:: #mythtv

Daily chat history

Current users (92):

aloril, Anssi, anykey_, Beirdo, brfransen, brtb, CaCtus491, cattelan_away, cesman, Chutt, clever, coling, Cougar, damaltor, danielk22, Dave123, dekarl, dlblog, Dr{Wh0}, ElmerFudd, eugo, foobum, foxbuntu, ghoti, gregL, GreyFoxx, Guest29306, highzeth, J-e-f-f-A, j-rod|afk, jams, jarle, jcarlos, joe_, joki, jpabq, jpabq-, jpabq_, jstenback, justinh, jwhite, jya, jya_, k-man, kenni, knightr, kormoc, kurre2, kwmonroe, laga, lolcat`, mag0o, Malard, MaverickTech, mike|2, mrand, MythBuild, MythLogBot, mzanetti, noahric, peitolm, pheld, poptix, purserj, rhpot1991, rsiebert, skd5aner, Slasher`, sphery, sraue, stuarta, stuartm, superm1, sutula, tgm4883, TheAsp, ThisNewGuy, ToadP, tomimo, tris, Unhelpful, vallor, wagnerrp, wahrhaft, wseltzer, xavierh, xavierh_, XDS2010_, xris, ybot_, zCougar, _charly_
Monday, March 5th, 2012, 00:02 UTC
[00:02:31] Dave123 (Dave123! has joined #mythtv
[00:03:23] MythBuild: Hey! build 0.24-freebsd-64bit #124 is complete: Success [build successful]
[00:03:23] MythBuild: Build details are at . . . t/builds/124
[00:04:19] wagnerrp: MythBuild: force build 0.24-freebsd-64bit
[00:04:20] MythBuild: build forced [ETA 39m54s]
[00:04:20] MythBuild: I'll give a shout when the build finishes
[00:09:43] MythBuild: Hey! build 0.24-freebsd-64bit #125 is complete: Success [build successful]
[00:09:43] MythBuild: Build details are at . . . t/builds/125
[00:37:04] danielk22: sphery: Should we do the DB creation/update update prior to the beta too?
[00:54:42] noahric (noahric!~noahric@ has quit (Quit: noahric)
[00:57:19] taylorr: danielk22: the point is that pkt->dts is the same as frame->pts – so using the dts seems to be more reliable in the wild.... I don't know of any sample we've received that fails with pkt->dts
[00:58:03] Beirdo: OK, nap is over
[00:59:01] taylorr: danielk22: if you read that section of the tutorial I sent you then it will explain it better
[01:00:17] Beirdo: danielk22: I see no reason to delay the beta. Delaying RC1, perhaps, but we'll see how the buglist goes.
[01:05:40] wagnerrp: Beirdo: looks like we no longer need the 'COMPILER_PATH' set to compile on freebsd
[01:05:51] wagnerrp: i forgot to do so, but its running through anyway
[01:06:24] Beirdo: no? sweet
[01:08:33] Dave123 (Dave123! has quit (Read error: Connection reset by peer)
[01:08:36] Dave-road (Dave-road! has joined #mythtv
[01:08:50] wagnerrp: must be something different with the FBSD9 environment
[01:12:01] Beirdo: so if we wanted to support both, we still need it?
[01:12:09] Beirdo: both 8 and 9 that is?
[01:17:31] wagnerrp: if we wanted to build on both, yes
[01:17:49] wagnerrp: but its building on 9.0 now
[01:18:03] wagnerrp: i could keep a copy of an 8.x buildbot around if you think it would be worthwhile
[01:18:40] Beirdo: hmm, worth thinking about. I don't know how many FreeBSD users there even are out there :)
[01:29:39] jya: so when are we upgrading to use C++11? one key feature looks very interesting to me :)
[01:36:08] Beirdo: that would depend on a lot of things, most importantly, which versions of gcc do and do not support it
[01:36:21] Beirdo: and certainly not before 0.25 release
[01:36:55] Beirdo: jya: you see the ticket you got sent last night about the plugins config being broken?
[01:37:17] Beirdo: #10406
[01:37:24] jya: yes
[01:37:27] jya: I've fixed it already
[01:37:33] jya: it's in my pushing queue
[01:37:43] jya: will push shortly..
[01:37:47] Beirdo: K. :)
[01:38:07] ** wagnerrp is still curious about this unmentioned 'key feature' **
[01:38:21] Beirdo: heh, yeah, me too :)
[01:39:20] jya: wagnerrp: it's in regards to initialisation of static variables. "such a variable is initialized the first time control passes through its declaration; such a variable is considered initialized upon the completion of its initialization. [...] If control enters the declaration concurrently while the variable is being initialized, the concurrent execution shall wait for completion of the initialization."
[01:39:47] jya: pretty much, the use of static variables is now thread-safe without having to think about it
[01:40:34] jya: and only for code 200% bigger ! :)
[01:41:05] wagnerrp: oh no! mythtv is going to be 6MB instead of 3
[01:41:25] Beirdo: I'd be more worried about it being slower than bigger
[01:41:58] jya: wagnerrp: 6? mythfrontend on my mac alone (Without lib) is 8.5MB, in universal mode it's 15MB
[01:42:52] jya: I actually wonder if it's worth my time struggling in building a universal application on the mac, it's makes the application over 350MB
[01:43:08] danielk22: jya: The big problem with static initialization in C++ is initialization order across compilation units is undefined. Is that fixed?
[01:43:31] Beirdo: hehe, boy did we prove that in this devel cycle :)
[01:43:32] jya: danielk22: i haven't read all of the specs yes
[01:43:50] jya: yet
[01:44:10] wagnerrp: jya: with debugging stripped out, the 64-bit 'mythfrontend' executable on gentoo is just under 3MB
[01:44:28] wagnerrp: with debugging left in, that 64-bit executable on freebsd is just under 50MB
[01:44:39] jya: i saw references about the order of members's constructors being called in the order they're declared
[01:45:14] Beirdo: we already have that, do we not?
[01:45:27] jya: for static var?
[01:46:27] jya: danielk22: I found this article interesting
[01:46:27] jya: . . . e.php/c12897
[01:48:48] Dave-road (Dave-road! has quit (Quit: Leaving)
[01:52:43] gregL (gregL! has quit (Read error: Connection reset by peer)
[01:52:45] danielk22: jya: Where I see devs get in trouble is with "ClassA a;" "ClassB b;" in separate cpp files where ClassB::ClassB() { a.print("B is init"; } .... this works fine when a gets initialized first but blows up when B is initialized first.
[01:53:54] danielk22: jya: I have seen the problem in that article, but only a couple times in the last decade.
[01:58:51] jya: well, they've had 8 years to work on the problem at hand.. hopefully it's been considered
[02:06:22] davide_ (davide_!~david@mythtv/developer/gigem) has quit (Remote host closed the connection)
[02:06:44] davide_ (davide_!~david@mythtv/developer/gigem) has joined #mythtv
[02:06:52] danielk22: <-- looks like it is mentioned in the C++ FAQ so they were probably at least aware of the problem.
[02:08:27] gregL (gregL! has joined #mythtv
[02:15:14] jya: danielk22: having said that, I would have done something like this instead: ClassB::ClassB(ClassA& a) { a.blah(); } ; ClassA a; ClassB b(a);
[02:15:51] Dave123 (Dave123! has joined #mythtv
[02:22:06] danielk22: jya: The problem really occurs when a is used implicitly. It's hard to come up with a simple example that is believable because it occurs in larger code bases... the official solution is to use lazy initialization. ClassA *a=NULL; ClassB *b=NULL; ClassB::ClassB() { A::instance()->print("B is init."); }
[02:22:55] jya: that's why I think we should all go back and do assembly instead
[02:23:06] jya: then you'll know exactly what you do
[02:23:15] Beirdo: ANSI C will do fine.
[02:33:29] stichnot (stichnot!chatzilla@mythtv/developer/stichnot) has quit (Ping timeout: 245 seconds)
[02:43:56] stichnot (stichnot!~chatzilla@ has joined #mythtv
[02:43:56] stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has joined #mythtv
[02:43:56] stichnot (stichnot!~chatzilla@ has quit (Changing host)
[03:03:58] k-man (k-man!~jason@unaffiliated/k-man) has quit (Quit: WeeChat 0.3.6)
[03:09:03] k-man (k-man!~jason@unaffiliated/k-man) has joined #mythtv
[03:22:29] ** knightr favorite language used to be assembly but he hasn't programmed anything in it in years... Last time he did it was a TSR (Terminate and Star Resident) program (that's how long ago it was...) **
[03:24:53] knightr: s/Star/Stay
[03:31:27] jams (jams! has joined #mythtv
[03:49:13] Seeker` (Seeker`!~cjo20@unaffiliated/seeker) has quit (Ping timeout: 245 seconds)
[03:51:00] Seeker` (Seeker`! has joined #mythtv
[03:51:00] Seeker` (Seeker`! has quit (Changing host)
[03:51:01] Seeker` (Seeker`!~cjo20@unaffiliated/seeker) has joined #mythtv
[03:55:30] knightr: Missed Future? How can something in the future be missed? Have Doc Brown and Marty been traveling again?
[03:56:06] knightr: (it's a recording status...)
[03:59:02] wagnerrp: knightr: to be fair, you still write such applications on graphing calculators
[04:08:28] knightr: wagnerrp, LOL, I guess I must buy myself one of those for old times' sake... I wonder what kind of mnemonics these things have though.. (I did mostly x86 assembler, some 68k assembler and a little 6502...)
[04:09:25] wagnerrp: the Ti hardware still runs 68ks
[04:10:08] wagnerrp: at least their older stuff
[04:10:13] wagnerrp: i think their new ones are ARMs
[04:17:05] knightr: LOL, if those 68k based one are still available that could definittely be interesting...
[04:19:26] wagnerrp: the -89 and -92 both run a 10MHz 68k with a few hundred kb or memory, and a few mb of flash
[04:40:37] ThisNewGuy (ThisNewGuy! has joined #mythtv
[04:41:36] Seeker` (Seeker`!~cjo20@unaffiliated/seeker) has quit (Read error: Operation timed out)
[04:42:24] ThisNewGuy1 (ThisNewGuy1! has quit (Ping timeout: 256 seconds)
[04:43:01] highzeth (highzeth! has quit (Ping timeout: 248 seconds)
[04:43:39] highzeth (highzeth! has joined #mythtv
[04:44:36] Seeker` (Seeker`! has joined #mythtv
[04:44:36] Seeker` (Seeker`! has quit (Changing host)
[04:44:37] Seeker` (Seeker`!~cjo20@unaffiliated/seeker) has joined #mythtv
[04:51:50] jeff999 (jeff999! has left #mythtv ("Leaving")
[05:48:52] jya: wagnerrp: they are mostly ARMs these days. Only the TI89 and Voyager 200 are custom 68k. All the other (83, 73, 84 etc) are Z80
[05:49:30] jya: the newer nspire range are all ARMs.. that includes the 83/84 clone
[05:50:55] wagnerrp: they still sell the 83/84?
[05:51:13] jya: they sure do.. it's still their biggest seller, and by far
[05:51:16] jya: over 1M a year
[05:58:13] xris: meta data grabber seems to be broken for me. what's the verbose flag to debug further?
[05:59:16] xris: looks like it's either not searching tmdb, or it's not getting rid of underscores
[06:46:46] Yancho (Yancho!~mpulis@unaffiliated/yancho) has quit (Read error: Connection reset by peer)
[07:16:27] jya: I hope releasing the beta doesn't mean we'll have to create a new fixes/0.25 branch.. maintaining the two will add unecessary overhead
[07:19:17] stoffel (stoffel! has joined #mythtv
[07:19:22] Beirdo: the branch is cut at release
[07:19:47] Beirdo: and really, once we release 0.25, we don't maintain 0.24
[07:20:12] Beirdo: at least by our current methodology
[07:20:29] jeff999 (jeff999! has joined #mythtv
[07:20:40] jeff999 (jeff999! has left #mythtv ()
[07:26:24] Beirdo: all we do at beta is drop a tag and archive
[07:26:32] Beirdo: AFAIK
[07:28:46] Beirdo: stuartm: were you planning on making the beta release?
[07:29:03] Beirdo: being 30min to midnight Pacific Time :)
[07:30:08] MaverickTech (MaverickTech! has quit (Ping timeout: 240 seconds)
[07:30:37] MaverickTech (MaverickTech! has joined #mythtv
[07:45:32] jya: Beirdo: just that for 0.24, I remember being cut a lot earlier, and having to make two commits instead of one for a while
[07:45:52] jya: is down for me
[07:46:43] Beirdo: works fine hee
[07:46:46] Beirdo: here rather
[07:47:44] jya: it blocks at hops 18 18 ( 218.483 ms 211.419 ms 211.375 ms
[07:48:23] Beirdo: anyways, AFAIK, fixes/0.25 starts at the point where we make v0.25 release
[07:48:51] jya: good.. so for 0.24 that was a one off I gather
[07:49:19] Beirdo: well, we are discussing on the mailing list how to do it ongoing
[07:50:03] Beirdo: but to my knowledge, 0.25 will be done the same as before, but with the fun of a complete freeze immediately after the release so we can swing commits over to our server
[07:50:16] Beirdo: which should take only a few minutes
[07:56:54] warped (warped!~piotro@ has joined #mythtv
[07:57:39] jya: is the new git server up and running already?
[07:59:11] Beirdo: yes, but I need to do the last-minute resync, then enable mirroring to github
[07:59:37] Beirdo: and we need ssh keys from all devs (don't worry, there will be an email with details soon)
[08:01:03] Beirdo: OK, generating tarfiles
[08:02:07] jya: no chance for a last commit?
[08:04:44] Beirdo: does it NEED to be in the beta rather than RC1?
[08:05:55] jya: not really
[08:06:11] jya: it's just a change required to build universal mac libraries
[08:06:27] jya: the flags used to link the ffmpeg dylib aren't good
[08:06:52] jya: I'm sure this has been discussed earlier, but don't we do an alpha normally before a beta?
[08:07:37] Beirdo: we've had an alpha for weeks, have we not? I don't recall releasing an alpha before
[08:07:59] jya: we've released an alpha?
[08:08:05] jya: news to me
[08:08:40] Beirdo: essentially, once we code freeze, isn't that an alpha with or without the release tarballs?
[08:09:01] jya: i don't see it that way...
[08:09:31] jya: FreeBSD have feature freeze, alpha release, beta release, rc and in all that complete code freeze
[08:11:19] Beirdo: you saw the same release schedule email that I saw. If we wanted to add more steps, that woulda been the time to add them. I'm going by what we agreed on :)
[08:11:52] Beirdo: now for *future* releases, yah, there may be some validity to it
[08:12:14] jya: oh well… I think I'm now done with the mac build…
[08:12:22] jya: perfect timing with a release..
[08:12:22] Beirdo: but again, that whole process is being discussed on the mailing list
[08:16:06] stoffel (stoffel! has quit (Remote host closed the connection)
[08:18:17] Beirdo: yeah, that will be in for rc1 in 2 weeks, I think it is?
[08:18:54] Beirdo: meanwhile we do the "whack-a-bug" for a bit longer :)
[08:19:46] jya: actually, I think 0.25 as it is now is more stable that 0.24 at release
[08:20:06] Beirdo: for sure
[08:20:38] Beirdo: I'm just hoping no massive blockers come up in the next month
[08:31:23] warped (warped!~piotro@ has quit (Remote host closed the connection)
[08:37:18] Beirdo: OK, I think I shall head to bed
[08:55:26] peitolm: jya do you need an extra tester for the mac build?
[08:55:44] jya: peitolm: sure… what's your environment ?
[08:55:53] peitolm: i managed to get your new bundler down, but couldn't get it to work
[08:56:09] peitolm: early 2008 xserve running 10.7
[08:56:13] jya: I'm trying to get it to compile for 10.5 while building on 10.7. Which according to Apple is not possible, but I want to prove them wrong
[08:56:38] jya: peitolm: no? why not? what was wrong with the one you got earlier?
[08:57:12] jya: I'm going to commit things shortily.. I'm doing a last round of tests… testing with various xcode, on various OS
[08:57:14] peitolm: i hit the same problems as sturta, I couldn't get qmake to work and return sensible answers
[08:57:39] jya: did you install the Qt SDK or the Qt lib package?
[08:57:54] peitolm: can't remembet, one sec
[08:58:20] jya: didn't know that it didn't work for stuarta . he hasn't told me is he even tried
[08:58:43] peitolm: this was a few weeks ago, before your work
[08:58:52] jya: ahhh that
[08:58:54] jya: no wonder
[08:58:57] peitolm: :)
[08:59:06] jya: it does work for me though :)
[08:59:08] peitolm: i've been away over the weekend, so haven't had much time to try
[08:59:16] jya: ok...
[08:59:25] peitolm: let me fire up the xserve and answer your SDK/lib question
[08:59:46] jya: for the new one, you need to go to the Qt web site, and download either the SDK (64 bits only) , or the Qt lib package.
[09:00:25] jya: 4.8 package is 64 bits only, I use the 4.7.4 package, because it's 32/64 and I wanted to get to compile a dual architecture package
[09:00:36] markk (markk! has joined #mythtv
[09:00:46] markk (markk! has left #mythtv ()
[09:01:15] jya: it's so dirty the way I do it, but can't think of another way. I compile for 32 bits first in a specific directory, then another image in a 64 directory, and when it's done, I run a big loop that combine all libs together to make a universal package
[09:01:23] peitolm: qt-mac-opensource-4.8.0-debug-libs.dmg
[09:01:31] peitolm: is what i have
[09:01:43] jya: ok… that will do, but it's 64 bits only
[09:01:47] jya: just so you know
[09:02:01] peitolm: (and the non -debug) one, following reports on the mailling list that something didn't work if it's not compiled with debug
[09:02:14] jya: i use the non-debug one just fine
[09:02:26] peitolm: not a problem for me, all the macs in the house are 64bit
[09:02:29] peitolm: i think
[09:02:30] jya: however, I have to build myth with --compile-mode=debug
[09:02:43] jya: i'm certain it's due to a bug in llvm-4.2 shipped by apple
[09:02:59] jya: if I build using llvm-g++ 4.2 from the llvm source tree it works
[09:03:33] jya: there's no need to worry about setting debug or not, the script check your version of c compiler and enable debug if needed
[09:03:44] peitolm: ok.
[09:04:03] peitolm: right, which bin and plugins do i need, i guess i need whereever qmake is
[09:04:45] jya: I put it in in help ...
[09:05:05] jya: do you have the patch I posted 2–3 days ago?
[09:05:27] peitolm: i have a patch that produced
[09:05:37] jya: if using the QtSDK, or Qt you've compiled from elsewhere, you use the -qtsdk option
[09:06:06] peitolm: i think i'm using the libs am i not?
[09:06:08] jya: for the QtSDK from Qt, that -qtsdk ~/QtSDK/Desktop/Qt/4.8.0/gcc
[09:06:31] jya: if one Qt you compiled yourself (or via macports), it's ~/QtSdk
[09:06:36] jya: if using the binary packages:
[09:07:04] stuarta: i haven't had a chance to try the build yet
[09:07:05] jya: it's -qtbin /usr/bin -qtplugins -qtplugins /Developer/Applications/Qt/plugins
[09:07:36] ** jya anxious.. trying the build on 10.5 ... **
[09:08:05] jya: it works !!!!
[09:08:07] jya: awesome
[09:08:44] jya: can't see my backend though
[09:10:15] jya: just UPnP not working on that VM… werid
[09:12:41] jya: peitolm: note that I haven't tried with fixes/0.24, and I'm pretty certain it won't compile
[09:12:49] jya: you need to use master HEAD
[09:12:53] peitolm: trying to build trunk
[09:13:17] jya: I also haven't tried the git part of it.. I kept the code from the earlier version
[09:13:24] jya: almost everything else has been changed
[09:13:45] jya: so I don't know if it does a git checkout properly or not
[09:13:53] jya: I always compile against my git tree
[09:13:54] ** peitolm will verify :) **
[09:14:06] peitolm: might be nice if it told you what it was fetching
[09:14:07] jya: and actually, I'm thinking of completely removing the git part
[09:14:28] jya: I don't see the problem in making people doing a git clone themselves first, and pointing the script to their checkout
[09:14:47] peitolm: hmm, might have found a typo 'rm: libtoolT no such file or directory"
[09:14:55] jya: if you run it with -verbose , it's very verbose
[09:15:10] jya: let me check
[09:15:23] jya: i have made a lot of mode since I posted the version you're using
[09:15:32] jya: but building the dependencies haven't changed
[09:16:18] jya: peitolm: there's no libtoolIT anywhere
[09:16:36] peitolm: libtoolT
[09:16:37] peitolm: no I
[09:16:51] jya: there's not custom libtool either
[09:17:50] jya: taglib was a bit of a bitch to compile in universal mode. they use an old configure and it's buggy. so I had to re-generate the configure and makefile, for that I download all the autotools packages and recompile them all.
[09:17:54] jya: just to fix their makefile
[09:18:31] jya: the newer taglib uses cmake instead.. I tried to compile cmake instead, but that stuff is monstruous, after 15 minutes it was still compiling
[09:18:34] peitolm: 'rm: libtoolT: No such file or directory'
[09:18:43] peitolm: that was part of the mysql build section
[09:18:54] jya: ah
[09:19:07] jya: has it downloaded the source code for the Qt already?
[09:19:15] jya: that itself takes 20–30 minutes here
[09:19:25] jya: I need the source code to compile the mysql plugin
[09:19:44] peitolm: the download or the build?
[09:19:55] peitolm: it's just build mysql
[09:19:57] jya: the download
[09:20:03] jya: the build is about 2–3s
[09:20:08] peitolm: and downloaded 227M of something
[09:20:10] jya: it's a tiny module
[09:20:34] jya: the qt package doesn't contain the mysql plugin, so I have to compile it
[09:21:31] jya: peitolm: I think the error you are seeing in libtool
[09:21:40] jya: is in the package itself
[09:21:47] jya: because I don't do any of that
[09:21:57] jya: probably them doing a make clean or something
[09:22:05] jya: anyhow.. you need the new version
[09:22:09] jya: hold a sec
[09:22:29] peitolm: it continued past it, so i don't think it's an issue
[09:22:38] peitolm: although i have no idea what it's currently doing
[09:23:02] peitolm: ah, it's unpacking qt
[09:25:06] jya:
[09:25:56] peitolm: IO on this box is stupid slow
[09:26:23] jya: yeah, doing a full build on my box, in universal mode is 9m31 all up
[09:26:42] peitolm: compile is quick, but writing is slow, the raid battery has failed
[09:26:56] jya: i build on a ssd
[09:27:09] jya: though having said that, it makes 0 difference in compile time
[09:27:34] jya: when i compile of the hard drive, or off the SSD it's almost identical time (give or take 1–2s)
[09:30:54] peitolm: if i get the battery fixed, then it should be a lot better
[09:31:12] jya: why would the battery make a difference?
[09:31:23] jya: isn't it only used for the case there's a power loss ?
[09:31:33] stuarta: because the controllers disable the write cache when the battery dies
[09:31:39] jya: i see
[09:31:43] jya: makes sense
[09:35:07] peitolm: jya: checking size of short... configure: error: cannot compute sizeof (short), 77
[09:35:13] jya: ahah
[09:35:22] jya: someone else on the list had this error
[09:35:26] jya: for libogg I think
[09:35:42] jya: you're using XCode 4.2 right?
[09:36:04] peitolm: how would i check
[09:36:16] jya: start xcode, About xcode ?
[09:36:30] peitolm: no access to screen at the moment
[09:36:59] jya: the guy on the user list saw that error for one of the dependency
[09:37:04] jya: i think it was libogg
[09:37:04] jya: xcode-select -version
[09:37:23] jya: xcode 4.3 is 2307
[09:37:30] jya: was just released a few days ago
[09:38:49] peitolm: xcode-select version 2003.
[09:41:21] jya: ah here:
[09:41:27] jya: xcodebuild -version
[09:42:30] peitolm: Xcode 4.2.1
[09:42:54] jya: try with the new script 've posted
[09:43:05] jya: I've upped the version on all dependencies
[09:43:10] jya: maybe they've fixed that one
[09:43:23] jya: otherwise I'm going to install xcode 4.2 here and test...
[09:44:35] jya: yeah, that was libogg
[09:44:44] jya: what libogg is it trying ?
[09:44:52] jya: the issue was with 1.2.1, I've upped to 1.3
[09:45:49] stuartm: <jya> FreeBSD have feature freeze, alpha release, beta release, rc and in all that complete code freeze << So nothing changes between the alpha and release?
[09:46:24] jya: stuartm: the code freeze is something that you need to get everything reviewed by a core panel before it's committed
[09:48:06] peitolm: trying again
[09:48:38] peitolm: new error, configure: error: No 16 bit type found on this platform!
[09:49:10] peitolm: that's with libogg
[09:49:14] stuartm: fwiw, that's a little too strict for mythtv, it would take the duration of the beta just to get agreement on some fixes :) Most people are too busy even to work on fixes, nevermind reviewing other people's work
[09:50:41] jya: is that with the latest script and libogg 1.3.0 ?
[09:51:11] peitolm: latest sript, yes
[09:51:15] jya: there are no particular build instruction for libogg, it's just the default, configure, make, make install
[09:51:52] peitolm: and i have 1.1.2 and 1.3.0
[09:52:34] jya: it probably started with 1.1.2 with the previous script
[09:52:43] jya: and after you changed to the new one, it used 1.3.0
[09:53:50] peitolm: yeah, probably
[09:54:19] peitolm: config.log has a nasty looking error 'ld: lto: could not merge in /var/folders/5t/_wyfr81n4zq2psqqv0dr5r7m0000gn/T//cctkVEPz.o because Unknown instruction for architecture x86_64'
[09:54:44] peitolm: although i get a couple of them
[09:56:42] peitolm: looks like you were part of a discussion about it
[09:56:47] peitolm: (on mythtv-users)
[09:59:36] jya: yeah, I got that with LLVM on 4.3 , but they disappear when using debug mode
[09:59:58] jya: i wouldn't be surprised if it was a bug in the 4.2 llvm-gcc
[10:00:16] jya: I'm going to download 4.2 once again, and try there
[10:01:17] jya: I'm going to commit the code now, and try again on my laptop…
[10:01:23] peitolm: o.k., if it is a bug in 4.2, then i'll bump it to 4.3 when i get home
[10:01:42] jya: i still would like to gind a way around it...
[10:01:57] jya: 4.3 is a bit different to 4.2
[10:02:04] jya: it's now a self contained package.
[10:02:37] jya: make sure you completely remove 4.2 first, the 4.3 when you first run it offers to uninstall it, but it doesn't do a great job and leave the command line tool (which is what we need)
[10:02:46] jya: to fully uninstall 4.2, you do:
[10:06:52] jya: sudo /Library/Developer/4.2/uninstall-devtools --mode=all
[10:09:43] jya: hum.. I think I found another bug on 32 bits build with lamemp3
[10:09:53] jya: allright.. I'm off from this machine, will come back later
[10:10:28] peitolm: o.k. catch you this evening maybe
[10:10:32] xavierh_ (xavierh_! has joined #mythtv
[10:14:42] ikonia (ikonia!~irc@unaffiliated/ikonia) has joined #mythtv
[10:16:53] jya_ (jya_!~jyavenard@mythtv/developer/jya) has joined #mythtv
[10:17:01] jya_ (jya_!~jyavenard@mythtv/developer/jya) has quit (Remote host closed the connection)
[10:17:17] ikonia (ikonia!~irc@unaffiliated/ikonia) has left #mythtv ()
[10:17:44] jya_ (jya_!~jyavenard@mythtv/developer/jya) has joined #mythtv
[10:20:04] hank (hank! has quit (Remote host closed the connection)
[10:22:14] dblain (dblain!~dblain@mythtv/developer/dblain) has quit (Ping timeout: 245 seconds)
[10:23:16] dlblog (dlblog! has quit (Ping timeout: 276 seconds)
[10:26:32] dlblog (dlblog! has joined #mythtv
[10:26:37] dblain (dblain! has joined #mythtv
[10:26:38] dblain (dblain! has quit (Changing host)
[10:26:38] dblain (dblain!~dblain@mythtv/developer/dblain) has joined #mythtv
[10:38:13] xavierh_ (xavierh_! has quit (Ping timeout: 276 seconds)
[10:44:56] stuartm: sphery: are you likely to look at again for 0.25?
[10:56:10] jya_: peitolm: i have installed 4.2
[10:56:16] jya_: and i see the problem there too
[10:56:31] jya_: libogg fails to compile
[11:05:02] Guest74934 (Guest74934! has quit (Remote host closed the connection)
[11:05:53] mike|2 (mike|2! has joined #mythtv
[11:05:54] jya_: this is very weird.. if i run the exact same commands manually, it works
[11:07:26] stuarta: environment then
[11:20:11] jya_: stuarta: I have reproduced precisely the same environment
[11:21:17] stuarta: i mean what is in the output of 'env'
[11:21:50] xavierh: stuartm: would you like me ta have a look at channel grouping from the Program guide ?
[11:29:27] stuartm: xavierh: if you have the time and interest, then sure, that would be great – ideally it should work exactly like recording groups in PBB
[11:30:29] xavierh: stuartm: I was thinking something like this: 10:56 < jya_> libogg fails to compile
[11:30:37] xavierh: damn
[11:31:04] jya_: ?
[11:31:23] stuartm: xavierh: that's not quite what I had in mind
[11:31:38] stuartm: how would that work exactly?
[11:31:44] xavierh: jya, stuartm: sorry bad paste :)
[11:31:48] stuartm: :p
[11:31:54] xavierh:
[11:32:06] xavierh: what is PBB so I can have a look ?
[11:32:42] stuartm: 'Watch Recordings', PBB (Playbackbox is it's historic name and the name it uses internally)
[11:32:54] xavierh: thx
[11:33:45] xavierh: stuartm: yes I am interested as my MythUI conversion works, but is not really ideal
[11:34:01] stuartm: in 'Watch Recordings' you can 'Change Recording Group' for a recording, that shows a menu with a list of available groups and at the top an 'Add new' option
[11:34:44] stuartm: xavierh: well I can do it easily enough, it's a pretty quick job, but not until we're out of feature freeze for 0.25
[11:35:02] xavierh: stuartm: a recording can be only in one group right ?, I don't have any recording here to check
[11:36:23] stuartm: yes, can channels be in multiple groups? I didn't know that (except maybe in both favourites and another group), I figured we'd handle that by having a separate 'Toggle favourite' option in the menu
[11:37:19] stuartm: well that's not too difficult, the same menu but with checkboxes enabled in the buttonlist instead
[11:37:33] xavierh: yes they can
[11:37:34] jya_: definitely seems to be a bug in xcode 4.2 ; i'm reading the same ld error with many projects, including libre office, cairo
[11:37:50] xavierh: ok for checkboxes, we have a go at it
[11:40:04] stuartm: xavierh: your proposal is fine, although a single 'Group Options' menu (wording subject to change) which shows a list with checkboxes allows additional and removal from a group
[11:40:19] stuartm: sound good?
[11:43:19] xavierh: stuartm: perfect
[11:44:21] xavierh: also I am surprise the the info key does not do anything in program guide already .
[11:44:33] xavierh: surprised that
[11:47:51] seeker (seeker!~seeker@unaffiliated/seeker) has joined #mythtv
[11:49:46] stuartm: it should ...
[11:50:08] stuartm: it does here, brings up the Program Details, as it does elsewhere
[11:55:32] zombor (zombor!~zombor_@kohana/developer/zombor) has joined #mythtv
[11:55:59] xavierh: sorry for the noise, I have got some channel here but no data so ...
[11:57:41] zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Remote host closed the connection)
[12:00:12] jya_: stuarta: this has to be the strangest thing ever… libogg: ./configure -> ok ; CFLAGS="" ./configure -> not okay
[12:00:57] stuarta: look in the config.ep
[12:02:27] jya_: i am...
[12:02:32] jya_: nothing special there
[12:03:51] jya_: not sure if it's worth bothering about the problem when it's fine with xcode 4.1 and 4.3, just not 4.2
[12:04:36] jya_: ah ah
[12:05:07] jya_: another case of llvm-gcc crashing when compiling without debugging option
[12:05:56] jya_: correction: CC="-O2" is what it needs
[12:18:42] stuartm: someone's had their email account compromised ...
[12:30:35] jya_: where about?
[12:33:23] jya_: now that's weird.. it's trying to use distcc, i have never set that up *ever*
[12:38:01] stuartm: jya_: spam to the -user and -dev lists came from a genuine user's account, he's been subscribed for 3+ years
[12:38:17] stuartm: and he's even been into IRC a few times
[12:38:55] stuartm: !seen nuonguy
[12:38:56] MythLogBot: nuonguy has not been seen here
[12:40:23] jya_: i guess it's been filtered before getting to the list, i'm not seeing any spam
[12:42:50] stuartm: I don't filter the mailing lists because I need to see the spam to ban the spammer
[12:43:39] seeker (seeker!~seeker@unaffiliated/seeker) has quit (Quit: Colloquy for iPhone -
[12:45:58] jya_ (jya_!~jyavenard@mythtv/developer/jya) has quit (Read error: Connection reset by peer)
[12:46:46] jya_ (jya_!~jyavenard@mythtv/developer/jya) has joined #mythtv
[12:52:11] knightr (knightr!~knightr@mythtv/developer/knightr) has quit (Read error: Connection reset by peer)
[12:54:43] sraue (sraue!~stephan@xbmc/staff/sraue) has quit (Ping timeout: 265 seconds)
[12:57:50] jya_ (jya_!~jyavenard@mythtv/developer/jya) has quit (Read error: Connection reset by peer)
[12:58:50] jya_ (jya_!~jyavenard@mythtv/developer/jya) has joined #mythtv
[13:06:04] davide_ (davide_!~david@mythtv/developer/gigem) has quit (Remote host closed the connection)
[13:06:25] davide_ (davide_!~david@mythtv/developer/gigem) has joined #mythtv
[13:17:24] xavierh: stuartm: I had a quid look athe the Guide on 0.22-???? (will have a look later on 0.25 to see if it is different), pressing the Info key bring the recording option. sticking the channel grouping functionnality will not make sense.
[13:19:15] stuartm: bah, I screwed up, deleted the 0.24.2 milestone when I meant to change the date
[13:19:56] stuartm: xavierh: it's different with 0.25, very different
[13:20:27] stuartm: and it's the Menu key you want, not Info
[13:20:34] xavierh: I need to populate may fake channel so see it then
[13:40:00] sraue (sraue!~stephan@xbmc/staff/sraue) has joined #mythtv
[13:59:15] jya: peitolm: If you give a try to the new mac build script. it's in packaging git
[14:02:25] peitolm: one sec
[14:02:43] peitolm: well, it will be longer than a sec, as i have to power the thing up, but you know what i mean :)
[14:02:55] stuarta: tickety tock
[14:04:32] peitolm: you committed the latest?
[14:04:47] peitolm: git's complaining, let me see if a force will fix it
[14:05:54] jya: i did commit it
[14:06:16] peitolm: cool
[14:06:17] jya: libogg and libvorbis make llvm-gcc crash unless you use -O2
[14:06:19] peitolm: right
[14:06:27] jya: from xcode 4.2 that is
[14:06:57] jya: so what I do is check if you're using 4.2 and force -O2.. and in the mean time I added a user option to configure that
[14:07:26] peitolm: configure: error: No 16 bit type found on this platform!
[14:07:33] jya: I should have actually checked the llvm-g++ build
[14:07:35] jya: ?
[14:07:44] jya: are you sure you're using the last one?
[14:07:54] jya: maybe I didn't push.. i 'm a bit tired
[14:07:59] jya: oh..
[14:08:12] peitolm: i did a git reset --hard HEAD && git pull
[14:08:16] jya: make sure you delete the libogg and libvorbis folder in .osx-packager/src
[14:08:32] jya: because the configure you ran earlier would have left crap there
[14:08:38] peitolm: it asked me to move the out of the way
[14:08:44] peitolm: i've removed libogg
[14:08:48] peitolm: i'll try libvorbis
[14:09:04] jya: i did push
[14:09:07] peitolm: don't have any libvorbis
[14:09:25] jya: right.. it would have failed for the same reason as libogg after
[14:10:25] jya: hum… -O2 cause the same error I saw when compiling the mythtv filters without the debug option...
[14:10:34] jya: Jeez… this compiler is terrible…
[14:10:39] peitolm: i don't get your error message
[14:10:59] jya: which one?
[14:11:14] jya: that you're using XCode 4.2 upgrade !
[14:11:15] jya: ?
[14:11:59] peitolm: yeah
[14:12:09] peitolm: i think your regex is wonky
[14:12:23] jya: how could I be clearer? 4.2 is buggy, upgrade...
[14:12:23] peitolm: if ( $XCODEVER =~ m/Xcode\s+([0–9]+\.[0–9]+(\.[0–9]+)?)/ )
[14:12:28] jya: to 4.3 for the time being
[14:12:35] jya: it handles X.X.X
[14:12:40] jya: where the last .X is optional
[14:12:56] jya: so 4.2.1
[14:13:02] peitolm: in which case your comparison is wrong
[14:13:07] jya: or 4.2
[14:13:08] peitolm: if ( $1 eq "4.2" )
[14:13:23] peitolm: because $1 is "4.2.1"
[14:13:26] jya: the xcode I downloaded from my SDK account was 4.2
[14:13:39] jya: for you it is.. but I couldn't find 4.2.1
[14:13:57] peitolm: but your check should probably be 4.2.X ?
[14:14:01] jya: ah yes..
[14:14:06] jya: for the time being then
[14:14:13] jya: just run the script with -olevel 2
[14:14:43] peitolm: you could do if ( $XCODEVER =~ m/Xcode\s+([0–9]+)\.([0–9]+)(\.[0–9]+)?/ )
[14:15:13] peitolm: if ($1 ==4 && $2 == 2
[14:15:19] jya: if the compilation fail later on the plugins, re-run the script without the -olevel
[14:15:37] peitolm: i'm not sure why you're nesting brackets
[14:15:42] jya: I'd do m/^4\.2/
[14:15:57] jya: so both the . and the [0–9] are optional
[14:16:12] Jordack (Jordack! has joined #mythtv
[14:16:20] jya: you like to give me crap on my regex you.. the other day as well
[14:16:38] stuarta: ah the fun of regexps
[14:16:45] peitolm: i'll stop if you want :) it's only good natured
[14:17:42] ** peitolm thinks (\d+)\.(\d+)\.(\d+) is cleaner but :) **
[14:17:58] jya: so we have the depenendencies not building without -O2 with xcode 4.2
[14:18:00] peitolm: you work really is appreciated here
[14:18:07] jya: but mythtv not compiling with -O2 with 4.3
[14:18:15] jya: how can I ever get it right ?
[14:18:29] peitolm: fix mythtv:) *ducks*
[14:18:40] jya: that's llvm-g++ crashing
[14:18:42] ** stuarta larts peitolm **
[14:18:51] peitolm: sorry
[14:19:09] jya: actually.. I had a look at how they compile ffmpeg in macports
[14:19:46] jya: what they've done, is put a dependency on llvm-4.2 upstream.. they compile llvm-4.2 themselves, then compile ffmpeg as xcode's one , while the same version, crashes
[14:20:05] jya: ahah
[14:20:42] jya: the taglib compiled with -O2, make llvm-gcc crash when linked to libmythmusic
[14:21:03] peitolm: right, plugins compiled, now for mythtv itself
[14:21:14] zombor (zombor!~zombor_@kohana/developer/zombor) has joined #mythtv
[14:21:39] jya: let see if it fails… I didn't wait til the end as the machine with 4.2 is a macbook air.. not the fastest beast
[14:22:17] peitolm: got to get the mythtv git
[14:22:24] jya: ouch
[14:22:29] jya: that will take a while
[14:22:50] peitolm: 300–400KiB/s
[14:22:53] peitolm: but yes
[14:29:39] xavierh: stuartm: I had a llok a 0.25, in the MENU menu in program guide, the is 2 entries, "add to channel group" and "choose Channel group". both pop a message "You don't have any channel group defined". So the work have been started but not finished ?
[14:29:55] jya: peitolm: I've pushed the change to cater for 4.2.1
[14:32:11] jya: i need to find a way to set the version number of the package properly
[14:32:53] j-rod|afk is now known as j-rod
[14:33:10] peitolm: jya i'll let you know how it goes once something happens :)
[14:33:29] peitolm: it just finished the git pull, so hopefully it'll start building
[14:34:29] jya: hopefully..I've redone mine after deleting the .osx-packager/src subfolders, and it took 7 minutes all up (both dependencies, and mythtv universal: so build twice)
[14:34:47] stuartm: xavierh: I guess so
[14:35:10] stuartm: xavierh: I hadn't noticed before, but I don't use channel groups so I wasn't really looking
[14:35:34] xavierh: sound strange to add the menu entry and ... that's it
[14:36:25] xavierh: or it might just be pbroken, I am a mount late so
[14:36:27] xavierh: ...
[14:36:35] xavierh: month
[14:36:38] stuartm: hmm, I don't even see those entries here, maybe because I have channel groups disabled in the settings or something
[14:36:55] peitolm: jya apart from stacks of messages about distcc (being unable to distribute), it seems to be o.k.
[14:37:03] jya: yeah
[14:37:08] peitolm: why are we using distcc again?
[14:37:11] jya: for some reasons, xcode 4.2 ships with distcc
[14:37:16] jya: so the configure tries to use it
[14:37:17] stuarta: --disable-distcc
[14:37:24] xavierh: stuartm: Can you disable it ? do remember seing this option
[14:37:24] jya: it doesn't come with 4.3
[14:37:27] stuarta: had to do that on the buildbot
[14:37:45] jya: what I did is just rename /usr/bin/distcc to /usr/bin/distcc.unactive
[14:37:45] peitolm: will this be a fully re-locatable bundle?
[14:37:50] stuartm: xavierh: maybe it's only shown if you have at least one channel group defined, testing that theory now
[14:38:02] jya: and I certainly didn't want to add an option on the builder for that
[14:38:08] stuartm: nope, that still doesn't work
[14:38:14] jya: peitolm: yes.
[14:38:24] jya: as long as it's 64 bits
[14:38:31] jya: by default it compiles for your host
[14:38:43] jya: you can set -universal
[14:38:49] stuartm: xavierh: I'll look at the code I guess, but in a few minutes, I'm preoccupied with something else atm
[14:39:12] jya: but if you've Qt is 64 or 32 bits only, it will not try to compile (I added a test checking for the format of the Qt libs)
[14:39:34] jya: peitolm: had you ever compiled mythtv on a mac before?
[14:39:40] xavierh: stuartm: I did have only Favorite group to start with and I had the same behaviour. No worry, I will have a look myself
[14:39:59] jya: there was a long bug that prevented to run the build, unless you renamed the .osx-packager package.. that got fixed today (it was a mythtv issue)
[14:41:00] peitolm: jya: yep, only for 10.6
[14:41:24] jya: so you probably hit the issue with messages about loading 2 versions of Qt
[14:41:25] peitolm: i had that, and then it wouldn't run on t'wifes machine, so i ended up doing a full build on there
[14:41:36] jya: really?
[14:41:39] peitolm: yep, and found the work-around (for the Qt issue)
[14:41:40] jya: that's surprising though
[14:42:00] peitolm: yeah, it was around the time I couldn't get audio working, nor her second screen
[14:42:10] peitolm: i fixed it by smacking the database
[14:43:18] jya: if you're running lion, and you want passthroug audio
[14:43:23] jya: lion is very buggy there
[14:43:35] jya: you can play digital audio once.. after that no more audio
[14:43:45] jya: (same issues for any applications on lion)
[14:44:10] peitolm: really, i use digital audio every day on my lion desktop
[14:44:15] jya: the trick to get audio again, is go into the system sound preferences, switch to another audio device if you have one, and go back to the original
[14:44:17] jya: or reboot
[14:44:20] peitolm: not sure it's passthrough though
[14:44:30] jya: the issue is with passthrough only
[14:44:33] peitolm: i'll test that before i give it to the wife
[14:44:34] jya: AC3 or DTS
[14:44:50] jya: if you hafven't seen the problem, you surely aren't
[14:44:54] peitolm: actually, she'll be on analogue, so that won't affect her
[14:45:08] stuartm: xavierh: ah, you need a group defined and at least one recording in that group (which is stupid for 'Add to Channel Group')
[14:45:32] peitolm: o.k. bad news, that blew up spectacularly
[14:45:44] jya: yeah?
[14:45:59] jya: in one of the plugin saying some symbols are missing?
[14:46:15] jya: I had 2 kernel panic on my mini compiling :)
[14:47:03] peitolm:
[14:47:15] jya: that's the one...
[14:47:55] jya: llvm-g++ drops some code it shouldn't
[14:48:05] jya: so to fix that try the following
[14:48:13] jya: git pull on packaging
[14:48:15] xavierh: stuartm: that what I just noticed, also there is no clear indication that the channel is or not in a group already
[14:48:32] jya: and run it with -olevel 1 and if that fails -olevel 0
[14:48:54] jya: ld: warning: -read_only_relocs cannot be used with x86_64 I'm pretty sure I fixed that warning messag
[14:48:56] xavierh: It seems that we can remove a channel from a group (looking at the code) but not really sure where in the menu
[14:48:57] peitolm: do i need to tell it to not do another git clone again?
[14:49:11] stuartm: xavierh: yeah, the UI needs re-writing along the lines we discussed, it's not good enough right now
[14:49:12] jya: do -nohead -force for that
[14:49:22] jya: this keep the git tree untouched
[14:49:52] stuarta: it shouldn't redo the git clone anyway, it will just update it iirc
[14:50:44] jya: stuarta: that's one thing i didn't test
[14:51:42] stuartm: so, 24 of the open tickets (or 8%) are for mythweb, most have been open for years, since mythweb is basically dead with no active maintainer should we just close them and let the new, completely re-written mythweb replace it?
[14:52:05] stuarta: huh?
[14:52:06] jya: works pretty well for something unmaintained !
[14:54:23] peitolm: what's the ETA for the new mythweb?
[14:54:53] stuarta: what new mythweb? have i been living under a rock again?
[14:56:01] peitolm: some embedded thing isn't it?
[14:56:05] stuartm: stuarta: we're going to use the built in http server in mythbackend, we're already serving up a simple collection of pages in 0.25, but nothing yet to rival mythweb – probably for 0.26
[14:56:44] stuarta: oh that thing...
[14:57:05] stuartm: jya: only significant code commits made have been to fix major breakage, although there's probably lots of minor breakage that has gone unfixed and several features which are ignored by mythweb at the moment
[14:57:58] jya: Well, I guess enabled x86_32 && append SHFLAGS -Wl,-read_only_relocs,suppress doesn't work :( it adds that option all the time
[14:58:19] peitolm: olevel 1 filed
[14:58:31] jya: you mean it worked?
[14:58:46] jya: I got an error in mythmusic when linking against taglib
[14:58:50] stuartm: only two dozen or so commits have been made to mythweb in the last six months, and 2/3s of those were translations
[14:59:08] wagnerrp: Captain_Murdoch: is there a known issue with the HLS stuff playing fast?
[14:59:17] peitolm: jya: failed, sorry
[14:59:22] stuartm: the six months prior to that were little or no better
[14:59:25] jya: I'm guessing you'll hit the same, so do rm -rf .osx-packager/src/taglib-1.6.3
[14:59:38] jya: peitolm: :(
[14:59:49] wagnerrp: this must be how sphery watches tv
[15:00:18] jya: hopefully -olevel 0 will work, otherwise I've have to come with another soltution for xcode 4.2 users
[15:00:38] ** stuarta quite likes mythweb **
[15:00:39] peitolm: if it doesn't, i'll upgrade to xcode 4.3, i'll be heading home in an hour or so anyway
[15:00:42] stuartm: mythweb has more tickets open against it than any other component except the catch-all 'General' component and 'Video Playback'
[15:00:59] stuartm: and that's despite the fact that I've just closed some
[15:01:12] stuartm: it's dead Jim
[15:01:21] ** peitolm notes if it was written a language he was more comfortable in, he'd happily prod it **
[15:01:23] jya: peitolm: if -olevel 0 doesn't work, edit the script and comment-out line 520
[15:01:46] stuartm: stuarta: if you love something enough, then you know when to let it go ;)
[15:02:13] stuartm: or something like that, I heard it in a movie once
[15:02:20] jya: i do like mythweb too.. it's one of the preferred feature when I show people..
[15:02:31] peitolm: mythweb is the only feature i show people
[15:02:48] jya: though I never try other than the recordings and tv listing bit
[15:03:01] stuarta: that why having a service interface in the backend is useful, mythweb can abuse it
[15:03:07] stuarta: use it ;-)
[15:03:28] stuartm: well it's not like we're not replacing it, we are, but at least when we do it will be maintainable by most devs and not just those with PHP experience (and perseverance)
[15:03:36] jya: peitolm: in case you don't see line 520, it;s " $OPT{'olevel'} = "2";"
[15:03:49] jya: to comment out, and don't forget taglib
[15:03:53] stuarta: i find php simpler that gui code
[15:05:06] stuartm: it probably is, but the new html interface isn't going to be gui code, it's html
[15:05:17] stuartm: and C++
[15:05:18] jya: interesting… the 32/64 universal code compress much much better than 32 bits alone… I didn'te expect to have that much duplicated code between the 32 and 64 bits opcodes
[15:05:49] jya: stuarta: php really is brilliant for web work.. no wonder it's so popular
[15:06:07] stuartm: stuarta: I have a lot of experience with php, but even so I avoid mythweb just because it's written in such a way that I find it difficult to impossible to find my way around
[15:06:36] stuarta: it's not the easiest that's for sure
[15:06:44] stuarta: but then nothing complex is
[15:07:01] stuartm: nothing is stopping someone writing a new mythweb in php around the services API, but since no-one seems willing to maintain the existing version I wouldn't get your hopes up
[15:07:09] jya: stuarta peitolm I put my universal packages there: . . . d1a6cc88.dmg . download will be completed in 8 minutes. it's 32/64 bits and works on 10.5 too. I'm off to bed
[15:07:17] stuarta: nn
[15:07:24] stuartm: stuarta: it could be much, much simpler, 'obfuscation' comes to mind
[15:07:45] jpabq_ (jpabq_!~jpabq@mythtv/developer/jpabq) has joined #mythtv
[15:07:46] jya: stuartm: to be honest, I never thought there was that much to do in mythweb, because it works so well as it is…
[15:07:57] stuarta: aye, wants porting away from PATH_INFO too
[15:08:56] wagnerrp: stuartm: #8801 can actually be closed with some level of completeness
[15:08:59] stuartm: jya: the volume of tickets suggests otherwise, but hey, I'm also not proposing to delete it, just close the ticket which have been open for years
[15:09:15] wagnerrp: the UPnP server is integrated with mythvideo, although i believe its still lacking in metadata access
[15:09:33] wagnerrp: alternatively, it can be closed as mythvideo no longer exists for it to be integrated with
[15:10:15] wagnerrp: i think beirdo may have wanted to keep it open until he gets time to do some of the metadata stuff
[15:10:54] stuartm: wagnerrp: ok, I will close
[15:11:26] jya: from the user list "About two weeks ago I started getting periodically long startup times for the frontend, usually 30 – 120 seconds.
[15:11:26] jya: I can't see anything in the frontend or backend logs."
[15:11:40] stuartm: the number of open tickets means that real bug tickets are getting lost in the noise, there's stuff out there with simple and good patches attached which have sat unnoticed for months or longer
[15:11:45] jya: I have seen those hangs on the mac from time to time, I always asumed it was a mac crap only
[15:15:00] peitolm: same error with -olevel 0
[15:16:07] jya: oh.. I know
[15:16:09] jya: do
[15:16:17] jya: -olevel blah
[15:16:29] jya: that's an invalid value, so it won't be used, as if there was nothing
[15:16:40] stichnot: xavierh: Besides the various settings pages, where is there still non-MythUI code? I'm asking because of the jump point execution code.
[15:17:44] peitolm: jya: i'm just commented out 521 and i'm trying with olevel 1
[15:18:04] jya: didn't you try -olevel 1 already?
[15:18:10] jya: you said it bombed out
[15:18:42] peitolm: i did, but as i have 4.2.1, isn't it overriden by #521?
[15:19:01] jya: didn't you do a git pull?
[15:19:04] jya: I've fixed that
[15:19:33] jya: but --olevel with any digits will do
[15:19:39] jya: non-digit
[15:19:44] jya: as if you had nothing
[15:19:46] peitolm: -- or -?
[15:20:03] peitolm: i'll tripple chek in a sec, if it breaks
[15:20:16] jya: any non digits
[15:20:53] ** jya really shouldn't watch the infomercial, I want to buy the hair straightner **
[15:25:19] stuartm: stichnot: part of mythgallery is perhaps the only non-settings screen
[15:28:09] peitolm: o.k. so commenting out #521 a&& rm taglib && olevel 1 failed
[15:28:52] stichnot: stuartm: oh yeah, I forgot about plugins.
[15:29:13] jya: peitolm: i don't get why you try again with -O1
[15:29:37] jya: and it's 520...
[15:30:05] jya: really, all you got to do, is remove taglib, -olevel A and that's it
[15:30:11] jya: pretty convinced it will work
[15:30:51] jya: did you see a "Using LLVM: Forcing debug compile."
[15:31:00] jya: in the log? if not, use -debug
[15:32:06] jya: right.. definitly off now
[15:32:29] peitolm: trying with olevel foo
[15:39:40] jya: peitolm: i just went to check on my MBA, it has failed as you.. -olevel anynondigit did the right thing and I'm fairly convinced it will work okay… otherwise 4.3 that is.. Ciao
[15:39:59] peitolm: ciao
[15:40:02] peitolm: thanks for the effort
[15:51:01] seeker (seeker!~seeker@unaffiliated/seeker) has joined #mythtv
[16:01:53] peitolm: jya: that built
[16:08:16] kwmonroe (kwmonroe!~kwmonroe@ has quit (Quit: Coyote finally caught me)
[16:11:17] kwmonroe (kwmonroe!~kwmonroe@ has joined #mythtv
[16:11:46] rsiebert (rsiebert! has joined #mythtv
[16:13:47] seeker (seeker!~seeker@unaffiliated/seeker) has quit (Quit: Colloquy for iPhone -
[16:14:09] xavierh: stichnot: nothing I am aware of
[16:14:28] stichnot: cool, thanks.
[16:15:14] xavierh: stichnot: I just read, stuartm comment, haven't been using mythgalerry for years
[16:15:34] rsiebert_ (rsiebert_! has quit (Ping timeout: 276 seconds)
[16:18:11] seeker (seeker!~seeker@unaffiliated/seeker) has joined #mythtv
[16:26:30] gregL (gregL! has quit (Read error: Connection reset by peer)
[16:27:23] danielk22: stuartm: Is libmythmpeg2 completely gone? I still see the library in my checkout
[16:28:21] stuartm: danielk22: it's still used for the lossless mpeg2 transcoder, but can't be selected for playback
[16:30:23] CyberKnet (CyberKnet!~s.blomfie@ has joined #mythtv
[16:31:04] CyberKnet: Just wanted to leave a quick note of gratitude for the services API and documentation... I know a lot of hard work goes into this stuff, and wanted to let you guys know it is really appreciated, thanks!
[16:31:10] CyberKnet (CyberKnet!~s.blomfie@ has left #mythtv ()
[16:32:12] danielk22: stuartm: ok. btw thanks for doing all that triage this weekend.
[16:34:11] stuartm: still working through it, makes the commits list pretty noisy but hopefully worthwhile
[16:35:45] stuartm: picked up a few tickets for myself to look at for 0.25, might be over-extending myself but it's better than leaving them to rot :)
[16:39:22] zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Read error: Connection reset by peer)
[16:39:35] zombor (zombor!~zombor_@kohana/developer/zombor) has joined #mythtv
[16:43:39] gregL (gregL! has joined #mythtv
[16:44:45] stuartm: danielk22: seems simple enough?
[17:15:31] xavierh: stuart
[17:16:37] stuartm: stuarta: ^^
[17:17:07] xavierh: stuartm: I had another quick look, you were right, only the non-empty group are shown. The logic is to not allow to change to a group which is empty. but the same list is use for both "add to channel group" and "change channel group"
[17:17:25] xavierh: I have to big fingers :)
[17:17:28] xavierh: too
[17:18:35] stuarta: stuartm: that really can't hurt
[17:18:50] stuartm: xavierh: yeah, they should use different checks
[17:19:23] stuartm: stuarta: that was my thinking, I just didn't want to steal it away from danielk22 without asking first
[17:19:38] stuarta: tbh, it's one debug string change and a tolower()
[17:20:24] xavierh: I will do a patch and submit it later on
[17:34:38] stichnot (stichnot!~chatzilla@mythtv/developer/stichnot) has quit (Ping timeout: 245 seconds)
[17:39:35] kormoc (kormoc!~kormoc@mythtv/developer/kormoc) has joined #mythtv
[17:49:25] stuartm: Beirdo: what's Jonatan's trac username?
[17:50:05] stuartm: assuming natanojl for now, can be changed later
[17:51:17] xavierh_ (xavierh_! has joined #mythtv
[17:55:48] zombor_ (zombor_!~zombor_@kohana/developer/zombor) has joined #mythtv
[17:59:11] zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Ping timeout: 260 seconds)
[18:04:59] davide_ (davide_!~david@mythtv/developer/gigem) has quit (Remote host closed the connection)
[18:05:22] davide_ (davide_!~david@mythtv/developer/gigem) has joined #mythtv
[18:10:24] noahric (noahric!~noahric@ has joined #mythtv
[18:12:30] xavierh: stuartm: fix for channel group in program guide
[18:38:34] sraue (sraue!~stephan@xbmc/staff/sraue) has quit (Read error: Connection reset by peer)
[18:40:20] sraue (sraue!~stephan@xbmc/staff/sraue) has joined #mythtv
[18:41:30] seeker (seeker!~seeker@unaffiliated/seeker) has quit (Remote host closed the connection)
[18:41:43] seeker_ (seeker_!~seeker@unaffiliated/seeker) has joined #mythtv
[18:44:40] knightr (knightr!~knightr@mythtv/developer/knightr) has joined #mythtv
[18:51:41] stuartm: xavierh: thanks
[18:52:34] xavierh_: stuartm: the channel grouping will still need to be reworked for 0.26, I prefer your idea about checkboxes as it does allow to add a channel to multiple group much quicker. also I would like to be able to add multiple channels to the current group (with check boxes too)
[18:54:44] stuartm: xavierh_: I think the UI for mapping multiple channels to different groups is best handled as part of backend setup, where the rest of the channel setup occurs, with the stuff in the frontend just for small tweaks or changes?
[18:56:43] xavierh_: stuartm: that is arguable, I would still like to be able to do it from a 10" UI
[18:57:09] zombor_ is now known as zombor
[18:58:11] xavierh_: also I notce something like this :menuPopup->AddButton(tr("Record This"));
[18:58:20] xavierh_: and then if (resulttext == tr("Recording Options"))
[18:58:57] xavierh_: this thing can break with the translation i.e. if two != string are translated to one
[18:59:06] xavierh_: is that a good practice ?
[19:00:10] david (david! has joined #mythtv
[19:00:33] david is now known as Guest29306
[19:01:55] stuartm: it's not great practice, but it won't break with translation, two identical strings in the same context appear just once in the translation
[19:03:09] stuartm: i.e. they are guaranteed to be translated identically
[19:03:35] stuartm: the translator doesn't see that same string twice in linguist, they see it just the once
[19:04:35] stichnot (stichnot!chatzilla@nat/intel/x-qfcxsodjbroyzpww) has joined #mythtv
[19:04:35] stichnot (stichnot!chatzilla@mythtv/developer/stichnot) has joined #mythtv
[19:04:35] stichnot (stichnot!chatzilla@nat/intel/x-qfcxsodjbroyzpww) has quit (Changing host)
[19:09:55] wagnerrp: ooh, a blocker
[19:10:45] stuartm: backend crashes generally merit blocker status
[19:11:06] danielk22: xavierh: In general those should be something like AddButton(tr("Record This"), QVariant("RECORD_THIS")); if (resulttext == "RECORD_THIS").. but as stuartm noted it will work as is as long as no one changes the text inside one of the tr()'s.
[19:14:40] wagnerrp: something going on with the server? its abnormally slow
[19:14:47] Beirdo: stuartm: you guessed the name correctly
[19:15:18] Beirdo: wagnerrp: load average: 1.45 9.23 7.93
[19:15:30] Beirdo: it looks like someone was hitting it hard for a bit
[19:16:14] danielk22: stuartm: yeah, seems simple enough I just didn't get a chance to look at it yet. feel free to take it.
[19:16:15] stuartm: probably when I clicked on the link to a commit in that new ticket, referenced the trac git browser and not github
[19:17:24] Beirdo: I thought we had disabled it... maybe jsut the link?
[19:18:25] wagnerrp: Beirdo: is there any way to make the logger run synchronously?
[19:19:34] Beirdo: wagnerrp: no, the whole purpose of making it async was to keep it from slowing down the program. What are you trying to accomplish, maybe there's another way to do it.
[19:20:04] wagnerrp: making it run synchronously, so it actually logs everything up to the line that makes it segfault :)
[19:20:31] Beirdo: nope
[19:20:37] xavierh_: stuartm, danielk22: You misunderstood me. you could have 2 != word in english which are translated in one word only. then it would break
[19:21:19] Beirdo: now, if Qt weren't such a PITA with gdb, you should be able to go through the logging queue to see what hasn't been output yet
[19:22:48] stuartm: xavierh_: yeah, possible I guess
[19:23:20] danielk22: xavierh: I still don't understand. As long as each button has unique text after translation it should work. But that code isn't how things are supposed to be done..
[19:23:54] wagnerrp: anyone know off hand what OS yiannividalis uses?
[19:24:24] stuartm: Beirdo: we disabled the direct link, but it's still linking any sha1 mentioned in the ticket or comments
[19:25:03] Beirdo: I thought we'd changed those to being github links
[19:25:09] Beirdo: hmm
[19:25:18] stuartm: danielk22: I believe he's asking what happens if two different buttons end up with the same translation, it's theoretically possible, not likely but possible
[19:25:44] danielk22: Beirdo: I use a similar async logging at work but have a run time switch that enables synchronous logging. Used only when a segfault happens and we need up to the last nanosecond debug. That way a sys-admin can grab a more useful log for the devs while the segfault causing data is still on the wire.
[19:26:38] MythBuild: build #2053 of master-linux-ppc is complete: Failure [failed compile core] Build details are at . . . /builds/2053 blamelist: Karl Dietz < >
[19:26:56] danielk22: stuartm: xavierh: yeah it would break in that case.. but that would be a bad translation, no? you have two options: "Yes, delete all files", "Yes, delete all files" :)
[19:27:43] Beirdo: danielk22: something to consider for 0.26, not happening for 0.25 :)
[19:27:50] stuartm (stuartm!~stuartm@mythtv/developer/stuartm) has quit (Remote host closed the connection)
[19:27:51] xavierh_: stuartm: yes that's what I meant
[19:27:52] MythBuild: build #3038 of master-linux-32bit is complete: Failure [failed compile core] Build details are at . . . /builds/3038 blamelist: Karl Dietz < >
[19:28:00] MythBuild: build #2079 of master-freebsd-64bit is complete: Failure [failed compile core] Build details are at . . . /builds/2079 blamelist: Karl Dietz < >
[19:28:06] danielk22: Beirdo: I agree.
[19:28:09] stuartm (stuartm! has joined #mythtv
[19:28:10] stuartm (stuartm! has quit (Changing host)
[19:28:10] stuartm (stuartm!~stuartm@mythtv/developer/stuartm) has joined #mythtv
[19:28:17] stuartm: Beirdo:
[19:28:33] Beirdo: it does sound like a reasonable plan to add though, I'll put it at the top of my list.
[19:29:34] stuartm: <stuartm> danielk22: right, it wouldn't make any sense to the end user anyway
[19:29:53] MythBuild: build #3293 of master-linux-64bit is complete: Failure [failed compile core] Build details are at . . . /builds/3293 blamelist: Karl Dietz < >
[19:30:05] Beirdo: stuartm: hmmm, that's a pretty useless backtrace
[19:30:07] stuartm: huh, shit
[19:30:31] stuartm: Beirdo: I posted that because it's the comments with the links to the trac browser
[19:30:36] Beirdo: oooh
[19:30:48] Beirdo: gotcha, let me look at THAT ;)
[19:31:25] Beirdo: yeah, those redirect to github when you click on em
[19:32:23] MythBuild: build #1804 of master-debian-stable-64bit is complete: Failure [failed compile core] Build details are at . . . /builds/1804 blamelist: Karl Dietz < >
[19:33:36] seeker_ (seeker_!~seeker@unaffiliated/seeker) has quit (Quit: Colloquy for iPhone -
[19:37:18] dekarl: whaaat, the buildbot calls me?
[19:37:33] stuartm: dekarl: you broke the build
[19:37:37] dekarl: how am I supposed to fix that before the slow build fail :(
[19:38:16] stuartm: Beirdo: hmm, so they do, I guess github was down earlier because it just sat spinning with the trac url in the address bar
[19:38:30] stuartm: Beirdo: ok, sorry for the noise
[19:38:39] kormoc: It was. Github was hacked, so they went defcon-1 for a little
[19:38:59] Beirdo: heh, it happens
[19:39:22] dekarl: stuartm: but thanks for commiting that fix
[19:39:49] stuartm: my fault in the first place for not spotting the missing (
[19:40:01] Beirdo: stuartm: #8801 was still open as there is a need for directory levels in the mythvideo stuff for UPnP. I guess that can be put in a new task :)
[19:40:23] Beirdo: i.e. all videos are lumped together in one listing, no heirarchy
[19:40:49] dekarl: how did you add the typo fix in my name? (I think it wasnt in the original patch) ... => don't trust git to much
[19:40:55] MythBuild: build #872 of master-vista-mingw-32bit is complete: Failure [failed compile] Build details are at . . . t/builds/872 blamelist: Karl Dietz < >
[19:41:08] Beirdo: I like the idea of closing it though, and a more specific one being opened for the remaining deficiency.
[19:52:19] sphery: stuartm: I'm hoping to look at #9819 before 0.25, but I don't know that I'll get to it (nor that I'll feel comfortable pushing the fix during feature freeze). All the patches there are broken, and fixing the issue without breaking other stuff requires someone to write code to use 2 events (do this and it's done events) to replace the dispatchNow() calls--the same calls that you said are "not going away any time soon" in ...
[19:52:25] sphery: ...  :)
[19:52:55] wagnerrp: kormoc: looks like my ebuild generator breaks on the new tags
[19:53:06] wagnerrp: doesnt know what to do with 'v0.25-beta'
[19:53:33] kormoc: youch
[19:53:45] wagnerrp: should be a simple regex fix
[19:54:09] Beirdo: wagnerrp: #10405... seems it's segfaulting during shutdown
[19:54:36] wagnerrp: yeah, i think it was davide that pointed out the issue a week or so back
[19:54:50] Beirdo: no idea why it's shutting down, though
[19:54:55] wagnerrp: the backend was not closing properly when it self terminated on startup
[19:55:11] Beirdo: yeah, that's happened before
[19:55:28] wagnerrp: the likely cause of self termination is not having the proper IPs available
[19:55:38] wagnerrp: due to being run on startup
[19:55:41] Beirdo: quite likely, yeah
[19:55:47] wagnerrp: on a distro that does not properly init their networks
[19:56:00] wagnerrp: (hence my query on trac)
[19:56:03] Beirdo: the m_priv there is probably null somehow
[19:56:20] Beirdo: he really needs to use a debug or profile build
[19:56:22] wagnerrp: the segfault is new, before it would just hang
[19:56:27] Beirdo: no locals on everything...
[19:58:12] MythBuild: build #3294 of master-linux-64bit is complete: Success [build successful] Build details are at . . . /builds/3294
[20:00:18] stuartm: sphery: when I said they weren't going away soon I didn't know there were alternative solutions already coded up (but not committed)
[20:01:09] noahric (noahric!~noahric@ has quit (Quit: noahric)
[20:03:13] sphery: stuartm: there are no coded solutions--all those patches are broken and none use the proper approach with 2 events . See conversation at
[20:03:14] wagnerrp: opensuse... anyone know how they deal with their networks by default?
[20:03:32] wagnerrp: last time i used it for 11.x, they had normal startup scripts
[20:03:38] sphery: stuartm: (that was in response to your "call to fix warnings" a month ago :)
[20:04:41] Beirdo: wagnerrp: no idea, they may have gone with upstart or the like
[20:05:52] sphery: FWIW, I *really* wish these conversations in github's "commit comments" would be held in any one of our multiple locations that are logged/archives (like mythtv-dev mailing list or here or ...)
[20:05:53] stuartm: ubuntu won't init wifi until long after the backend comes up in S5, although localhost should be available by that point
[20:06:29] stuartm: sphery: I don't read them, or pay much attention to what happens on github generally
[20:07:06] MythBuild: build #2080 of master-freebsd-64bit is complete: Success [build successful] Build details are at . . . /builds/2080
[20:07:21] stuartm: wagnerrp: ^^ I had problems with that even before the changes in master, so I wonder how it will cope with 0.25?
[20:07:32] sphery: stuartm: yeah, that's the problem
[20:08:07] stuartm: sphery: I can see if comments on github can be disabled
[20:08:15] sphery: and since the conversations aren't archive, searching the list or IRC logs or ... won't help you find that info, again, in the future when someone wants to know "why did the Mac OS folder for plugins change from PlugIns to Plugins") or whatever
[20:09:31] wagnerrp: stuartm: im actually considering writing some short loop, maybe 30 seconds or so
[20:09:36] stuartm: a quick look at their very basic settings page suggests no
[20:09:51] wagnerrp: where it will sit there for a bit waiting for those addresses to become available, before self terminating
[20:09:53] stuartm: wagnerrp: that seems like a sensible approach
[20:10:36] wagnerrp: but that still doesnt resolve the previous hang and current segfault
[20:10:39] stuartm: might need a little longer than 30 secs for wifi, it can be slow to handshake and assign an IP
[20:11:05] wagnerrp: a backend on wifi?
[20:12:56] wagnerrp: i suppose that could be an issue for the frontend web server, now that it is restricted to the same BackendServerIP
[20:13:21] wagnerrp: same on the notify stuff
[20:15:25] wagnerrp: has anyone besides mark actually tried the pulse-eight stuff?
[20:15:36] wagnerrp: or have any idea what its capable of in its current implementation in mythtv?
[20:16:11] wagnerrp: if were going to list something as supported, it would be nice to be able to tell people specifically what it does
[20:16:15] sphery: stuartm: probably would be nice (but don't waste a ton of time on it)--especially since if github becomes a secondary/"public" mirror of our own-hosted repo, it will just make those notes even more disconnected for us
[20:16:54] sphery: ah, just saw you said it looks like it's not possible... nvm
[20:20:11] MythBuild: build #3039 of master-linux-32bit is complete: Success [build successful] Build details are at . . . /builds/3039
[20:20:34] stuartm: yeah, you can't do a whole lot of customisation on github, at least not with a free account
[20:20:58] sphery: thanks for looking, though.
[20:23:46] stuartm: there might be a way to forward those comments via one of their hooks, Beirdo would know what actually triggers a hook, whether it's just commits or comments on commits too
[20:25:16] sphery: yeah, I'm thinking we need to get pull requests forwarded to commits list, too (like trac messages are), since it seems that users aren't creating tickets for their pull requests
[20:26:08] sphery: lists some methods for getting comments on commit in the repo or on a particular commit, but I'm guessing there's a notification, too?
[20:28:45] stuartm: I'd still prefer to encourage users to open tickets since there is a permanent record of those in trac and in the one place
[20:29:04] stuartm: but we could look at pull requests automatically opening a ticket on trac
[20:29:26] sphery: yeah, agreed that users opening or updating a ticket in Trac is best
[20:30:01] sphery: auto-opening a ticket would make triage harder, too, I'm guessing (since there's nothing in the requests, typically, referencing existing tickets and such, so...)
[20:30:08] k-man: why not use the github ticket system?
[20:30:09] sphery: but would at least keep a record
[20:30:31] sphery: too bad we can't just take the social out of github's "Social coding" ;)
[20:32:15] ToadP (ToadP!~jboyle@ has joined #mythtv
[20:34:49] jams: wagnerrp- i happen to have suse 12.1 here. what were you looking for?
[20:36:00] wagnerrp: when does the network configuration happen?
[20:36:25] wagnerrp: you know, now that i think about it, thats not it
[20:36:29] jams: give me a few minutes and i will take a look
[20:36:37] wagnerrp: because he already connected to the master backend, and the database
[20:36:45] wagnerrp: so the network had to be accessible
[20:37:04] stuartm: k-man: we like trac
[20:37:17] k-man: fair enough :)
[20:37:50] wagnerrp: stuartm: i thought it was more "we dislike the alternatives more"
[20:37:51] wagnerrp: :P
[20:37:54] stuartm: and we have 300 existing tickets in trac that we can't copy across, plus thousands of old tickets that still represent a useful archive we'd want to reference
[20:38:07] stuartm: wagnerrp: some of this, some of that ;)
[20:40:32] stuartm: if we switched to github's ticket system then we'd be stuck with github (not that great a prospect given their downtime issues), at least with trac we have options open to us – there are plenty of migration tools for trac to other systems and if nothing else we have the DB so we can write our own
[20:40:53] stuartm: so a few reasons there ;)
[20:43:23] Beirdo: *some* of us like trac. Those of us that don't... can't find a viable alternative, so are living with it :)
[20:43:59] Beirdo: I think most systems will have a sticking point that annoys at least one of us :)
[20:44:16] MythBuild: build #873 of master-vista-mingw-32bit is complete: Success [build successful] Build details are at . . . t/builds/873
[20:47:21] stuartm: I don't know anyone who has they don't like trac, but there are some systems they might like better
[20:47:34] stuartm: s/has they/has said they/
[20:47:47] ** Beirdo raises his hand **
[20:47:49] stuartm: everyone is comfortable with trac
[20:47:53] stuartm: except Beirdo
[20:47:56] stuartm: ;)
[20:48:00] Beirdo: but I can't think of a better alternative for the group
[20:48:10] Beirdo: so... trac it is :)
[20:48:37] Beirdo: every ticketing system will have sillinesses to them
[20:48:45] Beirdo: trac is pretty darn good though.
[20:48:54] stuartm: all the alternatives considered so far are more complicated, more features but more buttons and a less simple workflow
[20:49:01] Beirdo: yeah
[20:49:23] Beirdo: for some cases that is good, but on the whole... trac fits our needs fairly well
[20:49:49] MythBuild: build #2054 of master-linux-ppc is complete: Success [build successful] Build details are at . . . /builds/2054
[20:50:01] Beirdo: we could use ... bugzilla... and spend half our time messing with it
[20:50:03] Beirdo: hehe
[20:50:12] stuartm: I don't, for example, think anyone needs to be able to see pretty graphs of ticket activity
[20:50:28] stuartm: sounds good, but largely pointless
[20:50:37] Beirdo: they are nice, but not really needed, yeah
[20:51:13] stuartm: sort of thing designed for project managers for whom stats are more important than real results
[20:51:47] Beirdo: so, until someone finds something that we can agree is an improvement, I think trac is our tool for the near to mid future at least
[20:52:22] MythBuild: build #1805 of master-debian-stable-64bit is complete: Success [build successful] Build details are at . . . /builds/1805
[20:53:43] Beirdo: and a prereq to being a replacement would be to be able to port in all of the old ticket history
[20:53:57] Beirdo: which really cuts the list down significantly
[20:55:41] stuartm: a firewall/intrusion detection app decompressing a file I can understand, but sending it on de-compressed is stranger
[20:56:21] Beirdo: yeah, I;m thinking squid
[20:56:52] Beirdo: but even then, I don't think it decompresses it
[20:57:12] wagnerrp: stuartm: those graphs would be useful if we were using git "properly", with all sorts of development going on in separate branches and forks
[20:57:25] wagnerrp: they would help make sense of where what stuff exists
[20:57:38] wagnerrp: but could be done just as well (and probably better) using gitk locally
[21:00:12] stuartm: I'd rather people spent their time closing tickets and fixing bugs than going gooey over visual aides, but I'm a hard task master ;)
[21:00:21] dblain: Beirdo: You have to remember I'm a Microsoft user :/ My firewall is ISA server. I suspect it's not re-compressing it after inspection, but I haven't confirmed it. It could be that the initial request isn't honoring the gzip request.
[21:00:23] jams: wagnerrp- 12.1 uses systemd, so can' ttell you exactly when it starts
[21:00:42] Beirdo: ISA? ahhh
[21:01:00] Beirdo: just trying to think of what would be doing that to ya :)
[21:01:34] Beirdo: it could be stripping the gzip accept-encoding header out on the way out
[21:02:05] stuartm: how rude
[21:02:14] Beirdo: wagnerrp: cgit shows the branches like gitk :)
[21:02:35] Beirdo: dblain: looks good to me, we'll see what others think.
[21:02:36] dblain: Beirdo: Quite honestly, it doesn't bother me either way. I've always just patched locally, but since you got rid of wget, it makes the fix easy enough to request a perminate fix.
[21:02:46] Beirdo: aye
[21:02:48] wagnerrp: Beirdo: im talking more about the limitations (slowness) of a web UI, as opposed to a native application
[21:02:53] sphery: dblain: pretty sure ISA Server blocks the gzip encoding headers outright
[21:03:05] ** jams prefers cgit over gitweb **
[21:03:07] wagnerrp: not specifically the crappy horizontal scrolling in github's network view
[21:03:26] Beirdo: jams: we have cgit on the new server setup :)
[21:03:33] stuartm: run git run
[21:03:34] jams: nice
[21:03:35] dblain: sphery: could be... there is most likey a setting some place where I can turn it on. Just haven't been bothered enough by it to look.
[21:03:42] Beirdo: the network view is pretty, but clunky
[21:03:55] jams: it's also much faster the chiliprojects built in git browser
[21:04:39] Beirdo: yeah, there's a 1–2min delay to see new commits due to caching, but cgit is quite nice
[21:04:57] jams: use cgit for real adhoc lookups, and chiliprojects plugin/browser for linking&closing tickets
[21:05:11] stuartm: it's nice that ISA doesn't mind wasting other people's bandwidth :/
[21:05:13] Beirdo: that reminds me, I want to get it running with nginx. Still something to look into in the next few days
[21:06:05] sphery: stuartm: but, security! can't let some virus through just because it's gzip'ed, then let the browser gunzip it...
[21:06:32] sphery: (If you say "security", it's amazing what you can get away with...)
[21:07:18] stuartm: I suppose it just blocks https outright
[21:07:40] Beirdo: sphery: witness the TSA
[21:08:04] sphery: dblain: looks like you might be able to enable it with . . . tm#220016113 – (but then again, I don't know ISA--just remembered that it doesn't like gzip--so this is just a random result of a search :)
[21:08:29] sphery: that said, we /should/ be detecting whether the result of decompression is useful--this has caught other users in the past, too
[21:08:42] Beirdo: heh, we do
[21:08:55] Beirdo: it's 0 bytes if not decompressed properly
[21:09:09] Beirdo: OK, we do indirectly :)
[21:09:10] sphery: well, detecting and trying to do something about it :)... meaning do what he said
[21:09:19] Beirdo: yeah
[21:09:29] Beirdo: seems like a reasonable tactic
[21:09:53] dblain: sphery: thanks, I'll take a look
[21:10:16] stuartm: aye, 'security' is the sudo of the real world – "Goverment official: strip naked! – Citizen: No! – G-Man: It's for your safety – Citizen: Oh ok then"
[21:10:19] dblain: plus the gzip header doesn't have to be honored by the server (per spec)
[21:10:37] Beirdo: G-Man: <<wolf whistle>> thanks, babe!
[21:11:14] Beirdo: dblain: true, but in the case of TMS, it does, but we shouldn't depend on that in case they change it
[21:18:58] dblain: true
[21:24:35] noahric (noahric!~noahric@ has joined #mythtv
[21:24:46] jya_ (jya_!~jyavenard@mythtv/developer/jya) has quit (Quit: jya_)
[21:25:41] jya_ (jya_!~jyavenard@mythtv/developer/jya) has joined #mythtv
[21:31:29] wagnerrp: kormoc: theres also the minor issue of 'beta' being lower than 'pre'
[21:31:34] wagnerrp: whoops... :)
[21:34:32] zombor (zombor!~zombor_@kohana/developer/zombor) has quit (Remote host closed the connection)
[21:36:45] karl_k (karl_k!~karl@gateway/tor-sasl/karlk/x-26642041) has joined #mythtv
[21:37:20] karl_k: hi, i have a problem with dvb errors: anyone have time to look ?
[21:39:21] Beirdo: gah
[21:40:04] Beirdo: wagnerrp: you have a suggestion on better name for pre for 0.26 cycle?
[21:40:23] Beirdo: beta, rc1, rc2, are in the correct order already
[21:40:43] Beirdo: I'd hate to call it alpha all the way through the cycle
[21:40:50] wagnerrp: in gentoo, its _alpha, _beta, _pre, _rc, <nothing>, _p
[21:41:06] wagnerrp: so it considers the _beta a downgrade from the former pres
[21:41:11] Beirdo: what's the _p?
[21:41:16] wagnerrp: post, i suppose
[21:41:20] Beirdo: ah
[21:41:47] wagnerrp: theres no reason we cant just change the ebuild naming scheme in gentoo to suit
[21:41:54] Beirdo: how unfortunate.
[21:41:56] wagnerrp: and keep mythtv's own tagging in its current format
[21:42:15] Beirdo: K. Well, just looking to smooth things over time. Whatever works
[21:42:25] dekarl: stuartm: wrt #10408 should we simply filter out porn results as a workaround?
[21:42:28] karl_k (karl_k!~karl@gateway/tor-sasl/karlk/x-26642041) has quit (Quit: Verlassend)
[21:42:32] wagnerrp: arbitrarily call _pre _alpha instead
[21:42:46] wagnerrp: and then go from _beta to _rc, and skip _pre entirely
[21:43:20] wagnerrp: or, there might be a way to rejigger the ordering in portage, ill have to look
[21:43:31] jya_ (jya_!~jyavenard@mythtv/developer/jya) has quit (Quit: jya_)
[21:43:53] danielk22: taylorr: Can I reassign #10007 to you? I think dts based on an env variable is a reasonable solution for now.
[21:44:22] taylorr: danielk22: sure, is it OK to commit now... should be harmless
[21:44:36] taylorr: also might be worth backporting
[21:44:48] danielk22: yep
[21:45:12] taylorr: danielk22: ok, I'll get something together soon I hope
[21:46:07] taylorr: danielk22: does FORCE_DTS_TIMESTAMP sound reasonable? if you have a preference then I'll use it
[21:46:37] stuartm: dekarl: huh? What does porn have to do with that ticket?
[21:47:11] wagnerrp: stuartm: using the TTVDB ID to pull artwork from TMDB resulted in porn
[21:47:32] stuartm: not in that ticket ...
[21:47:47] danielk22: taylorr: no preference, whatever it is people will have to discover it though google..
[21:47:48] dekarl: correct
[21:47:53] wagnerrp: perhaps not there, but it is an issue raised (and experienced) in the past
[21:48:22] dekarl: but if the answer is "fix past 0.25" it might be nice to not surprise significant others
[21:48:45] stuartm: ah, well yeah, might be a good idea then
[21:49:34] wagnerrp: dekarl: it would be a fairly simple fix to filter 'adult' content out of tmdb for now
[21:50:03] stuartm: I think the best fix should have been implemented months/years ago when the possibility for confusion was first raised and that is prepending the inetref with a unique identifier for each source e.g. tvdb_12345
[21:51:00] sphery: danielk22: Regarding the db schema upgrade fix... I now know how to fix it to work with MySQL 5.5, even with old, pre-UTF-8 columns in the DB. Only reason I haven't pushed the change that requires 0.22-fixes schema version or higher for upgrades is because I want to make sure that we want to do that now, and not right after 0.25 release. I don't think people really should need a whole release cycle of warning, but want to make sure others ...
[21:51:06] sphery: ... agree. Captain_Murdoch seems to be OK with the idea, so checking with you one more time before pushing. (Can either leave ability to upgrade from ancient schemas and fix it to work with MySQL 5.5, or just limit the schemas we recognize for upgrade--and users upgrading from 0.21 and below can select an intermediate version before going to 0.25.)
[21:51:09] stuartm: that was rejected out of hand and now we have a situation where if people change from one grabber to another .e.g. tmdb to omdb their database and artwork needs to be cleaned out first
[21:52:02] stuartm: IMHO that's still doable with a db update (but no schema change) and we should be seriously thinking about doing that instead of whatever iamlindoro has planned
[21:53:49] danielk22: sphery: I agree, we've trimmed the upgrade list in the past, but we haven't in the last few releases so we support upgrading way more releases than we have in the past.
[21:54:16] sphery: OK, so I'll push the change trimming the list to 0.22 schema version and above
[21:54:17] stuartm: fwiw, I also agree
[21:54:42] stuartm: and from this point on we should habitually do it before every new release
[21:55:15] stuartm: or even immediately after a release so that it doesn't happen at the last minute when we're in a freeze
[21:57:00] sphery: and I think it's actually good to do it now, since some of that old upgrade code works differently depending on Qt and MySQL version (i.e. whereas it was tested with certain behaviors in Qt or MySQL, Qt/MySQL have changed their behaviors in ways that affect those upgrades--for example, our "check to see if the DB has invalid data before attempting conversion to UTF-8" no longer works since Qt changed Qt/MySQL drivers to /always/ use prepared ...
[21:57:06] sphery: ... statements--even when the dev doesn't request them--with Qt 4.5+, so we never see the warnings that tell us there's bad data, then we hit the upgrade, start converting, and fail mid way, leaving the DB in a corrupt state).
[21:58:17] sphery: stuartm: Sounds good. I'll stick with 0.22 schema for now (since all upgrades past then should be safe enough), but for 0.26, we can decide how many versions back we want to support
[21:59:01] sphery: And sorry for taking so long with the fix. It was one part being busy with other projects and one part being a bit trigger shy causing me to spend more time on those other projects. :)
[22:06:34] joki (joki! has quit (Ping timeout: 276 seconds)
[22:08:15] Jordack (Jordack! has quit ()
[22:28:42] danielk22: sphery: where does 6200ch.c live these days?
[22:29:35] danielk22: sphery: nm looks like its in the wiki
[22:30:00] sphery: danielk22: it's at and can be downloaded all at once using mythwikiscripts. (makes sense to check edit history before downloading, though)
[22:30:51] stuartm: heh, yeah, download and run it only to find it's executed "rm -rf /"
[22:31:01] sphery: danielk22: I keep meaning to create a new github project for "channel change and firewire scripts"--something like we're doing with MythTV-Themes, so we can add maintainers who actually have/can test with the given equipment--but haven't gotten around to it, yet
[22:31:19] stuartm: the danger of keeping scripts in a publicly editable wiki
[22:31:19] sphery: actually scripts and programs
[22:31:32] joki (joki! has joined #mythtv
[22:33:38] danielk22: Well for my purposes if it has been demoted to the wiki it is no longer supported and I'm not going to try to update it with the data in #10367.
[22:33:40] sphery: though I'm considering other places to host it... seems a separate repo per script could be annoying and since github only seems to allow per-rep permissions...
[22:34:17] sphery: danielk22: yeah, I couldn't tell if that was referring to 6200ch or if we also needed to update the internal firewire channel changer (which, TTBOMK, is the basis of 6200ch)
[22:36:23] danielk22: sphery: I've updated the firewiredevice.cpp list. But there are way to many steps involved to update 6200ch.c when it is in the wiki.
[22:37:45] ** dblain with be afk for ~1 hour and will check back as soon as possible to make sure the commits compile clean on all environments. **
[22:39:30] danielk22: sphery: I had considered pulling 6200ch.c support into something like mythutil, but firewire just doesn't look like it has any future.
[22:39:43] joki (joki! has quit (Ping timeout: 276 seconds)
[22:51:14] stuartm: danielk22: what's the logic behind kMaxUIWaitTime when checking the availability of a recording? If it takes more than 100ms to check the availability we abort playback of the next file in a playlist, even if the file was found to be available and the delay has already happened so there's nothing lost by continuing?
[22:52:33] jya_ (jya_!~jyavenard@ has joined #mythtv
[22:52:36] jya_ (jya_!~jyavenard@ has quit (Changing host)
[22:52:36] jya_ (jya_!~jyavenard@mythtv/developer/jya) has joined #mythtv
[22:53:06] stuartm: this is PBB, ln 4044 onwards
[22:55:32] stuartm: if there is a delay checking availability it's almost certainly because we had to spin up a drive or similar to check the file was there, so when we start playback the same delay shouldn't occur
[22:55:42] danielk22: stuartm: Basically the UI is frozen while we wait for the result. The idea is if we've waited this long and not gotten a response there is something wrong and it's better to just give up than prevent the user from doing anything.
[22:56:40] danielk22: stuartm: but 100 ms may be too short.. I think I wasn't able to get it to take more than 30 ms or so here, but I left it as it was because I thought it might need to be increased at some point. I just didn't know to what..
[22:56:52] stuartm: danielk22: but we have got a response, that's where we're aborting, after getting the response back
[22:57:20] danielk22: really? Then it doesn't make much sense...
[22:57:25] stuartm: so we get a response saying the file is available, but it took more than 100ms to check so we'll abort
[22:57:49] natanojl (natanojl! has joined #mythtv
[22:57:50] danielk22: Probably that was just supposed to prevent the playlist from kicking in long after the user had moved on to something else.
[22:59:52] xavierh_ (xavierh_! has quit (Ping timeout: 276 seconds)
[23:00:18] danielk22: i.e. the playlist is playing some items, we hit an offline item and return to the menu, 2 hours later the server comes back online, we don't want to keep playing from the playlist..
[23:00:34] joki (joki! has joined #mythtv
[23:01:29] j-rod is now known as j-rod|afk
[23:02:43] danielk22: hmmm, honestly looking at that code, i'm not sure why it is there.
[23:03:58] danielk22: Actually, that last theory holds some weight. We clear the playlist if this timeout has been reached and then set playnext to false if the playlist is empty.
[23:04:38] clever (clever!~clever@ has quit (Ping timeout: 260 seconds)
[23:05:17] danielk22: It could easily be 5 seconds without causing any end user confusion. Probably even larger times are ok. But say 10 minutes would violate the least surprise principle.
[23:05:58] danielk22: How long do drives take to spin up?
[23:06:18] stuartm: I'll increase the threshold then, 5 seconds ought to be enough
[23:06:21] danielk22: Maybe 3x that value is a good #?
[23:06:39] stuartm: danielk22: 2–3 seconds sometimes
[23:06:54] stuartm: from a sleep state
[23:06:57] danielk22: So make it 10 seconds.
[23:07:28] danielk22: This isn't freezing up the UI like I first thought; just queueing up the video
[23:07:57] clever (clever!~clever@ has joined #mythtv
[23:09:12] sphery: stuartm: if is not a crash (and, it makes sense that it's not), then the only problem here is that we lack a UI for the schema wizard, which will require a complete refactoring of the schema wizard to change it to work in an event-driven manner with mythui widgets/dialogs, right? If so, there's no way I'm finishing that before 0.25. So, it's probably either wait for the refactor or set up a fake ...
[23:09:18] sphery: ... qApp->exec() loop just to show a popup saying the exact same thing that the log file says--that it's shutting down because it's not allowed to upgrade the database.
[23:10:29] sphery: this is why I requested a log file a month ago--to confirm that it's just shutting down with a log message saying why
[23:11:34] stichnot: danielk22, stuartm: if you get playlists working, can you give me a heads-up if jump points don't properly quit the playlist? That was an old bug that I wasn't able to test with the jump point fix.
[23:12:49] danielk22: wagnerrp: I'm getting a readyRead segfault.. do you need a stacktrace did you already get that far in the debugging?
[23:13:25] stuartm: stichnot: I can try, but testing playlists is long-winded, the only realistic way of doing it is to create a bunch of very short recordings which probably means manually scheduling each one
[23:14:35] stuartm: stichnot: I had no plans to test whether this fix works, because I'd lay money that it will fix the bug in question and testing it myself is difficult
[23:15:42] stichnot: stuartm: I would just create a playlist with 2 items, start playing, then do a jump point. It should go straight to the main menu. The old broken code would go to the next item in the playlist instead.
[23:16:38] stichnot: also, can't you test playlists efficiently by jumping close to the end of the recording?
[23:17:02] jya: noticed something a tad annoying occurring on my mac, and I had it happening on my linux box too. For some reasons, I've been unable to download any themes using the theme manager. Usually, it goes on displaying the progressing bar "downloading…" and is stuck there. There's no way to interrupt that window, so the only way to exit when it occurs is force quit
[23:17:04] stuartm: stichnot: true, I was thinking in terms of testing/reproducing my bug, for yours it's enough to test a playlist of just two items
[23:18:18] taylorr: question, with git can you look at the branching history of just one file. ie : Like the file tree you get with tkcvs for svn or cvs?
[23:18:27] Beirdo: sphery: I don't think adding a qApp->exec() loop would be a good plan at this juncture. That's a feature a:)
[23:18:45] noahric (noahric!~noahric@ has quit (Quit: noahric)
[23:18:57] Beirdo: taylorr: you can do "git log filename"
[23:18:58] stichnot: in any case, I'll test when I get the chance. (at the moment, my testing is limited to the network control interface...)
[23:19:02] Beirdo: if that's what you mean
[23:19:11] stuartm: sphery: I'll get back to you in a second, I just want to commit this change then I can look at the ticket
[23:19:57] sphery: no hurry--and I have to admit that I'm still not positive what the reporter was trying to say, so I may be missing something
[23:21:41] danielk22: wagnerrp: I attached the bt to #10405. I honestly don't see how your commit could have caused that. But we've had issues like that in the past due to not closing the sockets before deleting MythContext.
[23:23:01] jams: sphery- about 10312..i have seen that type of behaviour when opengl doesn't work. not saying thats what the ticket is about, but it could be.
[23:23:03] clever (clever!~clever@ has quit (Ping timeout: 265 seconds)
[23:23:19] natanojl (natanojl! has quit (Ping timeout: 246 seconds)
[23:23:33] joki (joki! has quit (Ping timeout: 248 seconds)
[23:24:09] jya: Beirdo: from a git commit, can you get back what branch or tag you are using ?
[23:24:34] joki (joki! has joined #mythtv
[23:24:48] Beirdo: git describe says what the tag upstream from it is
[23:25:12] Beirdo: you can't really do that for branches as one commit can be on several branches (due to merges)
[23:25:49] jya: Beirdo: but they wouldn't have the same commit number then...
[23:25:51] Beirdo: well, git describe says the *annotated* tag upstream
[23:26:00] Beirdo: sure they will
[23:26:29] jya: ah, ok.. i'm confusing with the cherry-picked one
[23:26:32] Beirdo: if you work on a side branch, and then merge the branch down to master, every commit on that branch is also now on master
[23:26:35] clever (clever!~clever@ has joined #mythtv
[23:26:37] stuartm: there, a 2 year old bug figured out in a couple of minutes (committing the fix took a little longer, but not much more)
[23:27:09] Beirdo: cool ;)
[23:27:41] wagnerrp (wagnerrp!~wagnerrp_@mythtv/developer/wagnerrp) has quit (Remote host closed the connection)
[23:27:46] jya: stuarta: is there a way to try the mac buildbot with a specific compile script? I want to amend the builder so you can still build qt from source from a version of your choosing.
[23:28:35] Beirdo: run a local builbot master, put in our configs and modify?
[23:28:46] Beirdo: that's the simplest way
[23:29:13] jya: Beirdo: yeah, but reproducing the build environent stuarta is using: that is a core 1 duo, on 10.6
[23:29:14] Beirdo: the hooks won't hit it, so you'd have to force builds though
[23:29:20] Beirdo: ooooh
[23:29:33] Beirdo: I getcha
[23:30:02] Beirdo: you could get him to retarget his slave to another master maybe?
[23:30:05] jya: last time I replaced the existing builder script (which really is a drop-in replacement) I got 1000s wtf
[23:30:26] Beirdo: you'd need to coordinate, of course
[23:30:38] stuartm: it's a shame really, but I'm sure there are a few tickets like that one, where the fix is a simple one-liner or less but no-one really bothered to investigate at the time and they've sat open for months or years
[23:30:43] jya: yeah, but he's a busy fella
[23:30:45] taylorr: since AV_NOPTS_VALUE is defined as "#define AV_NOPTS_VALUE INT64_C(0x8000000000000000)" we don't need to do "if (pkt->dts != (int64_t)AV_NOPTS_VALUE)", right?
[23:31:39] ** jya wondering how to read a windows phone store backtrace.. **
[23:31:58] danielk22: taylorr: nope, probably an old typecast.
[23:32:26] taylorr: ok, I'm going to remove them – should make the code a little neater – I'll do in a separate commit
[23:35:33] jya: so it took two months for the windows app hub to provide the ability to download backtraces. and what you get is an excel spreadsheet… wtf?
[23:35:41] Beirdo: that is likely necessary still, taylorr
[23:35:44] wagnerrp (wagnerrp!~wagnerrp_@2001:470:1f11:12f::a27) has joined #mythtv
[23:35:44] wagnerrp (wagnerrp!~wagnerrp_@2001:470:1f11:12f::a27) has quit (Changing host)
[23:35:45] wagnerrp (wagnerrp!~wagnerrp_@mythtv/developer/wagnerrp) has joined #mythtv
[23:35:45] stichnot: ok git experts, can someone explain this?
[23:35:59] Beirdo: you may need to make that 0x800000.... LL
[23:36:10] Beirdo: I think that was to make the compiler shut the hell up
[23:36:41] Beirdo: stichnot: what about it?
[23:36:41] wagnerrp: gah!
[23:37:00] stichnot: ugh, never mind
[23:37:05] wagnerrp: down in the basement, logged into my server, sshed into another machine
[23:37:13] wagnerrp: type 'reboot'
[23:37:15] stichnot: I did "git pull -a" (whatever that does) instead of "git pull --rebase"
[23:37:16] Beirdo: look at it on the network and you'll see what happened :)
[23:37:19] stuartm: INT64_C doesn't guarantee int64_t though? only int_least64_t – for pretty much every platform we target they are equivalent I believe, but it's still a potential gotcha
[23:37:29] wagnerrp: oh wait... im not ssh'd into another machine!
[23:37:52] Beirdo: stuartm: might work, worth trying to remove it, but beware of new warnings if ya do :)
[23:38:17] taylorr: Beirdo: why, it's not necessary to cast something to int64_t is it's already defined and an int64_t
[23:38:34] ** stuartm notes that Beirdo was saying something similar **
[23:38:39] taylorr: the ffmpeg codebase doesn't have the typecasts either
[23:38:54] Beirdo: In theory it shouldn't :)
[23:39:03] Beirdo: just try it, see if gcc whines, basically
[23:39:19] Beirdo: but that would be the reason it was put in in the first place.
[23:39:32] jya: taylorr: but how is the the INT64_C macro defined?
[23:39:33] stuartm: taylorr: it's not defined as int64_t, it's defined as the type available on that platform offering at least 64bit, but it could be int72_t
[23:39:36] taylorr: danielk22: what do you think?
[23:39:41] Beirdo: no issues with silencing the compile warnings (if they happen) in a different way
[23:39:59] jya: Beirdo is right, it should finish with LL if you want to guarantee not using precision
[23:40:10] Beirdo: it all comes down to the definition of INT64_C
[23:40:20] stuartm: google agrees –
[23:41:04] Beirdo: anyways, try it, compile, check for warnings
[23:41:13] Beirdo: if no warnings, you are likely good.
[23:41:37] Beirdo: it was just to keep the compiler from whining, originally
[23:41:40] ** stuartm wonders how many 9bit systems actually exist **
[23:41:46] Beirdo: heh
[23:41:54] Beirdo: or 72bit
[23:42:17] jya: me guessing that doing (int64_t) (big number) would not always result as big_number, as it would have used big_number to an int first
[23:42:32] stuartm: vanishingly small number I'd imagine
[23:43:09] Beirdo: no, It won't downconvert to int as an intermediate AFAIK
[23:43:14] jya: so my guess is the most portable approach is : INT64_C(0x8000000000000000LL)
[23:43:30] Beirdo: but if you don't tell it, it may get confused
[23:43:44] jya: Beirdo: I've seen it occurring on gcc in the audio class, and why I added a few LL here and there
[23:43:56] Beirdo: yeah, it does get confused at times :)
[23:44:05] Beirdo: being explicit often does help
[23:44:32] jya: because while it should have used it as a long long to start with, it actually generated int -> LL
[23:45:00] jya: so for the sake of clarity, and as it's subject to interpretation, INT64_C(0x8000000000000000LL) that is
[23:47:49] xavierh_ (xavierh_! has joined #mythtv
[23:50:24] noahric (noahric!~noahric@ has joined #mythtv
[23:57:25] knightr: +

IRC Logs collected by BeirdoBot.
Please use the above link to report any bugs.