00:08:18  * FearTheCowboyquit
00:11:51  * mloskotquit (Quit: mloskot)
02:02:08  * saivertquit (Read error: Operation timed out)
02:07:45  * saivertjoined
02:08:44  * gixquit (Ping timeout: 272 seconds)
02:12:51  * gixjoined
02:23:26  * saivertquit (Ping timeout: 276 seconds)
02:23:38  * saivertjoined
04:43:18  * madewokherdquit (Remote host closed the connection)
07:48:03  * ender`joined
08:08:36  * piscisaureus_joined
08:18:04  * HamishCjoined
09:46:53  * stalledjoined
10:19:06  * ssam2joined
10:23:29  * HamishCquit (Read error: Connection reset by peer)
10:23:45  * HamishCjoined
11:25:32  * HamishCquit (Read error: Connection reset by peer)
11:41:18  * HamishCjoined
12:28:30  * Kitten_Basketquit (Read error: Connection reset by peer)
12:34:48  * Kitten_Basketjoined
13:45:59  * FearTheCowboyjoined
13:45:59  * FearTheCowboyquit (Changing host)
13:46:00  * FearTheCowboyjoined
13:47:30  * madewokherdjoined
13:56:06  * piscisaureus_quit (Ping timeout: 240 seconds)
14:00:23  * piscisaureus_joined
14:03:57  * piscisaureus__joined
14:04:00  * piscisaureus_quit (Read error: No route to host)
14:04:11  * piscisaureus__changed nick to piscisaureus_
14:08:42  <madewokherd>oh, I see why I didn't build pango/cairo
14:08:50  <madewokherd>it's because they require pixman and pixman needs perl to build
14:09:25  <auroraeosrose>it does?
14:09:41  <madewokherd>according to my notes, it does
14:11:13  <virmitio>a functioning perl package should be available shortly (though it's going to be kindof a beast and the current package will only be temporary)
14:16:49  * piscisaureus_quit (Ping timeout: 248 seconds)
14:19:38  * piscisaureus_joined
14:20:55  <madewokherd>is there a policy on when to update the package version?
14:21:21  <madewokherd>every push?
14:22:40  <auroraeosrose>madewokherd: you don't have to use perl
14:22:56  <auroraeosrose>it generates some C files (like php does with bison/re2c) but they commit the generated files as well
14:23:07  <auroraeosrose>so if you want to make changes and regenerate you need perl
14:23:12  <virmitio>present policy: major, minor, and release version numbers should match the upstream version forked. build version should increment at least once on each push where it is desired for the automation to build msi packages
14:23:13  <madewokherd>I feel like we had this conversation before
14:23:20  <auroraeosrose>we did ;)
14:23:42  <madewokherd>I think I didn't want to mess with the build process to get rid of the dependency
14:24:43  <virmitio>madewokherd: from another angle, you can push as often as you'd like to the repo, but the automation will only try building it if the latest push contains a change to the version.inc file (or if that file is missing entirely)
14:25:12  <madewokherd>mind if I work on adding the flavor into the libiconv assemblies?
14:25:18  <madewokherd>assembly names
14:26:10  <virmitio>already have patches in place and staged to push on my machine
14:26:15  <madewokherd>ok
14:26:56  <virmitio>I'm standing by for the new devtools that will not die horribly with some of the changes I made
14:27:17  * piscisaureus_quit (Ping timeout: 245 seconds)
14:27:38  <madewokherd>how about a gettext package? anyone working on that?
14:27:59  <virmitio>not me (yet at any rate)
14:28:09  <auroraeosrose>I think everyone is afraid of it
14:28:12  <auroraeosrose>but it is needed
14:28:21  <virmitio>you're welcome to tackle it if you want
14:28:50  <madewokherd>well the hard part will be ensuring gettext can find the applications' .mo files
14:29:27  <auroraeosrose>that's kind of application specific though
14:29:58  <madewokherd>maybe
14:30:14  <auroraeosrose>bindtextdomain handles that
14:30:18  <auroraeosrose>or should ;)
14:31:05  * piscisaureus_joined
14:32:00  <auroraeosrose>is there some kind of a default location libintl wants to use?
14:32:15  <auroraeosrose>last time I looked at that code I was shaking my head sooo
14:33:42  <madewokherd>oh
14:33:57  <madewokherd>I need a mingw build of gettext to get the header
14:34:17  <virmitio>?
14:34:28  <madewokherd>never mind
14:34:54  <madewokherd>at some point I'll have to make a mingw package of libiconv, but I don't think I need to just yet
14:34:59  <auroraeosrose>you mean config.h?
14:35:05  <madewokherd>no, I mean libintl.h
14:35:29  <auroraeosrose>ah - yeah gettext has that circular dep on it doesn't it
14:35:34  <auroraeosrose>always hated that
14:35:48  <madewokherd>does it?
14:36:24  <auroraeosrose>yeah, you had to build it twice
14:36:40  * virmitioblinks
14:36:43  <auroraeosrose>my brain is foggy a bit but you built it once without iconv or something and then again with
14:36:43  <madewokherd>I haven't yet, but then I haven't been building the whole thing
14:36:46  <madewokherd>just gettext-runtime
14:36:48  <virmitio>that'll play hell with the automation
14:36:58  <auroraeosrose>virmitio: yeah it's a bitch
14:37:21  <madewokherd>I'd better go back and check how I set this up
14:38:10  <auroraeosrose>I think it's not libintl (runtime) but gettext .exe itself that needs that
14:38:38  <madewokherd>oh
14:38:49  <madewokherd>I think we just never bothered to build a libiconv that uses gettext
14:38:56  * piscisaureus_quit (Read error: Connection reset by peer)
14:39:14  * piscisaureus_joined
14:40:04  <auroraeosrose>yeah
14:40:19  <auroraeosrose>because you built one way, then the other, then back again ;)
14:40:25  <auroraeosrose>stupid circular dep
14:40:33  <auroraeosrose>although they should be different flavors then
14:40:38  <auroraeosrose>virmitio: will that help?
14:40:42  <auroraeosrose>we can dep on flavors right?
14:40:49  <virmitio>yes
14:41:09  <madewokherd>so we'll end up with a libiconv[vc10-bootstrip], which gettext will build against but link to libiconv[vc10] ?
14:41:28  * virmitiostarts digging out the hatchet to go deconstruct the automation again...
14:41:30  <auroraeosrose>well - so you'll do libiconv no gettext
14:41:42  <auroraeosrose>then build gettext with libiconv no gettext
14:41:47  <auroraeosrose>then build libiconv with gettext
14:41:54  <madewokherd>well, how would you do it?
14:41:58  <auroraeosrose>then probably should do a new build of gettext against libiconv with gettext
14:42:00  <madewokherd>wait
14:42:03  <auroraeosrose>blah
14:42:06  <auroraeosrose>my brain is fried
14:42:12  <auroraeosrose>I need to go read the makefiles again
14:42:22  <auroraeosrose>whoever decided to do this circular nightmare should be SHOT
14:42:24  <madewokherd>having gettext or not having it won't change the abi, right?
14:42:29  <auroraeosrose>it shouldn't
14:42:32  <madewokherd>so we can link to one and use the other
14:42:37  <auroraeosrose>so the second rebuild isn't technically necessary
14:42:45  <auroraeosrose>but we'll probably want two packages for it
14:42:49  <auroraeosrose>yes?
14:42:58  <madewokherd>well
14:43:29  <madewokherd>if we depend on libiconv targets with uses=, we won't need to package the libiconv that doesn't have gettext
14:43:46  <madewokherd>then I think maybe the automation doesn't need to change
14:43:49  <auroraeosrose>thing is someone may want to NOT use gettext
14:43:55  <virmitio>stupid question time: can we merge the source repos to build both at once in one build step? I could then package up the separate outputs as the appropriate packages.
14:43:56  <auroraeosrose>so they dep on a libiconv without it
14:44:13  <madewokherd>well, it's technically possible
14:44:18  <madewokherd>if you like subtree merges
14:44:21  <auroraeosrose>ewwww
14:44:28  <auroraeosrose>I'm letting you do that ;)
14:44:36  <madewokherd>I don't know how to do subtree merges
14:45:12  <auroraeosrose>.... I don't want to learn ;)
14:45:31  <madewokherd>me neither
14:46:00  <madewokherd>I was inclined to just use a libiconv without gettext
14:46:16  <virmitio>FearTheCowboy: are you actually here at the moment?
14:47:14  <auroraeosrose>madewokherd: unfortunately the minute we do that SOMEONE will complain
14:47:25  <FearTheCowboy>now I am
14:47:42  <madewokherd>ok
14:48:03  <auroraeosrose>FearTheCowboy: dont' know if you remember - but libiconv and libintl circular dep each other
14:48:04  <virmitio>FearTheCowboy: do we have an environment variable similar to %COAPP% but with backslashes instead or slashes?
14:48:07  <auroraeosrose>I had almost forgotten
14:48:24  <FearTheCowboy>%ALLUSERSPROFILE% ?
14:48:37  <virmitio>auroraeosrose: we're already complaining, and we haven't even done it yet
14:48:37  <madewokherd>I think uses= may actually be the way to go here
14:48:43  <madewokherd>though it'll mean we build a bunch of stuff twice
14:48:53  <virmitio>FearTheCowboy: but doesn't that only apply in a default setup situation?
14:48:56  <auroraeosrose>eh, building it twice is doable
14:49:08  <FearTheCowboy>virmitio -> Yeah.
14:49:12  <auroraeosrose>after all - we have dedicated build box ;)
14:49:18  <FearTheCowboy>I don't have an environment variable at this time
14:49:21  <madewokherd>gettext will just depend on targets that make libiconv packages, and libiconv will depend on targets that make gettext packages
14:49:26  <auroraeosrose>yup
14:49:29  * auroraeosroseshudders
14:49:32  <auroraeosrose>build all then package?
14:49:35  <virmitio>FearTheCowboy: k, I'll see if I can convince it to be happy with it
14:49:38  <auroraeosrose>or maybe put all the packages in gettext?
14:49:42  <auroraeosrose>that would be evil
14:50:04  <madewokherd>well
14:50:20  <madewokherd>we'll eventually have to deal with this sort of thing for language bindings
14:50:27  <auroraeosrose>this is true
14:50:38  <madewokherd>gettext has other things in it that indirectly depend on gettext
14:50:40  <auroraeosrose>virmitio and FearTheCowboy: any suggestions?
14:50:47  <auroraeosrose>gettext is spoiling the parade
14:50:53  <virmitio>we can make lots of packages from a single build process, that's pretty trivial (as long as the libraries themselves aren't mashed together into a single file)
14:50:56  <auroraeosrose>(there's a suprise)
14:51:18  <auroraeosrose>virmitio: can you do more then one build process per thing?
14:51:34  <auroraeosrose>cause really it needs to build in iconv _> build in gettext -> build in iconv -> package
14:51:40  <auroraeosrose>ewwww
14:52:02  <virmitio>auroraeosrose: honestly? haven't tried yet, but I don't see it as being more than a moderate nuisance to set up
14:52:43  <madewokherd>for now I'm going to go forward with packaging gettext depending on non-gettext libiconv; it's a step in the right direction
14:53:22  <auroraeosrose>sounds good
14:53:30  <auroraeosrose>virmitio: sounds swiffy
14:53:59  <auroraeosrose>virmitio: I think you're going to have a "packages I hate" list - PHP, gettext, iconv
14:54:04  <auroraeosrose>I'm sure we'll find more
14:54:24  <madewokherd>perl
14:54:33  <virmitio>after the first set of packages are built, the automation should be able to run through future builds without trouble
14:55:02  <virmitio>perl is only being a problem if I try building as x64 or if I try it under vc8
14:55:17  <auroraeosrose>hmmm, both sound like odd issues
14:55:36  <madewokherd>I'm sure I didn't have problems with x64
14:55:40  <virmitio>that and the section of the build script that rewrites the constructed batch files doesn't like me right now
14:55:57  <madewokherd>other than ending up with build files in the same place
14:56:34  <virmitio>madewokherd: I've taken care of that issue by telling it to go ahead and install the output files
14:57:08  <madewokherd>how'd you get it to not put random things in c:\perl?
14:57:13  <virmitio>and since it doesn't seem to honor an alternate path from whatever root drive I tell it to use, I just mapped a new drive to the correct output folder
14:57:49  <virmitio>it at least sends everything to the right drive
14:57:57  <virmitio>yay for small favors
15:01:07  <virmitio>madewokherd: any idea what compiler you used to make your x64 build?
15:03:23  <madewokherd>vc10
15:11:41  * wwahammyjoined
15:32:30  <madewokherd>hmm
15:32:35  <madewokherd>gettext is version
15:32:41  <madewokherd>we've run out of numbers
15:32:57  <auroraeosrose>LOL
15:40:49  <madewokherd>gettext is gpl3, really?
15:41:16  <madewokherd>oh
15:41:23  <madewokherd>gettext is, libintl isn't
15:41:48  <auroraeosrose>yeah
15:42:01  <auroraeosrose>the gettext-tools will need ot be shipped separately from the lib ;)
15:42:13  <madewokherd>no problem, I'm not even building those yet ;)
16:14:57  <madewokherd>virmitio: is the problem you're having that assembly["library[${flavor}]"] isn't a thing you can do?
16:15:10  <FearTheCowboy>it will be later today
16:15:23  <virmitio>madewokherd: yes, until the next tools come out
16:15:45  <madewokherd>well, now I am also having that problem
16:15:48  <virmitio>yep
16:16:13  <virmitio>is the reason I have a bunch of projects staged for me to push once the upgrade is out
16:16:14  <madewokherd>I'll change it to a hard-coded flavor for now
16:16:30  <virmitio>the real issue is the square brackets
16:17:09  <madewokherd>oh
16:17:14  <virmitio>yep
16:17:18  <madewokherd>dammit
16:17:19  <virmitio>just like that
16:17:28  <virmitio>will be fixed soon
16:18:07  <madewokherd>well that means I can't push even with a hard-coded flavor :(
16:18:12  <virmitio>true
16:18:31  <virmitio>but think happy thoughts
16:19:02  <virmitio>like lighting computers on fire
16:19:31  <madewokherd>why the heck did I name my dll libintl-8-vs10.dll
16:19:34  <madewokherd>instead of vc10
16:19:55  <virmitio>it seemed like the thing to do at the time?
16:21:47  <auroraeosrose>heh - DO NOT DO
16:22:10  <virmitio>little late for that
16:23:03  * madewokherdchanges it to libintl.dll
16:24:04  <auroraeosrose>hurray
16:24:12  <auroraeosrose>cause SxS is supposed to be able to figure it out anyway right ;)
16:24:21  <madewokherd>yeah
16:24:25  <virmitio>that's the theory
16:24:50  <madewokherd>I think I was afraid people would link things wrong if I didn't include that stuff in the name
16:26:02  <FearTheCowboy>uploaded newer coapp.toolkit and coapp.devtools ...
16:26:12  <FearTheCowboy>they are not 100% at this point.
16:26:38  <FearTheCowboy>`coapp list ` sometimes hangs, use `coapp list --verbose`
16:26:39  <FearTheCowboy>and
16:26:52  <FearTheCowboy>and the admin required is not fixed yet
16:27:08  <madewokherd>will I be able to put the flavor in my assembly names?
16:27:56  <virmitio>should be able to, I'll be checking that in a moment
16:29:09  <FearTheCowboy>Weekly #CoApp conf call meeting starts in 1 hour: http://t.co/4YydgYu4 join anytime, everyone welcome!
16:29:46  <madewokherd>I need to bookmark the conference call url
16:39:49  * wwahammyquit (Quit: Few women admit their age. Few men act theirs.)
16:40:42  <virmitio>autopackage stuff.autopkg --dep=zlib[${flav}]-${arc}-<pkt>
16:48:26  * _Andrewjoined
17:11:54  * wwahammyjoined
17:26:05  <virmitio>Weekly #CoApp conf call meeting starts in 5 minutes: http://t.co/4YydgYu4 join anytime, everyone welcome!
17:31:06  * aangjoined
17:37:34  * aangquit (Ping timeout: 245 seconds)
17:39:52  <auroraeosrose>https://gist.github.com/53201bd866b30cea523a - HOLY CRAP
17:40:32  <auroraeosrose>http://www.engadget.com/2012/03/23/35-foot-angry-bird-for-launch/
17:40:34  <auroraeosrose>dagit
17:40:37  <auroraeosrose>wrong link
17:40:44  <auroraeosrose>WHO PUTS ANGRY BIRDS ON SPACE NEEDLE
17:41:32  <mgdm>:D
17:49:10  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
18:08:35  * piscisaureus_joined
18:21:09  * cH40z-Lordquit (Ping timeout: 246 seconds)
18:21:27  * cH40z-Lordjoined
18:21:51  * piscisaureus_quit (Ping timeout: 246 seconds)
18:32:09  * ssam2quit (Quit: Leaving)
18:42:10  * remy_ojoined
19:15:54  <ender`>...and the cats are happily eating
19:18:53  <madewokherd>my error went away
19:18:58  <madewokherd>with new toolkit
19:19:30  <madewokherd>now I've got a worse error
19:19:44  <ender`>but... it gets worse?
19:20:07  <madewokherd>oh wait
19:20:11  <madewokherd>my build succeeded anyway
19:20:14  <virmitio>lol
19:20:36  <ender`>haha
19:20:41  <madewokherd>good enough
19:21:17  <virmitio>new zlib, lua, berkeleydb, libiconv, and pcre5 are all building
19:21:32  <virmitio>should auto-push the packages to the site soon
19:21:50  <madewokherd>I guess I should add the flavor to my libiconv manifest before pushing then
19:21:57  <virmitio>perl is running, but I expect it to be another hour or so before the automation gets done packaging it
19:22:16  <virmitio>probably
19:23:36  <FearTheCowboy>hmm. Yeah, the stale check was only going to be stale if the package feed had been read in the next twelve hours in the future.
19:23:41  <FearTheCowboy>o_O
19:26:38  <ender`>hm, my msdn expires in a month
19:27:24  <ender`>this year's gone by fast
19:29:57  <HamishC>I think pcre5 will be redundant soon, I've just merged in the latest httpd trunk which appears to add support for the latest pcre
19:30:34  <HamishC>just making projs for pcre8 now cos I dont wants to use cmake...
19:31:14  <virmitio>HamishC: works
19:33:55  <madewokherd>why don't you want to use cmake?
19:34:23  <madewokherd>would it help if I packaged it?
19:35:12  <virmitio>grr, zlib and a few others aren't packaging because I forgot to update the manifest references
19:35:18  <virmitio>going to fix them after lunch
19:35:31  <madewokherd>eh, I'm unblocked, works for me
19:37:54  <madewokherd>seriously, we have a build files upstream that work with msvc, why not use them?
19:38:34  <HamishC>I dont understand it. sure you can package it, I'm not going to push my vs projects or anything, I just need something to link against for now
19:38:59  <HamishC>only takes an hour to make new projects for it anyways not a big deal
19:39:13  <madewokherd>this is all theoretically automated so you can grab pcre and cmake and to a ptk build from pcre
19:39:20  <madewokherd>*do
19:39:37  * mloskotjoined
19:39:37  * mloskotquit (Changing host)
19:39:37  * mloskotjoined
19:41:06  <madewokherd>"Assembly Reference for libiconv[vc10] not found in this package, or any dependencies"
19:41:25  <madewokherd>and I have a dependency on libiconv[vc10]-
19:41:52  <FearTheCowboy>I'll see about takin' alook after I grab some lunch
19:43:59  <madewokherd>oh our developer libraries need flavors too?
19:44:14  <auroraeosrose>possibly
19:44:31  <auroraeosrose>unless the flavor doesn't have different dev stuff I would assume
19:44:55  <auroraeosrose>I would think 90% of binary ones do
19:46:17  <madewokherd>I wouldn't be surprised if .lib files are interchangeable
19:46:51  <auroraeosrose>are you sure they are?
19:46:57  <madewokherd>no
19:46:58  <auroraeosrose>in some cases they won't be
19:47:10  <auroraeosrose>(think a thread safe vs. non-thread safe build)
19:47:41  <madewokherd>but the exports are supposed to be the same
19:48:16  <auroraeosrose>dunno - would have to ask someone who does more evil
19:48:43  <madewokherd>well, I'm happy to just put the flavor in the name
19:49:05  <madewokherd>just didn't know I needed to
19:49:23  <madewokherd>I guess I assumed there'd be magic?
19:49:53  <auroraeosrose>heh, the install magic is there good - the dep and building packages magic not so much
19:50:52  * mloskotlikes magic as long as he can understand what it does :)
20:07:55  * Scotisis no longer away : Gone for 21 hrs 8 mins 49 secs
20:08:11  <Scotis>if you understand it - it is no longer magic ;)
20:25:25  * mloskotquit (Ping timeout: 246 seconds)
20:31:39  * virmitiois magic
20:31:57  <virmitio>although I suspect that FearTheCowboy is even more magic
20:32:40  <madewokherd>I'm interested in seeing whether your manifest references work after you update them
20:33:04  <virmitio>so am I
20:37:14  <virmitio>I can't tell, coapp is still not handling the force-scan switch as expected
20:38:00  <virmitio>wait... looks like the internal manifest worked this time
20:38:12  <madewokherd>I ended up just installing the package I needed to depend on
20:40:43  * _Andrewquit (Quit: bed time)
20:50:55  <madewokherd>should I be worried that the libiconv installer says it's going to install " -" ?
20:52:48  <virmitio>is that the version you built?
20:53:28  <madewokherd>no
20:53:31  <madewokherd>I didn't build libiconv
20:53:41  <madewokherd>at least, not the package
20:53:48  <virmitio>hmm
20:54:25  <virmitio>ah, hadn't updated my copy of the feed
20:54:48  <virmitio>yes, that's the most recent stable build (should work without exploding anything)
21:04:02  <madewokherd>screw it, I'll just push this and hope it works for the rest of the world, as usual
21:04:11  <virmitio>lol
21:04:20  <virmitio>might as well
21:04:34  <virmitio>if it doesn't work I'll fix it up in a bit
21:20:57  <madewokherd>I'd really like to make a package and have everything work as expected
21:21:02  <madewokherd>someday
21:34:38  * piscisaureus_joined
21:46:58  * qwebirc21897joined
21:47:20  * qwebirc21897quit (Client Quit)
22:10:41  * remy_oquit (Ping timeout: 244 seconds)
23:05:41  * piscisaureus_quit (Quit: ~ Trillian Astra - www.trillian.im ~)
23:14:30  * ender`quit (Quit: Consultation, n. Medical term meaning "to share the wealth.")
23:22:06  * mloskotjoined
23:44:23  * Dark-Starquit (Ping timeout: 260 seconds)
23:51:23  * HamishCquit (*.net *.split)
23:51:23  * FearTheCowboyquit (*.net *.split)
23:51:24  * sungamiquit (*.net *.split)
23:51:24  * cH40z-Lordquit (*.net *.split)
23:51:24  * Afalquit (*.net *.split)
23:51:24  * stalledquit (*.net *.split)
23:51:24  * gixquit (*.net *.split)
23:51:24  * Scotisquit (*.net *.split)
23:51:24  * voltagexquit (*.net *.split)
23:51:24  * BuckRogers25quit (*.net *.split)
23:51:24  * ender|quit (*.net *.split)
23:51:25  * auroraeosrosequit (*.net *.split)
23:51:25  * TReKiEquit (*.net *.split)
23:51:25  * mpreislerquit (*.net *.split)
23:51:25  * ChanServquit (*.net *.split)
23:51:25  * madewokherdquit (*.net *.split)
23:51:25  * virmitioquit (*.net *.split)
23:51:25  * mgdmquit (*.net *.split)
23:51:26  * Kitten_Basketquit (*.net *.split)
23:51:26  * saivertquit (*.net *.split)
23:51:27  * bobquit (*.net *.split)
23:51:27  * mloskotquit (*.net *.split)
23:51:27  * wwahammyquit (*.net *.split)
23:51:28  * sejoquit (*.net *.split)