...
17:10:11 garyo-home Then let's do it in 1.x early on
17:10:24 garyo-home like 1.x p1?
17:10:35 stevenknight yes, 1.x p1
17:10:39 GregNoel done
17:10:47 GregNoel next is consensus
17:10:53 garyo-home yes.
17:10:55 stevenknight 2067: consensus dup
17:11:01 garyo-home I'll take 2068, good idea.
17:11:11 stevenknight 2068: cool, thanks
17:11:16 garyo-home I'll take up details on the list.
17:11:27 stevenknight i'd like p2 (since I need it too :-))
17:11:45 garyo-home OK, fine w/ me.
17:11:55 stevenknight excellent, we're cruising
17:12:02 GregNoel done with the "current" spreadsheet, then; move on to the next?
17:12:03 stevenknight on to 2007 q4?
17:12:47 stevenknight 1740: consensus research, David
17:12:51 GregNoel done
17:13:15 stevenknight 1741: 1.x p3, stevenknight
17:13:27 garyo-home ok w/ me
17:13:28 GregNoel ok
17:14:05 garyo-home 1742 is a subprocess issue or something?
17:14:36 GregNoel I saw it as an issue with assuming that setting CC forced the C compiler selection
17:14:48 stevenknight 1742: i'm concerned it's a real problem that happens to be triggered by his weird stripped down CC = ''
17:14:59 stevenknight hmm, let me look at it again w/that in mind -- hang on...
17:15:26 garyo-home That code doesn't look right to me; the high 8 bits are supposed to be spawn status, the low 8 bits are return code
17:15:54 garyo-home (or other way around, sorry)
17:16:03 stevenknight but if it that code were that blatantly wrong, a lot of stuff would fail, not just this edge case
17:16:20 garyo-home stevenknight: yeah, I take it back -- it's OK as written
17:16:30 stevenknight i think the real problem here is that this compilation setting definitely shouldn't succeed
17:16:42 stevenknight but we pass back a return value that suggests the test passed
17:16:56 GregNoel No, setting CC is ignored, so TryXXX will succeed.
17:17:14 GregNoel He's expecting that setting CC will _disable_ the C compiler
17:17:30 garyo-home anyway, research is needed.
17:17:58 GregNoel I'll take it.
17:18:01 garyo-home But has to be fixed by 1.x one way or another, so that's my vote.
17:18:12 stevenknight okay, i can go with 1.x
17:18:16 GregNoel done
17:18:39 garyo-home 1745, VS junk
17:18:54 stevenknight 1745: basically, i'm going to take everything VS-related as research
17:19:04 garyo-home OK w/ me, this is super low pri.
17:19:24 GregNoel ok, stevenknight, research
17:19:39 stevenknight yeah, i'm just going to revamp VS support pretty heavily
17:19:44 garyo-home 1746: untangle threaded output
17:19:55 garyo-home This is really hard, and error-prone.
17:20:15 GregNoel It's a dup; mark it and triage that one.
17:20:21 stevenknight i'm okay with dup
17:20:21 garyo-home 2.x p3?
17:20:28 stevenknight 2.x p3
17:20:29 garyo-home (ok, dup)
17:20:37 * bdbaddog (n=bdeegan@adsl-71-131-1-136.dsl.sntc01.pacbell.net) has joined #scons
17:20:38 stevenknight on 1183 is fine
17:20:44 stevenknight hey bill
17:20:48 GregNoel hi, bill
17:20:50 stevenknight we're on th 2007 q4 spreadsheet
17:20:52 bdbaddog Hi.
17:20:57 stevenknight #1746, line 51
17:21:00 garyo-home Hi, Bill.
17:21:02 bdbaddog oh yeah. forgot there's a bug party.
17:21:22 bdbaddog I've gotta hit the road in like 10 minutes. So I'll be of no help today. sorry.
17:21:22 stevenknight no problem, you're obviously welcome if you have cycles
17:21:34 stevenknight that's cool
17:21:34 GregNoel 2.x p3 on 1183; agreed.
17:21:40 stevenknight done
17:22:08 stevenknight 1747: documentation, 1.0 p5 (like a lot of other doc issues)
17:22:17 GregNoel ok, done
17:22:24 garyo-home I seem to have a few doc things; assign it to me.
17:22:47 stevenknight 1748: 1.x p2 -- our code looks like it handles this right
17:22:56 stevenknight i really suspect this is in custom code for this project
17:23:15 garyo-home Push back, ask for testcase?
17:23:24 stevenknight hmm, not a bad idea
17:23:41 stevenknight oh, wait, unfortunately i think this is one where the *user* of a project submitted something to us
17:23:54 stevenknight that does suggest pushing it back, having him contact the original project
17:24:17 stevenknight i'll go ahead and answer the bug to that effect
17:24:26 GregNoel how about me, research, and I'll untangle it.
17:24:32 GregNoel I can try a test case.
17:24:55 stevenknight if you want, sure, go ahead
17:25:21 GregNoel done; next?
17:25:22 stevenknight 1.x, p2, greg -- done
17:25:40 stevenknight 1751: 1.x p3, me
17:25:49 stevenknight i think it's related to the other above, and #2015
17:25:50 GregNoel done
17:25:50 garyo-home agreed
17:26:12 stevenknight 1753: visual studio: research, stevenknight
17:26:15 garyo-home 1753: dup?
17:26:27 stevenknight maybe, just assign it to me and i'll take care of it if so
17:26:34 GregNoel done
17:26:55 stevenknight 1754: i think i put my comment on the wrong item, i think i intended that for 1753
17:27:26 garyo-home 1754 looks right to me.
17:27:29 garyo-home not a bug.
17:28:03 GregNoel not a bug, a feature request.
17:28:18 stevenknight right, maybe for a --clobber that will remove .sconsign*
17:28:21 stevenknight or some such
17:28:23 garyo-home (I put my .sconsign and .sconf_temp stuff in my build dir, that way if I wipe that out I start from scratch.)
17:28:29 stevenknight good idea
17:28:39 garyo-home stevenknight: ok w/ that I guess
17:28:52 GregNoel scons -ccc
17:28:58 garyo-home :-/
17:29:06 stevenknight sure, i could go with that
17:29:15 stevenknight either way, FEATURE...
17:29:16 stevenknight 2.x?
17:29:20 garyo-home 2.x
17:29:24 stevenknight p3
17:29:25 GregNoel We discussed this once before
17:29:34 GregNoel ok, 2.x p3
17:29:44 stevenknight probably, they all start to blur after a while... :-)
17:30:05 stevenknight 1755: 1.x p4, Greg
17:30:14 GregNoel 1755, consensus
17:30:30 garyo-home ok
17:30:34 stevenknight 1760: research, Rob
17:30:38 GregNoel yes
17:30:56 garyo-home ok
17:30:56 stevenknight 1761: gary, you okay with 1.x p3?
17:31:01 stevenknight and still on your plate
17:31:15 garyo-home ok, I'll try to do it.
17:31:19 garyo-home It would be cool.
17:31:28 stevenknight yes
17:31:27 GregNoel done
17:32:02 GregNoel 484 (actually 1762)
17:32:09 garyo-home 1762: general problem with ancient OSes (IRIX tar is even worse)
17:32:18 stevenknight 1762: 1.x p4, Greg, dup to 484 as you see fit
17:32:38 GregNoel Tar will be replaced by tarfile as soon as 1.5.2 is obsolete
17:32:43 GregNoel I already have it working
17:32:47 garyo-home yay
17:33:08 stevenknight i thought tarfile didn't show up until like Python 2.4
17:33:22 GregNoel I have backported it to 2.2
17:33:38 stevenknight GregNoel++
17:33:59 stevenknight and i was wrong anyway, it's 2.3
17:34:01 stevenknight cool
17:34:06 GregNoel but I couldn't backport it to 1.5.2; too many @staticfoo annotations
17:34:14 stevenknight makes sense
17:34:27 stevenknight so this definitely 2.x, but high priority
17:34:35 GregNoel agreed
17:34:26 garyo-home 1763: I think is user error.
17:34:35 garyo-home He wants this to work:
17:34:42 garyo-home cplusplus = __import__('g++', globals(), locals(), [])
17:34:51 garyo-home (sorry I'm getting ahead)
17:34:56 stevenknight 162: 2.x p2, or even p1, your call
17:34:59 stevenknight 1762 that is
17:35:21 stevenknight 1763: oh, did i misread it?
17:35:31 stevenknight I thought he was complaining about the Tool() call within the .generate() function
17:35:44 garyo-home Seems like he thinks tools should appear in sys.path.
17:35:48 GregNoel 1763, agreed
17:36:06 garyo-home Give 1763 to me and I'll reply to it, see if I can clear up the confusion.
17:36:15 stevenknight 1763: okay
17:36:20 garyo-home I have new doc for site_scons which should help anyway.
17:36:38 GregNoel done
17:36:57 stevenknight 1764: solaris
17:37:12 stevenknight and shall we see if maxim can become the solaris guy?
17:37:25 stevenknight meant to say: 1764: research
17:37:29 GregNoel works for me; will you talk to him?
17:37:36 stevenknight ok
17:38:04 GregNoel research, stevenknight, hand off to maxim
17:38:27 stevenknight done
17:38:51 stevenknight 1766: 2.x p3?
17:39:03 garyo-home what about 1765
17:39:10 garyo-home future/p1?
17:39:25 stevenknight sorry, 1765:
17:39:47 stevenknight future p1 stevenknight
17:39:48 GregNoel 1765, future, p1
17:40:14 * bdbaddog has quit ("Leaving.")
17:40:30 stevenknight done
17:40:41 GregNoel 1766
17:40:45 stevenknight 1766: 2.x p3?
17:40:50 garyo-home 1766: not really a bug, is it?
17:40:59 stevenknight it's weird
17:41:16 garyo-home stevenknight: sure is, but is it a bug?
17:41:16 stevenknight he gives it one .class file as a target and --debug=tree shows him the dependencies of another
17:41:37 garyo-home ok, I can see it being confusing.
17:41:48 GregNoel They're circularly dependent, so it's doing the right thing there
17:42:07 stevenknight i think it's because it's showing the "primary" dependency of the executor that creates all of the [ABC].class files
17:42:08 GregNoel but it should also report on B.class and C.class with the same tree
17:42:20 garyo-home GregNoel: ideally yes
17:42:23 stevenknight right
17:42:43 garyo-home anyway, I can't see putting it in 1.x
17:42:55 stevenknight 2.x p3, me?
17:42:56 GregNoel Make it dependent on batch builders and review when that is fixed.
17:43:09 garyo-home ok w/ me
17:43:49 garyo-home 1769: 1.x p2?
17:43:55 GregNoel do we want to triage 1086 now (batch builders) as 2.x p3?
17:44:31 garyo-home GregNoel: batch has to be in 2.x IMHO
17:44:31 stevenknight 1086: i see that as 1.x
17:44:56 GregNoel a small conflict ...
17:45:06 garyo-home stevenknight: is it possible to get into 1.x? If so, go for it!
17:45:14 stevenknight i think so
17:45:23 garyo-home Huge performance win
17:45:23 stevenknight people have been waiting a long time for it
17:45:29 stevenknight and it would be a huge performance win
17:45:33 garyo-home :-)
17:45:39 stevenknight give it to me for 1.x
17:45:48 stevenknight and i'll definitely push it out (again) if it's too hairy
17:45:46 GregNoel what priority?
17:45:49 stevenknight p2
17:45:53 GregNoel done
17:46:21 * GregNoel thinks stevenknight is crazy...
17:46:43 * stevenknight thinks so, too
17:46:13 stevenknight 1769: greg and i said future, gary you suggested 1.x
17:46:45 garyo-home Greg wants to do it right, I want to hack it so it works. :-)
17:47:03 garyo-home ... but then let Greg do it right later.
17:47:10 GregNoel hmmm...
17:47:27 garyo-home but whatever you guys think on this one.
17:47:36 stevenknight i'm agnostic, so i'm content letting you two fight it out... :-)
17:47:49 garyo-home in that case Greg it's up to you.
17:47:49 GregNoel Maybe Gary and I should talk about this off-line
17:47:58 garyo-home So future it is.
17:48:02 GregNoel done
17:48:09 stevenknight okay
17:48:35 stevenknight 1772: this sounds pretty serious, but beyond 1.0
17:48:39 stevenknight so 1.x p2 (if not p1)
17:48:52 garyo-home agreed
17:48:56 GregNoel 1.x p2
17:49:00 stevenknight done
17:49:11 stevenknight 1831: realized we can probably close this out with reference to Progress()
17:49:21 garyo-home yes, I do it now that way.
17:49:22 stevenknight which provides a hook for the user to print out the target
17:49:40 garyo-home I've even posted my progress func on the list iirc.
17:49:49 stevenknight I'll go ahead and close this out real time while we continue
17:49:57 GregNoel done
17:50:30 stevenknight 1832: moot, David Cournapeau already dup'ed it to 2004
17:50:49 garyo-home good.
17:51:17 GregNoel 1833
17:52:07 stevenknight 1833: assign to me (i have some other --debug=explain work already on my plate)
17:52:12 stevenknight 1.x
17:52:22 GregNoel ok, what priority?
17:52:22 stevenknight p4 because it's back-burner for David?
17:52:38 GregNoel done
17:53:18 garyo-home 1838 seems familiar?
17:53:31 stevenknight 1838: think i fixed that when I did Value nodes recently
17:53:41 stevenknight i'm inclined to close it on that basis
17:53:42 garyo-home ah yes, now I remember.
17:53:57 stevenknight but that's without hard evidence that it's the same problem
17:54:12 garyo-home Sohail can reopen if it doesn't work on next release.
17:54:24 GregNoel ok
17:55:08 GregNoel 1842
17:55:21 garyo-home is that Fortran problem or something else?
17:55:23 stevenknight okay, I'll close it -- i didn't notice it's Sohail, that makes it all right to close it unilaterally... :-)
17:55:45 stevenknight 1842 sounds really weird to me
17:56:09 garyo-home Must be Fortran; I say David should look at it.
17:56:22 garyo-home 1.x p3 for him
17:56:30 GregNoel done
17:56:34 stevenknight done
17:56:59 stevenknight 1844: 1.x p2, we should do right by 64-bit systems
17:57:08 stevenknight i'll be glad to take it
17:57:12 garyo-home ok w/ me
17:57:35 GregNoel ok, make 20xx a dup?
17:58:04 garyo-home ok
17:58:37 stevenknight done
17:58:41 GregNoel 1862
17:58:42 garyo-home 1862: 1.x p3, consensus?
17:58:46 stevenknight yes
17:58:49 GregNoel done
17:59:06 garyo-home 1869: 2.x p3?
17:59:07 stevenknight 1869: i said 1.x but could easily go 2.x
17:59:10 stevenknight done
17:59:11 stevenknight 2.x p3
17:59:27 GregNoel ok
17:59:38 stevenknight 1771: same, i put down 1.x but don't feel strongly about it
18:00:15 garyo-home I'm not a Java guy so I'll stay out of 1771, and now I'm about to turn into a pumpkin. I'll leave my window open so I can review the rest of the goodies :-) See you guys later...
17:59:40 Azverkan brandon here, fyi re 1844 the entire windows registry is screwy in 64 bit python, not just the visual studio stuff
18:00:26 Azverkan it should probably fixed in the upstream registry package somehow
18:00:39 garyo-home Azverkan: that is a good idea.
18:00:43 stevenknight hi brandon
18:01:00 Azverkan at work so I'm just watching
18:01:01 stevenknight agree re: some more comprehensive fix in how we deal with the registry
18:01:27 stevenknight wrap up all of these in a function that will look in both 32-bit and 64-bit locations
18:01:36 stevenknight without having to sprinkle that logic all over the rest of the modules
18:01:36 GregNoel (1771 isn't a registry problem, do you mean 1869?)
18:01:50 Azverkan 1844
18:03:31 GregNoel ah, way back there...
18:04:36 stevenknight okay, back to 1771:
18:04:51 stevenknight 2.x p2?
18:05:40 GregNoel I'm torn
18:06:32 GregNoel It does seem specialized, so 2.x p2 is reasonable.
18:06:50 stevenknight okay, let's go with that
18:06:54 GregNoel ok
18:07:07 GregNoel Next spreadsheet?
18:07:07 stevenknight on to 2007q3?
18:08:01 stevenknight 1687: INVALID or else a doc issue w.r.t. SideEffect() files not getting cleaned
18:08:16 stevenknight i didn't look to see what (if anything) we say about that...
18:08:49 GregNoel However, the TeX builders now are using SideEffect to specify optional files; that was in a REVIEW not too long aga
18:08:53 GregNoel ago
18:09:08 stevenknight um, grep SideEffect Tools/*tex*.py turns up nothing
18:09:16 stevenknight the TeX tools are using emitters, not SideEffect
18:09:29 GregNoel Well, it's what Rob said he was doing...
18:09:53 stevenknight oh, maybe that's in a pending patch -- let me do another quick search
18:10:32 stevenknight hmm, still not finding anything like that
18:10:35 * GregNoel is doing a search of his own...
18:10:56 stevenknight was he saying SideEffect as in the function, or "side effect" as in colloquial expression for "additional files created by TeX"
18:12:50 GregNoel Hmmm... Not sure. I read it as "SideEffect" but he could have just been imprecise.
18:13:46 GregNoel There's no internal API for side effects; the only entry is SideEffect(), so your search is sufficient
18:14:20 stevenknight well, even if TeX starts using it (i could see that for things like logs) i think the right way to handle it would be to also specify Clean() on the SideEffect() files
18:14:21 GregNoel I guess that makes it a doc issue.
18:14:32 stevenknight okay, 1.0 p3 doc
18:14:53 GregNoel ok, I'll write it up that way.
18:15:08 stevenknight done
18:15:21 stevenknight 1689: 1.x p2, who?
18:16:31 GregNoel not me. I'm curious about it, but I think I'm too UNIX-centric
18:16:58 stevenknight hmm, i think Gary might be off with the kids, and we're only on our second issue in this spreadsheet
18:17:08 stevenknight shall we call it a night?
18:17:16 GregNoel I'm willing
18:17:38 stevenknight okay, sounds good
18:17:52 GregNoel With three sets of relatives in town these past few days, I didn't get much farther than we are now
18:18:05 stevenknight not bad, though, we made pretty good progress
18:18:19 stevenknight any conflict for you w/next Monday same time (17h00)?
18:18:39 GregNoel No, I don't think so; let me check
18:19:22 GregNoel Monday the 2nd is good for me
18:19:43 stevenknight okay, that'll be the stake in the ground
18:19:58 GregNoel OK, I'll publish it
18:20:12 stevenknight do you have cycles to update the bugs or shall I handle that translation?
18:20:26 GregNoel I'll get some spreadsheets for the next couple of times as well
18:20:38 GregNoel No, I can handle it as long as my network is alive
18:20:45 stevenknight still flaky?
18:21:07 GregNoel Much better, but bandwidth is down
18:21:34 GregNoel There were probably some burned wires that haven't been replaced yet
18:21:38 stevenknight yow
18:22:01 stevenknight all right, i'm off to get back to other things
18:22:11 stevenknight many thanks...
18:22:11 GregNoel ok, cul
18:22:27 * GregNoel has been marked as being away
18:22:34 * stevenknight has quit ("Leaving")
18:54:11 * garyo-home has quit ("ChatZilla 0.9.82.1 [Firefox 2.0.0.14/2008040413]")
21:28:01 * Azverkan has quit ("[BX] Time to make the donuts")

BugParty/IrcLog2008-05-27 (last edited 2008-05-28 10:18:04 by GregNoel)