00:00:42  * mloskotjoined
00:14:50  * ender`quit (Quit: The first 90% of the code accounts for the first 90% of the development time. The remaining 10% of the code accounts for the other 90% of the development time.)
00:17:59  * drdanzquit (Read error: Connection reset by peer)
01:36:30  * mloskotquit (Remote host closed the connection)
02:15:38  * auroraeosrosequit (Remote host closed the connection)
02:29:36  * jgmdevquit (Remote host closed the connection)
02:41:51  * madewokherdjoined
05:12:16  * madewokherdquit (Remote host closed the connection)
06:50:59  * ender`joined
07:24:27  * drdanzjoined
07:32:37  * ender|quit (Ping timeout: 264 seconds)
07:44:27  * ender|joined
10:01:56  * ender|quit (Ping timeout: 246 seconds)
10:04:22  * ender|joined
10:09:17  * ender|quit (Ping timeout: 246 seconds)
10:14:14  * ender|joined
11:16:08  * ender|quit (Ping timeout: 246 seconds)
11:16:34  * ender|joined
11:25:34  * ender|quit (Ping timeout: 246 seconds)
12:29:41  * piscisaureus_joined
12:58:00  * ender|joined
13:02:55  * ender|quit (Ping timeout: 246 seconds)
13:15:06  * auroraeosrosejoined
13:18:01  * ender|joined
15:00:13  * virmitiojoined
15:00:21  * jgmdevjoined
15:03:40  * madewokherdjoined
15:08:49  * ender1joined
15:10:50  * ender`quit (Read error: Operation timed out)
15:46:46  <madewokherd>if a project's source control has updates that are 4 years newer than the latest release, I should probably go with the source control, right?
15:47:15  <FearTheCowboy>yeaaah.
15:48:06  <madewokherd>then should I also add it to nuget?
15:54:04  <virmitio>um... huh? that question doesn't parse
15:56:07  <madewokherd>if I'm downloading a thing from source control and building and using it myself, and not publishing the results of that work for other people to consume more easily, that seems anti-social
16:00:00  <virmitio>well, is the copy you grabbed from source control able to be referenced by an explicit and unique version number? and is the source you're building the confirmed final state referenced by that version number? if so, then go ahead and push a nuget package of it by that version reference. if not, I can't see it being a stable representation of the library, and likely not a good candidate for a nuget package
16:00:26  <madewokherd>I have no reason to believe it is
16:00:45  <madewokherd>but this is what strong-naming is for
16:01:07  <virmitio>if it's versionable but non-final for that version, you can still push a nuget package with a version semantic identifying it as a pre-release/development package
16:01:26  <virmitio>yeah, nuget doesn't use strong naming really
16:01:36  <madewokherd>oh
16:01:49  <virmitio>currently it's using semver version 1
16:02:13  <auroraeosrose>strong naming is a LIE
16:02:45  <madewokherd>well how am I supposed to publish anything without stepping on someone's toes if I don't have strong-naming?
16:02:58  <madewokherd>except things I author myself I guess
16:04:02  <FearTheCowboy>lets bring this up in the meeting.
16:04:27  <madewokherd>unless coapp is going to fix it, let's not
16:14:30  * drdanzquit (Remote host closed the connection)
16:37:53  <virmitio>Weekly Conf Call 10:30 PDT (55 mins from now) Connect w/Lync Attendee http://j.mp/KP0Zht to http://j.mp/MsWUKA Join Anytime
16:43:07  * piscisaureus_quit (Ping timeout: 258 seconds)
16:45:38  <ender1>i won't be joining tonight - i've got a bad cold and keep sneezing
17:18:15  * piscisaureus_joined
17:25:44  <FearTheCowboy>auroraeosrose ... gonna join? We'll try to keep it brief this week
17:25:50  <auroraeosrose>coming
17:26:02  * virmitiodoesn't believe that claim...
17:26:04  <auroraeosrose>my MIL dropped by
17:45:01  * jgmdevquit (Remote host closed the connection)
18:15:43  * hanrahatjoined
18:15:52  * piscisaureus_quit (Ping timeout: 258 seconds)
18:28:37  <madewokherd>I guess if that doesn't get solved it means we can't package any libraries statically that have dependencies that aren't self-contained (i.e. if there's a library that expects its users to pass it a zlib stream, it can't be statically linked if we can't guarantee there won't be separate zlibs in the binary)
18:38:54  <FearTheCowboy>Hmmm. It does appear possible to link a .lib file so that it doesn't contain the dependencies, and that makes it the responsibility of the thing that links *that* in to make sure that the dependencies are satisfied at link time.
18:38:59  <FearTheCowboy>Interesting.
18:39:36  <FearTheCowboy>that means that it's actually possible to ship an openssl.lib file that still has link-time dependencies that aren't in the lib file.
18:40:35  <auroraeosrose>wow that's
18:40:36  <auroraeosrose>nuts
18:40:39  <auroraeosrose>but kind of cool
18:41:00  <FearTheCowboy>that's really good. I think that we can use that.
18:44:27  <madewokherd>that's how I originally thought it worked
18:45:22  <FearTheCowboy>Yeah, but when you build a lib file, the LIB tool doesn't have a way of saying "Get XYZ() from OTHER.LIB
18:45:37  <FearTheCowboy>but I just figured out how to arrange that.
18:45:57  <madewokherd>well, sure, but if you don't specify that you get maximum flexibility
18:46:52  <FearTheCowboy>at the expense of the possibility that the consumer might not know how to solve the issue correctly...
18:46:58  * FearTheCowboyis experimentin'
18:55:58  * piscisaureus_joined
19:19:45  * piscisaureus_quit (Ping timeout: 252 seconds)
19:26:47  * drdanzjoined
19:36:20  * mloskotjoined
19:39:13  * hanrahatquit (Ping timeout: 245 seconds)
21:05:00  * [[zzz]]joined
21:08:36  * [[zz]]quit (Ping timeout: 264 seconds)
21:36:17  * piscisaureus_joined
21:36:42  * drdanzquit (Remote host closed the connection)
21:37:32  * piscisaureus_quit (Client Quit)
21:37:58  <FearTheCowboy>*whew*... Good news--we've pretty much worked out how to allow intermediate packages to delegate linking decision to the consumer.
21:38:30  <FearTheCowboy>Should allow you to make a libpng package that *needs* zlib, but lets the end dev choose dynamic or static.
21:39:13  <FearTheCowboy>the CallingConvention (on X86) can be forced for a dependent package by the intermediate tho'
21:39:48  <FearTheCowboy>I'd highly recommend that package devleopers stick with cdecl, as that will leaves everyone with the most flexibility
21:46:36  * virmitiopart
21:47:30  <auroraeosrose>most people aren't silly enough to use stdcall
21:47:34  <auroraeosrose>very very small set that do that
21:47:40  * ender1quit (Quit: I believe in free will, but I don't really have a choice.)
21:47:56  <FearTheCowboy>Yeah, I agree. And that'll help a lot.
21:58:02  * madewokherdjust tried calling MessageBoxA on windows rt as cdecl and stdcall to see if one of them crashes
21:59:51  <madewokherd>only use case I can think of for stdcall is vb6, and somehow I'm hoping vb6 and coapp never have any reason to interact
22:02:08  * mloskotquit (Quit: Leaving)