From: IRC <wt...@us...> - 2003-12-25 06:41:24
|
******************************************************************* [03:05] <Misirlou> What do --enable-plugin-builddir and --disable-plugin-builddir do? [03:06] <Misirlou> I don't understand the message given after ./autogen.sh. [03:11] walters (wa...@ve...) left irc: "bbl" [03:17] mathrick|sleep (~mathrick@Zietka-18.a-inter.net) got netsplit. [03:18] mathrick|sleep (~mathrick@Zietka-18.a-inter.net) returned to #gstreamer. [03:22] walters (wa...@ve...) joined #gstreamer. [03:25] root__ (~ro...@cu...) left irc: Read error: 60 (Operation timed out) [03:28] <ds-work> walters: how does volume get linked to osssink? [03:29] <walters> ds-work: gst_element_link_many [03:30] <walters> ds-work: player/rb-player-gst.c:580 [03:32] foser (d0h@131.211.232.225) left irc: "[ I want to believe ]" [03:34] <ds-work> somehow, the pipeline goes from PAUSED to NULL [03:35] <walters> ds-work: eos? [03:35] <ds-work> it never gets to playing [03:35] <walters> oh [03:36] <ds-work> INFO GST_STATES(22713) gstbin.c(748):gst_bin_change_state: [pipeline]: done changing bin's state from READY to PAUSED, now in PAUSED [03:36] <ds-work> INFO GST_STATES(22713) gstelement.c(2552):gst_element_change_state: no state change needed for element pipeline (VOID_PENDING) [03:36] <ds-work> DEBUG GST_STATES(22713) gstelement.c(2373):gst_element_set_state: [bin0] setting state from PAUSED to NULL [03:39] <walters> --- gstelement.c.~1.227.~ 2003-12-23 21:33:15.000000000 -0500 [03:39] <walters> +++ gstelement.c 2003-12-23 21:38:20.000000000 -0500 [03:39] <walters> @@ -1461,6 +1461,8 @@ [03:39] <walters> /* try to create a new one */ [03:39] <walters> /* requesting is a little crazy, we need a template. Let's create one */ [03:39] <walters> if (filtercaps != NULL) { [03:39] <walters> + if (GST_RPAD_CAPS (pad) == NULL) [03:39] <walters> + return NULL; [03:39] <walters> templcaps = gst_caps_intersect (filtercaps, (GstCaps *) GST_RPAD_CAPS (pad)); [03:39] <walters> /* FIXME */ [03:39] <walters> if (templcaps == NULL) [03:39] <walters> does that look OK? [03:39] <walters> it makes the pipeline sorta work anyways. [03:49] <ds-work> ok, I'm starting to see the problem [03:49] <Misirlou> GAR [03:49] <Misirlou> This may be the problem. [03:49] <ds-work> it helps that you've pointed me directly at it about 10 times :) [03:49] <Misirlou> Somehow I had automake1.4 installed. [03:49] Action: Misirlou kicks the fuck out of old versions [03:50] <walters> ds-work: glad to help :) [03:51] <walters> i forget who it was that observed that a good way to get a problem fixed in the linux kernel was to post a very broken patch for it. then people would chime in and say how broken the patch was, and in the course of that figure out a real solution. [03:52] <Misirlou> heh [03:53] <Misirlou> That is always a good plan. [03:55] <ds-work> finding pad in bin0 compatible with volume:src [03:56] <ds-work> bin0 is osssink, no? [03:56] <walters> ds-work: most likely [04:01] Marsupilami23 (~Mar...@15...) joined #gstreamer. [04:05] <Misirlou> It would be nice if ./autogen.sh erred if things are "not ok". [04:06] <ds-work> it usually does [04:07] ChrisHJW (~chr...@p5...) joined #gstreamer. [04:07] <Misirlou> That doesn't sound encouraging. [04:11] ChrisHJW_lol (~ch...@p5...) joined #gstreamer. [04:11] <ds-work> woohoo! [04:12] <ds-work> rb plays not [04:12] <ds-work> hahah [04:12] <walters> ds-work: awesome :) [04:12] <ds-work> now [04:12] <ds-work> get lots of [04:12] <walters> ds-work: what was the bug? [04:12] <ds-work> ** (rhythmbox:17745): WARNING **: couldn't set requested OSS parameters, enjoy the noise :) [04:13] <ds-work> mistake in logic in gst_pad_can_link_filtered() [04:14] <thaytoo> oh, good show :) [04:14] <ds-work> it was saying that the pads weren't compatible if not in READY [04:14] <thaytoo> seems to be a slow day here today [04:14] <thaytoo> various ppl are playing RTCW [04:14] <thaytoo> maybe I'll just go home :) [04:15] <walters> ds-work: so, commit and stuff :) [04:15] <ds-work> I need to come up with a solution for the 'oss parameters' problem soon [04:15] <Misirlou> I have that problem [04:15] <Misirlou> is it gone now? [04:16] <ds-work> no [04:16] <ds-work> now it's worse [04:17] <Misirlou> sweet! [04:19] <ds-work> SHIT! [04:20] <ds-work> vorbisfile ! volume ! osssink [04:20] <Misirlou> sweet! [04:20] <ds-work> negotiation happens on vorbisfile ! volume, and the osssink fixate function is ignored [04:21] <Misirlou> It's nice to know that OSS is still broken. [04:21] <thaytoo> sounds like the same thing as videotestsrc ! colorspace ! ximagesink [04:21] <ds-work> yep [04:22] <ds-work> proxying fixate functions is stupid, but that might be what we need to do [04:29] <thaytoo> it doesn't sound that stupid [04:29] <thaytoo> if you want the proxied pad to behave the same as the real one, it needs to AFAICS [04:32] freyr (~freyr@4.10.235.125) joined #gstreamer. [04:48] <thaytoo> woot! Christmas, here I come [04:48] <thaytoo> see you all in a few days :) [04:48] thaytoo (ja...@ad...) left irc: "away" [04:52] <taaz> oops. missed thaytoo by a minute [04:53] Action: ds-work is now (conditionally) happy with negotiation [04:53] <ds-work> ok, I'm leaving for Indiana for 5 days [04:53] <ds-work> probably won't be online [04:53] <taaz> i'm sure one of those pesky users will break it for you [04:53] <ds-work> later [04:53] <taaz> l8r [04:54] <walters> ds-work: cool, thanks again [04:58] <Misirlou> Ooh, "couldn't set requested OSS parameters" is fixed? [04:58] Miko5881 (~Mik...@68...) left irc: "ChatZilla 0.8.31 [Mozilla rv:1.4/20030928]" [05:01] <Misirlou> Party on, Garth. [05:15] Marsupilami23 (~Mar...@15...) left irc: "I don't wanna grow up, I'm a Toys R Us Kid. There's a million toys at Toys R Us that I can play with!" [05:15] <walters> ds-work: still here? :) [05:16] <taaz> so walters... i suppose you need me to make some 0.7 debs some day ;) [05:17] <Misirlou> That'd be swell. [05:17] <taaz> how long till rb 0.7.x overtakes 0.6.x? [05:18] <walters> taaz: 0.7 will be a development release [05:19] <walters> sweet, segfault in opt. [05:25] aldug (~al...@al...) left irc: Read error: 54 (Connection reset by peer) [05:28] <taaz> huh... so gstreamer-mpeg2dec is the only user of libmpeg2. so i guess upgrading to 0.4.0 only involves coordination with myself ;) [05:28] <taaz> in debian that is... [05:46] <Misirlou> ugh, boring [05:46] <Misirlou> The following packages have unmet dependencies: [05:46] <Misirlou> libspeex-dev: Depends: libspeex1 (= 1.0.rel.3-1) but 1.1-1 is to be installed [05:46] <Misirlou> E: Broken packages [05:51] <Misirlou> I suppose that explains why the Speex plug-in hasn't been buildin'. [05:56] <Misirlou> Where would be the best place to explain that gst-launch-0.7 filesrc location="/home/aew/Audio/FLAC/Rancid/...And Out Come the Wolves/04 - Time Bomb.flac" ! flacdec ! osssink caused my tympanic membrane to rupture? [06:03] <Misirlou> I'm telling you, this perforation is very painful. [06:17] <taaz> walters: know anything about Bug#224978? [06:19] <Misirlou> zuh [06:19] <Misirlou> uh [06:19] <Misirlou> Bug not found [06:19] <Misirlou> There does not seem to be a bug numbered 224978. [06:19] <taaz> debian bug [06:20] <walters> 2.6.0-test9-mm2-mm2-mppe-cap-2 [06:20] <Misirlou> Okay, Mr. Fancypants. :) [06:20] Action: walters chuckles [06:22] <walters> taaz: if you are running gst-register, then i dunno. [06:23] <walters> can anyone else try a pipeline similar to this with the latest cvs? [06:23] <walters> /build/gstreamer-0.7/bin/gst-launch-0.7 -v gnomevfssrc location=/media/music/Evanescence/Fallen/Tourniquet.ogg ! spider ! osssink [06:23] <walters> for me it sits there and allocates memory like mad [06:24] <walters> hm, it's annoying that the docs are checked into cvs. [06:24] <Misirlou> confirmed [06:24] <walters> Misirlou: thanks. [06:24] <Misirlou> er, duplicated [06:24] <Misirlou> weird [06:25] <Misirlou> I was doing that with no knowledge of your similar doing [06:26] <Misirlou> I think it's definitely spider. [06:26] <Misirlou> No, I think it's osssink. [06:26] <Misirlou> I think I have no clue what I'm talking about. [06:40] <taaz> sening to fakesink gets some CRITICALs [06:40] <taaz> sending too [06:41] <Misirlou> NO taaz [06:41] <Misirlou> sending to [06:42] <taaz> i meant what i said [06:42] <Misirlou> i said what i meant [06:43] <walters> i what meant said i [06:44] <Misirlou> I will fight you to the DEATH [06:44] <taaz> Homer no function beer well without. [06:45] <Misirlou> hmm [06:45] <Misirlou> It seems that <http://bugs.gnome.org/showattachment.cgi?attach_id=22534> applies finely, save for this: [06:46] <Misirlou> - $(FLEX_PATH) -P_gst_parse_yy $^ [06:46] <Misirlou> + $(FLEX_PATH) -P_gst_parse_yy $< [06:46] <Misirlou> Wait, that makes no sense out of context. [06:46] Action: Misirlou bashes head on desk, takes cold shower [07:01] Action: Misirlou spams multiple channels [07:02] <Misirlou> Could someone add the PATCH keyword for <http://bugs.gnome.org/show_bug.cgi?id=129600>? I'm pretty sure the patch applied gracefully when he made it, but something seems to have changed since. [07:07] freerock (~fre...@oo...) joined #gstreamer. [07:08] <Misirlou> Hey, freerock. [07:09] <Misirlou> Are you having problems? [07:10] <freerock> i was figure it out though, gstautoplug.h was looking for that file, saw that it changed in 0.7.3 [07:10] <Misirlou> I think something got lost in transmission or something. . . . [07:12] <freerock> ah you mean the move to freedesktop? [07:12] <Misirlou> er, no, not that [07:13] <Misirlou> All I saw of your message (messages?) was "i was figure it out though, gstautoplug.h was looking for that file, saw that it changed in 0.7.3". Presumably there was a sentence or more preceding it. [07:14] <freerock> hehe, ah, yeah, something needed gstautoplug.h, which wasn't in 0.7.3 but I saw that it was in 0.7.2 so I guess the thing I was building wasn't updated yet [07:17] ds-uhv (~ds...@ad...) joined #gstreamer. [08:00] <ds> walters: mmm? [08:03] <walters> ds: just sent a mail to gstreamer-devel... [08:07] <ds-uhv> walters: btw, are you going to be around Columbus for the next few days? [08:07] <walters> yeah, should be [08:07] <walters> you going to be in the area? [08:07] <ds-uhv> Indianapolis [08:08] <walters> ah, ok [08:10] <ds-uhv> interested in meeting up somewhere along I-70? [08:11] <taaz> i bet ds is better at keysignings than i am ;) [08:11] <ds-uhv> heh [08:11] <taaz> i still have my stack of like 150 keys to sign... sigh [08:11] <walters> ds-uhv: well...i'll be busy the 24th and the 25th. the 26th is possibly doable [08:12] <taaz> python bindings are building again [08:12] <taaz> at least on my box [08:12] <taaz> i don't think i can commit to fdo yet though [08:12] Action: taaz looks at thomasvs [08:12] <walters> ds-uhv: what works for you? [08:13] <taaz> how much of the GstStructure should be wrapped? [08:13] <ds-uhv> walters: don't know yet [08:13] Action: taaz no longer knows what this code does ;) [08:21] <taaz> shoudl gst_structure_set_valist use that gst_value_init_and_copy helper function? [08:21] Action: taaz randomly guessing based on reading commit logs... ;) [08:22] <ds-uhv> no [08:22] <walters> ds-uhv: oh, i remembered what i wanted to ask [08:22] <walters> anyone mind a 20-line flood? [08:22] <walters> er, 30 actually [08:22] <Misirlou> Will Freenode let you live? [08:22] <walters> --- gsttypefindelement.c.~1.5.~ 2003-12-22 16:53:38.000000000 -0500 [08:22] <walters> +++ gsttypefindelement.c 2003-12-23 23:17:32.000000000 -0500 [08:22] <walters> @@ -151,7 +151,7 @@ [08:22] <walters> caps_str = gst_caps_to_string (caps); [08:22] <walters> GST_INFO_OBJECT (typefind, "found caps %s", caps_str); [08:22] <walters> g_free (caps_str); [08:23] <walters> - gst_caps_replace (&typefind->caps, caps); [08:23] <walters> + typefind->caps = gst_caps_copy (caps); [08:23] <walters> if (gst_pad_try_set_caps (typefind->src, caps) < GST_PAD_LINK_OK) { [08:23] <walters> gst_element_error (GST_ELEMENT (typefind), "could not set caps on source pad"); [08:23] <walters> } [08:23] <walters> @@ -195,8 +195,8 @@ [08:23] <walters> gst_type_find_element_signals[HAVE_TYPE] = g_signal_new ("have_type", [08:23] <walters> G_TYPE_FROM_CLASS (g_class), G_SIGNAL_RUN_LAST, [08:23] <walters> G_STRUCT_OFFSET (GstTypeFindElementClass, have_type), NULL, NULL, [08:23] <walters> - gst_marshal_VOID__UINT_BOXED, G_TYPE_NONE, 2, [08:23] <walters> - G_TYPE_UINT, gst_caps_get_type()); [08:23] <walters> + gst_marshal_VOID__UINT_POINTER, G_TYPE_NONE, 2, [08:23] <walters> + G_TYPE_UINT, G_TYPE_POINTER); [08:23] <walters> [08:23] <walters> gstelement_class->change_state = GST_DEBUG_FUNCPTR (gst_type_find_element_change_state); [08:23] <walters> } [08:23] <walters> @@ -669,7 +669,8 @@ [08:23] <walters> break; [08:23] <walters> case GST_STATE_PAUSED_TO_READY: [08:23] <walters> stop_typefinding (typefind); [08:23] <walters> - gst_caps_replace (&typefind->caps, NULL); [08:23] <walters> + if (typefind->caps) [08:23] <walters> + gst_caps_free (typefind->caps); [08:23] <walters> break; [08:23] <walters> default: [08:23] <walters> break; [08:23] <walters> ok so [08:23] <walters> two issues here [08:24] <walters> first...caps were getting double-freed [08:24] <walters> and the second...i think that jump-to-0x0 crash is in one of the gstcaps boxed functions [08:24] <walters> ds-uhv: basically anyone using gst_caps_replace needs to be checked, right? [08:27] <ds-uhv> why? [08:27] <ds-uhv> _replace should be ok [08:27] <walters> ds-uhv: because they're not refcounted anymore? [08:27] <walters> ds-uhv: then what's up with the /* FIXME */ ? :) [08:28] <ds-uhv> oh, that comment should be removed [08:28] <ds-uhv> I already fixed it [08:28] <taaz> what gcc opt gets rid of the type-punned spewing? [08:28] <walters> ds-uhv: so the refcounting isn't needed? [08:29] <ds-uhv> no [08:29] <ds-uhv> we don't do refcounting with caps anymore [08:29] <walters> taaz: -Wno-strict-aliasing iirc [08:29] <ds-uhv> there's a distinct owner for every caps [08:29] <ds-uhv> there could be problems with calling gst_caps_replace() with caps you don't own [08:30] <walters> ds-uhv: well, typefindelement.c seems to do it wrong [08:30] <taaz> walters: didn't seem to work... [08:31] <taaz> err... -Wall was after it... hrm [08:31] Action: taaz grumbles at autotools [08:33] <ds-uhv> yeah, _have_type() should have a const GstCaps * prototype [08:33] <ds-uhv> line 154 should have a gst_caps_copy() [08:34] <walters> ds-uhv: that's what my patch does [08:34] <ds-uhv> that should be the only change, i think [08:35] <walters> ds-uhv: ok, well what do you think about the BOXED -> POINTER? [08:35] <walters> ds-uhv: the only value in using _BOXED is that it's (marginally) typesafe, right? [08:36] <ds-uhv> BOXED copies the value, POINTER doesn't [08:36] <walters> ah. [08:36] <ds-uhv> we never really decided how to deal with it [08:37] <walters> so it should be pointer then. [08:38] <walters> i'll fix that. [08:41] <ds-uhv> yeah, probably [08:42] <ds-uhv> it just requires apps to not be stupid [08:42] <taaz> bindings are neato [08:43] <walters> ds-uhv: you see that opt problem? [08:44] <ds-uhv> yeah, I don't understand it [08:45] <walters> possibly related to that, the tag-finding pipeline doesn't give me any tags anymore either [08:47] <taaz> blah! don't you hate it when you edit a generated file then make clean && make and wonder where your changes went? [08:48] <walters> ds-uhv: with 0.7.3: [08:48] <walters> walters@nexus> /build/gstreamer-0.7.3/bin/gst-launch-0.7 -t gnomevfssrc location=/media/music/Evanescence/Fallen/Tourniquet.ogg ! typefind ! spider ! application/x-gst-tags ! fakesink [08:48] <walters> RUNNING pipeline [08:48] <walters> FOUND TAG : element "vorbistag0" [08:48] <walters> title: Tourniquet [08:48] <walters> artist: Evanescence [08:48] <walters> album: Fallen [08:48] <walters> track number: 6 [08:48] <walters> execution ended after 2 iterations (sum 252670000 ns, average 126335000 ns, min 38896000 ns, max 213774000 ns) [08:48] <walters> cvs just iterates until the end [08:49] <walters> interestingly it seems to do so fairly slowly [08:49] <walters> just piping things into fakesink is waaay faster [08:50] <walters> my guess is that vorbistag is examining every single buffer [08:54] <Misirlou> woooohooo [08:54] <Misirlou> aew@duende:~/Audio/FLAC/Rancid/...And Out Come the Wolves$ gst-launch-0.7 -t gnomevfssrc location=04\ -\ Time\ Bomb.flac ! typefind ! spider ! application/x-gst-tags ! fakesink [08:54] <Misirlou> RUNNING pipeline [08:54] <Misirlou> Segmentation fault [08:57] <walters> ah [08:57] <walters> DEBUG GST_ELEMENT_PADS( 3158) gstelement.c(1691):gst_element_link_pads_filtered: no link possible from vorbisfile0 to src_0 [08:57] <walters> whereas 0.7.3 seems to link them ok: [08:58] <walters> DEBUG GST_ELEMENT_PADS( 3159) gstelement.c(1597):gst_element_link_pads_filtered: linked pad vorbistag0:src to pad src_0:sink [08:58] <walters> actually [08:58] <walters> hm. it's using vorbisfile instead of vorbistag [08:58] <Misirlou> I can get it to show tags after removing typefind. [08:58] freerock (~fre...@oo...) left irc: Remote closed the connection [08:59] Action: Misirlou wonders when all of this broke [08:59] <ds-uhv> isn't that obvious? [09:02] <Misirlou> Also, (process:4670): GStreamer-CRITICAL **: file gstcaps.c: line 401 (gst_caps_is_always_compatible): assertion `!gst_caps_is_empty (caps1)' failed [09:02] <Misirlou> seems to be prevalent. [09:02] <walters> ds-uhv: ok, narrowed it down a bit more...seems to be a spider issue. [09:03] <walters> /build/gstreamer-0.7/bin/gst-launch-0.7 -t gnomevfssrc location=/media/music/Evanescence/Fallen/Tourniquet.ogg ! typefind ! oggdemux ! vorbistag ! fakesink [09:03] <walters> that works ok [09:03] <walters> i guess i don't need typefind in there actually. [09:04] <Misirlou> Heh, you get rid of the problem by removing spider and I get rid of it by removing typefind. [09:05] <walters> ds-uhv: could it be that before, the application/x-gst-tags filter was affecting the whole pipeline, but now it isn't? [09:07] <Misirlou> Oh, well the iterations are caused by spider I think; the segfault's caused by typefind. [09:11] swentel (~swentel@D576894C.kabel.telenet.be) joined #gstreamer. [09:13] <walters> Misirlou: i checked in some fixes to typefind a bit ago [09:28] pb_ (~pb@2002:5160:45ef:0:240:95ff:fe30:aa84) joined #gstreamer. [09:38] walters (wa...@ve...) left irc: "z" [10:21] thomasvs (~th...@17...) left irc: Read error: 110 (Connection timed out) [10:26] pb_ (~pb@2002:5160:45ef:0:240:95ff:fe30:aa84) left irc: "brb" [10:35] thomasvs (~th...@po...) joined #gstreamer. [10:37] ds-uhv (~ds...@ad...) left irc: Read error: 110 (Connection timed out) [10:54] <thomasvs> morning [10:54] <swentel> io t [11:00] pb_ (~pb...@ds...) joined #gstreamer. [11:05] <thomasvs> jo swentel [11:05] <thomasvs> hehe, for once a slashdot thread is actually useful to me [11:06] Action: thomasvs is looking for a good laptop bag to use while rollerblading [11:06] <swentel> heh [11:07] <swentel> you don't use your euh 'rugzak' ? [11:07] <swentel> die ge vroeger altijd gebruikte? [11:10] <Misirlou> I only understood "rucksack". :) [11:14] <thomasvs> swentel: yeah, I use it [11:14] <thomasvs> but it's not sturdy enough for falling down :) [11:14] <swentel> hehehe :) [11:14] <thomasvs> I need something that won't scare the shit out of me when I fall [11:15] <swentel> you always fall on your back ? :p [11:20] <thomasvs> nope, I don't fall actually, but I'm pretty sure I will as soon as I have the laptop with me :) [11:24] <swentel> btw, is the chbg.rpm online or didn't you had any time this weekend? [11:25] Nick change: mathrick|sleep -> mathrick [11:25] <mathrick> morning [11:25] <mathrick> and merry Xmas it seems :) [11:26] <thomasvs> swentel: the iso's are online, but still need to unpack stuff and discuss with wiebe where to host it [11:27] <thomasvs> so, I need about half a day to organize it [11:27] <swentel> ah k [11:27] <thomasvs> which could be coming up given it's christmas [11:27] <swentel> oh yeah, there's file missing in my DAD, don't know if it's in your package too [11:27] <thomasvs> mathrick: same to you [11:27] <thomasvs> swentel: what file ? [11:28] Nick change: mathrick -> mathrick|bath [11:28] <swentel> ratings_store.inc.php iirc [11:28] <thomasvs> hm, pretty sure I've added that to the release [11:29] <thomasvs> you're going to have to upgrade anyway, you had a prerel [11:29] <swentel> well, not on my machine :) [11:29] <swentel> yeah, still some things to fix :) [11:29] <thomasvs> hm, looks like I need a reboot [11:29] <thomasvs> suck [11:29] thomasvs (~th...@po...) left irc: "Client exiting" [11:41] <sxpert> fuck, ivorbis is broken :( [11:46] thomasvs (~th...@po...) joined #gstreamer. [11:57] <sxpert> thomasvs: ivorbis is broken [11:57] <sxpert> :( [11:58] <thomasvs> so fix it [11:59] <sxpert> I have several issues with config.h, things that are also in gst/gst.h [12:04] BBB (~rb...@st...) joined #gstreamer. [12:04] swentelke (~swentel@D576894C.kabel.telenet.be) joined #gstreamer. [12:10] <thomasvs> well, unless you log specific bugs you're going to be stuck with them :) [12:18] Nick change: mathrick|bath -> mathrick [12:18] <mathrick> ah, that feels so much better :) [12:20] swentel (~swentel@D576894C.kabel.telenet.be) left irc: Read error: 110 (Connection timed out) [12:25] <mathrick> mathrick@megumi:~/Dev/gstreamer/gst-plugins/gst-libs/ext/ffmpeg$ make install [12:25] <mathrick> Patching ffmpeg if necessary ...done. [12:25] <mathrick> make[1]: Entering directory `/home/mathrick/Dev/gstreamer/gst-plugins/gst-libs/ext/ffmpeg' [12:25] <mathrick> make[1]: Nie nic do roboty w `all-am'. [12:25] <mathrick> make[1]: Leaving directory `/home/mathrick/Dev/gstreamer/gst-plugins/gst-libs/ext/ffmpeg' [12:25] <mathrick> and install-{data,exec}-am rules seem to be empty [12:26] <mathrick> thomasvs? [12:26] <mathrick> s/Nie nic do roboty w/Nothing to be done in/ [12:36] <thomasvs> mathrick: looks like you somehow disabled ffmpeg from the build ? [12:38] Company (~Co...@p5...) joined #gstreamer. [12:39] <BBB> videotestsrc ! ffcolorspace ! ximagesink is broken too now [12:40] <BBB> I fixed that yesterday, and I fixed that three days ago [12:40] <BBB> WHY WHY WHY [12:40] Action: BBB cries [12:41] <Company> you got it wrong :p [12:42] <Company> _every_ time :p [12:42] <BBB> better yet, it works but there's no data passing [12:42] Action: BBB gives up [12:43] <BBB> or it's simply not being displayed [12:45] <Company> does x(v)imagesink check for successful caps nego? [12:45] <BBB> I think so... [12:45] <BBB> it used to [12:45] Action: BBB doesn't know what you guys broke in the caps branch [12:46] <BBB> framerate=(double)0.000000 <- ? [12:46] <Company> me neither [12:46] <Company> i only made stuff compile ;) [12:47] <Company> hey, i couldn't test the new colorspace caps nego code, because nothing worked at that time :p [12:47] <Company> and the old code didn't do renego [12:48] <Company> i'll make the new code work though :) [12:49] <BBB> the old code *did* do renego [12:49] <BBB> I added that to ffcolorspace [12:49] <BBB> at worst, it might have needed one or two bugfixes [12:49] <BBB> it did do renego [12:50] <Company> ok, then i had code from before you fixed that :) [12:50] <Company> the advantage of my code was that it was much shorter [12:51] ain (~ia...@us...) joined #gstreamer. [12:51] Nick change: ain -> iain [12:51] <Company> the disadvantage obviously was that it didn't work ;) [12:51] <BBB> d'oh [12:51] <iain> morning [12:51] <BBB> I'm fine with new code, but it needs to work [12:52] <BBB> it sucks badly if one branch doesn't contain one, but rather 20 new subsystems [12:52] <BBB> please, one at a time [12:52] <BBB> make it work [12:52] <BBB> and move on to the next [12:52] <BBB> that way we might actually end up with something that works up to a certain amount [12:52] <BBB> morning iain [12:52] <Company> yeah, it's ok to revert that stuff :) [12:52] <BBB> hm [12:52] <Company> we need a videotestsink [12:52] <BBB> is GST_PAD_CAPS (pad) always NULL? [12:53] Action: BBB notes that that breaks ffmpeg [12:53] <Company> maybe docs/random/negotiation or how it's called says something about that? [12:54] sxpert_ (~sx...@sx...) joined #gstreamer. [12:57] <BBB> Incoming: video/x-divx, divxversion=(int)5, framerate=(double)0.000000, width=(int)16, height=(int)16 [12:58] Action: BBB thinks he needs to kick ds [12:58] sxpert (~sx...@sx...) left irc: Operation timed out [12:59] <thomasvs> did ds say he was going to be awol for five days ? [13:00] <BBB> AVI caps: video/x-divx, divxversion=(int)5, framerate=(double)25.000000, width=(int)640, height=(int)464 [13:00] <BBB> Incoming: video/x-divx, divxversion=(int)5, framerate=(double)0.000000, width=(int)16, height=(int)16 :? [13:00] <BBB> how on earth is that possible? [13:00] <thomasvs> why did we gtk-doc comment the cothread stuff ? That's not part of the public API is it ? [13:01] <thomasvs> BBB: does your locale use . or , in general for decimal points ? [13:01] <BBB> I only know I'm going to be away for five days too if things stay like this [13:01] <BBB> probably , [13:01] <BBB> but width/height is wrong too [13:01] <thomasvs> that'll make them stay like this for longer :) [13:01] <BBB> I can't fix them [13:01] <thomasvs> BBB: let's just debug these issues [13:01] <thomasvs> BBB: what exactly are you running ? [13:02] <BBB> our greatest hackers continuously only rewrite working subsystems, that annoys the hell out of me [13:02] Action: BBB runs RH90 [13:02] <thomasvs> BBB: heh, well, what can you do ? :) [13:02] <thomasvs> BBB: btw, I have quilt packages and a nice quilt user pdf [13:02] <Company> BBB: i know it annoys you, but those subsystems did not work [13:03] <thomasvs> BBB: I'll send you the pdf and paste the links to quilt itself [13:03] <Company> there are subtle issues in each of those subsystems [13:03] <BBB> yeah, thanks thomas [13:03] <BBB> Company: subtle enough to be worked on long before an API freeze next time please [13:04] Action: BBB notes that gst-rec also broke [13:04] <thomasvs> BBB: http://thomas.apestaart.org/download/fedora/redhat-9-i386-fedora-stable/quilt-0.21-0.fdr.2.rh90/ [13:04] <Company> BBB: you don't know how long ds worked on it beforehand ;) [13:04] <BBB> I think I have the pdf already [13:05] <BBB> it's the suse pdf, right? [13:05] <thomasvs> yeah [13:05] <thomasvs> I didn't know that was online, damnit :) [13:10] <thomasvs> Company: do you see any reason not to remove cothread docs from gtk-doc stuff ? [13:10] <Company> thomasvs: as long as none of the docs are lost, no [13:10] Action: BBB is badly confused [13:10] <Company> thomasvs: so if there's stuff in the .sgml files, make sure it ends up somewhere [13:11] <BBB> ohwell... has it ever been different [13:11] Action: BBB is off [13:11] <BBB> university needs one more visit [13:11] <BBB> bbl! [13:11] BBB (~rb...@st...) left irc: "Client exiting" [13:11] <iain> do you need to check gst_caps_get_size >= 1 before you access it? [13:12] sxpert_ (~sx...@sx...) left irc: Read error: 110 (Connection timed out) [13:12] <Company> iain: when? [13:12] <Company> iain: in link functions you are guaranteed to get a fixed caps with size 1 [13:12] <iain> ok [13:12] <iain> great :) [13:13] <Company> that's one big advantage of the new caps stuff: link functions are only called with fixed caps [13:13] sxpert_ (~sx...@sx...) joined #gstreamer. [13:14] <iain> :D [13:14] Action: iain deletes if (GST_CAPS_FIXED (caps)) [13:15] <Company> we should deprecate all those macros [13:15] <Company> it's called gst_caps_is_fixed now ;) [13:17] <mathrick> thomasvs: yeah, looks so, but i specifically ran ./configure after slurping ffmpeg by hand, and 'make' actually built it. It's only install-* that's empty [13:18] <iain> Company: and what do you replace gst_caps_set with? [13:18] <mathrick> Company: did you do anything more about mms libification after talking with that xine guy? [13:18] <iain> ah, set_simple [13:19] <Company> iain: normally you get the structure and use gst_structure_set [13:19] <Company> mathrick: nope, been busy enough with caps merges. Why? [13:19] <mathrick> Company: b/c I may try doing it [13:20] <mathrick> Company: I presume SDP is the only 'spec' we have? [13:21] <Company> mathrick: yeah [13:22] <Company> mathrick: the biggest issue is that all that code is GPL [13:22] <Company> mathrick: and gstreamer is the only media framework that is LGPL... [13:22] <mathrick> Company: yeah, I think I might try getting in touch with that guy, but he's not on #xine atm [13:22] <mathrick> Company: gonna mail xine-devel probably [13:23] <mathrick> but first, I have to have a look at that SDP docs :) [13:23] <Company> mathrick: if you get in touch there, please CC gst-devel [13:23] <mathrick> Company: k [13:24] <mathrick> Company: do you think it's worthy to nag mplayer team too? [13:25] <iain> #1 0x401ab7b6 in gst_thread_catch (thread=0x8298efc) at gstthread.c:372 [13:26] <iain> whats gst_thread_catch? [13:27] <Company> iain: it makes sure a GstThread isn'T running [13:27] <Company> so you can safely access data that thread owns [13:27] <sxpert_> Company: ivorbis is b0rken [13:28] <iain> Company: so a hang there would mean a deadlock? [13:28] <Company> mathrick: i would try to find a solution with the xine guys first and then expand - too many cooks and so on... [13:28] <Company> iain: probably [13:28] <iain> bollocks, this all used to work (before the caps merge) [13:28] <mathrick> Company: is '(apostrophe) on your keyboard accesed with Shift? :) [13:28] <Company> iain: you could try to find out what the thread that should be catched currently does... [13:28] <mathrick> Company: ok, one thing at a time :) [13:29] <Company> mathrick: yes, shift-# ;) [13:29] <mathrick> Company: whoah, that's almost as insana as Polish Typist layout ;) [13:29] <mathrick> s/insana/insane/ [13:29] Nick change: sxpert_ -> sxpert [13:30] <iain> Company: it was a load pipeline that was paused, did some connecting, and then it hung/crashed when the state was set to PLAYING [13:31] <Company> iain: uh, "nice", you do dynamic replugging :) [13:31] <Company> iain: good luck ;) [13:32] <iain> yes [13:32] <iain> I reconnect things on the new-pad signal [13:32] <iain> well, I connect the new pad on the new-pad signal [13:32] <Company> i don'T think you need to pause the pipeline for that [13:32] <Company> spider doesn't... [13:33] <iain> hmm, I used to have to [13:33] <iain> I'll remove it then [13:33] <Company> sxpert: i don't use ivorbis so that's entirely possible [13:33] <sxpert> :( [13:33] <Company> since sometime in 0.7 spider doesn't anymore [13:33] <sxpert> Company: why wouldn't it ? [13:33] Nick change: TD[zZz] -> TD [13:34] <sxpert> ivorbis is extra-important for arm based pdas such as ipaq and zaurus [13:34] <sxpert> (no FPU there) [13:35] <mathrick> btw, why videotestsrc ! ximagesink window rescales prettily, and ! xvimagesink just does bilinear filtering? Is that ximagesink doing renego on size change? [13:35] <Company> that last line was for iain [13:35] <Company> mathrick: yes [13:36] <Company> mathrick: and since xvideo does that itself (via hardware) we don't do renego [13:37] <iain> Company: sweet, that fixed it more [13:37] <iain> now I just get GStreamer-WARNING **: trying to set non fixed caps on pad intfloat.src_0:sink, not allowed [13:37] <mathrick> Company: you mean, in case of real video feed, you always give it source size, and it scales down if necessary? [13:38] <Company> mathrick: xvimagesink scales to window size by itself [13:38] <Company> mathrick: ximagesink tries renego and failing that just outputs the video as is [13:38] <Company> so you might end up just seing the top left pixels [13:39] <mathrick> that reminds me me... where can I find explanation of what really overlay is? That is, more in-depth one [13:39] <Company> whaddaya mean "overlay"? [13:40] <mathrick> Company: well, we have some {x,fb,whatever}overlay ifaces? These ones [13:40] <Company> those are for embedding inside applications [13:41] <Company> we need interface docs anyway... [13:41] <Company> s/interface// [13:43] <mathrick> Company: yep. but anyway, there's always some "overlay" at one point, isn't it? [13:44] <mathrick> Company: AFAIK, overlay is some hardware assisted video display, but i don't really know much more [13:45] <Company> mathrick: that's where the name comes from, but in the gst case it's not definitely [13:45] <Company> mathrick: in the gst case it is used to negotiate between app and element the space where to put the video [13:46] <Company> and since that works different depending on windowing system you need different overlay interfaces [13:47] <mathrick> Company: ah, ok. But still, I think that some reading about these (hardware) overlays wouldn't hurt either :) [13:47] <Company> reading never hurts :p [13:47] <mathrick> Company: and one thing that popped about mms yesterday - would it be possible to make gnome-vfs mms module? [13:48] <Company> possibly [13:48] <iain> I was just thinking that :) [13:48] <Company> that's why i want it libified ; [13:48] <Company> ;) [13:49] <mathrick> iain: the question is - just how much is mms multimedia optimized, if it's as much as rt(s)p, IMO it wouldn't be feasible [13:49] Action: iain notes its all nice and quiet on the helix front :) [13:49] <mathrick> Company: and yeah, we love libs :) [13:49] <iain> mathrick: dunno... [13:49] <iain> mathrick: it should be possible though [13:49] <Company> well, the early mms interfaces had open(), read() and seek() [13:50] <mathrick> Company: we just hate lib symbols binary incompatibility ;) [13:50] <Company> i think that hasn't changed much, maybe transport selection [13:50] <iain> mathrick: and then, once we have an mms gnome-vfs-thing we can make our own streambox :) [13:50] Action: mathrick kicks ld still not supporting elf symbol namespaces [13:50] <mathrick> iain: yes, that's the plan :) [13:51] <iain> cant wait for wget to implement it :) [13:51] <iain> wget mms://porn.wmv [13:51] <Company> use gst-launch-0.7 mmssrc ! filesink :o [13:51] <iain> Company: yeah yeah ;p [13:52] <Company> you can even use tee ! spider ! xvideosink to watch what you download :p [13:53] <mathrick> and then, it's SDP's dream come true - downloading ASF streamed video in one simple cmdline :) [13:53] <thomasvs> sxpert: if it's extra important, FILE A BUG REPORT :) [13:54] <mathrick> thomasvs: any hope for me other than ./configure and doing everything from scratch? [13:54] <thomasvs> sxpert: you've been complaining about ivorbis for a few weeks now, but it's not going to change unless you do the usual :) [13:55] <mathrick> ./configure is just sooo slooow [13:55] <thomasvs> mathrick: well, personally I want to move gst-ffmpeg to a separate module, but not everyone seems to agree [13:55] <thomasvs> mathrick: you could try convincing the nay-sayers though, so I can go ahead :) [13:55] <Company> i do agree :) [13:55] <thomasvs> yeah, but BBB doesn't [13:55] <sxpert> thomasvs: oki. will see what I can do to correct the bugs. I may be asking questions to help me correct said thing [13:56] <thomasvs> and, I kinda feel that if we do we should rip out ffcolorspace and move it to gst-plugins [13:56] <mathrick> thomasvs: if that frees me from "ffmpeg failed, please rebuild everything", I'm all for it :) [13:57] <thomasvs> mathrick: it would very much do that, yes [13:57] <mathrick> thomasvs: you have my vote :) [13:57] <thomasvs> mathrick: any way, I don't need convincing. Just send out a mail, complain about the ffmpeg stuff breaking the rest of the build, and ask for it to be moved to a separate module :) [13:58] <mathrick> thomasvs: yes, that's an idea :) [13:58] Action: mathrick fires up evo [13:59] Action: thomasvs is studying applied politics these days [14:09] Action: sxpert hacks away at ivorbis [14:21] <TD> thomasvs: heh :) [14:22] <mathrick> hi, TD, mind some wine assistance? :) [14:22] <TD> what's up? [14:22] <mathrick> TD: making it work, i think :) [14:23] <TD> heh, no surprise. what proggie do you want to run? [14:23] <mathrick> TD: i need it for gaming, but so far failed miserably ;) [14:23] <mathrick> TD: GTA VC is top vote [14:23] <TD> oh yes [14:24] <TD> i took a look at that a few weeks ago. got distracted by the installer :) [14:24] <TD> well, you may need to break out your debugger, i have no idea if there is a "recipe" to make that game work or not. iirc winex "supports" it. [14:24] <mathrick> TD: first thing: .debs or from CVS? [14:24] <TD> supports in quotes because i've heard reports that some of the games they claim work, apparently don't [14:24] <TD> oh, CVS [14:25] <TD> debs tend to be out of date. you usually would want a no windows install [14:25] <mathrick> TD: ok, i'll slurp it right away :) [14:25] <TD> if you have windows, though, you might want to install the game there first and set up wine to use that. the GTA installer is a pain in the ass [14:25] <TD> we have problems with the kernel locking the drive and stuff, whereas windows doesn't [14:26] <mathrick> TD: i try to avoid installing windoze, so would rather try no-win first :) [14:26] <TD> ok. you might need two cd drives [14:26] <TD> that's how i did it. no idea if there is another way or not [14:27] <iain> does wine work nicely without windows being installed? [14:27] <iain> seeing as I've never been able to set it up right for a windowless system [14:27] <TD> it often works better in a no windows setup [14:27] <TD> most windows installs are full of crud that can confuse it [14:27] <iain> hmm, maybe I'll try again [14:28] <TD> it's better to start clean and install what you need. the only disadvantage with that approach is that most win32 installers cause Pain [14:28] <TD> iain: i have written a simple program to install IE into a clean no-windows install, if you need such a thing [14:28] <iain> no, I want to play commandos and starcraft and can't be bothered to move my desk around so I can reach my laptop keyboard :) [14:29] <mathrick> TD: if also have a copy of IE <5.5, it'd be great, seems my mom's on-line banking is written for micros~1 java :\ [14:29] <TD> yeah, games. i should get into that area really :) i don't have either of those sorry. i know starcraft has worked at one point [14:29] <iain> which is better then...wine or the transgaming version? [14:29] <TD> mathrick: no idea whether ms java in wine will work, but IE does. i think my scripts need a bit more tweaking to get the browser itself stable though. [14:29] <mathrick> s/if/if you/ [14:30] <TD> iain: well, for games i guess winex is better, but it depends on your ideals :) often it depends on the program. applications almost always work better in wine. games can work better in wine sometimes, but usually work better in winex [14:30] <mathrick> TD: "fun" part is, it doesn't even work in IE 6 ;) [14:30] <TD> mathrick: does it need to be 5.0? [14:30] <TD> wow, that sucks [14:30] <iain> I'll go with winex then [14:30] <TD> well, i guess you could try ie 5, but i wrote the script for ie 6 [14:30] <TD> iain: remember it's proprietary and pay-for software [14:31] <mathrick> TD: and at one point, I installed it for Acrobat Reader 6 [14:31] Action: sxpert copies whatever is in vorbis and puts it in ivorbis, with the exception of using tremor calls [14:31] <mathrick> TD: or it may be also Sun's Java I (de)installed, but anyway, now even IE is unable to access that damn site [14:32] <iain> TD: is it? did they not used to have a free version? [14:32] <TD> iain: you can build from CVS, but apparently that version is often broken and lags behind what they actually sell by quite a long way [14:32] <TD> it's also missing some important features [14:32] <iain> oh right [14:33] <TD> mathrick: not sure. i did a fair bit of work on making suns JVM work on wine over the summer [14:33] Action: TD gets a demo of starcraft to try [14:34] <iain> starcraft runs fine in normal wine...I had it running years ago :) [14:34] <mathrick> iain: and you can drop commandos, it sucks anyway ;P [14:34] <TD> i thought it did. i want to play it anyway :) feel like games at the moment [14:34] <iain> I like it [14:34] <TD> jazz jackrabbit "holiday hare" works great in dosbox :) [14:35] <iain> I just sucked at it [14:41] <TD> ah great, another kernel upgrade [14:41] Action: TD sticks it on ignore [14:42] <iain> TD: is winehq's cvs good or should I stick to tarballs? [14:43] <TD> cvs is ok at the moment [14:43] <TD> releases are just cvs snapshots anyway [14:43] <TD> what do you want to play though? [14:44] <iain> starcraft [14:46] <iain> and if i can get GTA 1 installed that'd rule too [14:47] <thomasvs> GTA1 even works under dosemu and so on [14:50] <TD> yeah [14:50] <TD> for very old games you want a dos emulator [14:50] <TD> jazz jackrabbit doesn't even start on wine, but runs perfectly in dosbox :) [14:50] <iain> I never had any luck with dos emulators :/ [14:50] <sxpert> yay, I have a working ivorbis [14:51] <sxpert> (at least, it compiles [14:51] <sxpert> thomasvs: should I commit ? [14:51] Action: sxpert tries it first, just in case... [14:52] <TD> well, dosbox cvs works ok. eats cpu time though [14:52] <TD> it emulates a full cpu :/ [14:52] <sxpert> TD: bochs style ? [14:52] <TD> i assume so [14:53] <TD> it's the best way, for real dos apps [14:53] <TD> wine contains some basic cpu emu code for windows apps that still make dos calls [14:53] <sxpert> TD: dosemu is nice [14:54] Action: iain checks out dosbox [14:56] <TD> so i've been told. i need to try it sometime [14:59] <Company> starcraft works here [14:59] <Company> with wine + some copied DLLs (i dunno which though, haven't looked in a long time) [15:00] <Company> i didn't install it in wine hough [15:00] <mathrick> TD: what about all that upcoming longhorn and it's api change? Is wine way to go, or will Mono make it? [15:00] <iain> yeah, I'd got starcraft to run years ago [15:00] <mathrick> s/it's/its/ [15:00] <iain> /usr/local/src/wine/dlls/glu32/../../include/winbase.h:1949: undefined reference to `HeapAlloc' [15:00] <iain> :/ [15:00] <Company> how long is that header? :o [15:01] <iain> 1994 lines [15:05] <TD> iain: compiling which file? [15:05] <TD> mathrick: longhorn isn't our problem, all their new apis will probably be .net based by then, ie it's monos job :) [15:06] <TD> besides, that's years away [15:06] <iain> gcc -shared -Wl,-Bsymbolic,-z,defs glu32.spec.o glu.o glu32.dll.dbg.o -o glu32.dll.so -L../../libs/wine -lwine -L/usr/X11R6/lib -lSM -lICE -lXxf86dga -lXxf86vm -lXrandr -lXrender -lXv -lXext -lX11 -lGL -lGLU -L../../libs/port -lwine_port -lm -lc [15:06] <TD> we don't fully implement all the new XP apis [15:06] <iain> glu.o: In function `wine_gluNewTess': [15:06] <TD> hmm [15:06] <TD> glu32 huh. that's a new one. [15:07] Action: TD does a cvs update [15:07] <TD> it builds ok here,. i've never built it in /usr/local/src thouygh [15:08] <TD> building with any strange config options? [15:08] <TD> you could always use a binary i guess. for games not that much changed since the last release [15:09] Action: mathrick wonders what does MMS need 3 length fields for [15:09] <TD> mathrick: we don't even fully implement the XP apis yet :( [15:10] <mathrick> TD: right now, I don't think I need any XP-specific apps, so it's not a problem. Stable implementation of earlier apis is of more importance now :) [15:10] <TD> winamp 5 isn't xp specific, but it uses xp apis anyway if it sees them [15:11] <sxpert> thomasvs: I have restored tremor to a working state. will commit as soon as cleaned up [15:11] <iain> TD: --prefix=/usr --sysconfdir=/etc are the only configure options [15:11] <mathrick> TD: hmm, but you don't have them, so it does not i guess? [15:11] <iain> TD: and I can't use a binary...I've broken RPM [15:11] <TD> iain: not sure. i only use /opt/wine as my prefix. [15:11] <TD> mathrick: we have some. and besides, some are there but not fully implemented, or buggy, or whatever [15:12] <TD> winamp5 is broke because it mixes together an XP technique and a pre-xp technique [15:12] <TD> we support both but not when used at the same time :( [15:12] <TD> works ok with the classic skin though [15:12] <mathrick> where is winamp4 though? [15:12] <iain> there isn't one [15:12] <Company> you don't emulate the bugginess of windows good enough apparently [15:12] <Company> winamp 5 = winamp 2 + winamp 3 [15:13] <mathrick> hehe :) [15:13] <Company> they do it the fibonacci way [15:14] <TD> bah. starcraft installer crashes on startup [15:15] <mathrick> TD: hmm, and wouldnt --winver=win2k help w/ winamp5? [15:15] Action: mathrick notes that --winver option vanished somewhere on the road [15:16] <TD> yes, it's only in the config file now [15:16] <TD> i doubt it, i think it uses the equivalent of dlsym() given that my wine is set to win98 :) [15:17] <mathrick> TD: why not --winver? Setting it in config for one-time try is PITA [15:19] <TD> yes, i know. aj removed --dll too. at some point i'm going to write a wrapper script that puts them all back [15:19] <TD> the problem is that they interfere with command line parsing of the app itself, and also dll separation [15:19] <TD> (apparently) [15:19] <mathrick> TD: i se [15:19] <mathrick> s/se/see/ [15:20] <mathrick> s/se/see/seee/seeee/... ;) [15:22] <TD> hmm, i see the crash but don't understand it [15:22] Action: TD goes to get a cup of tea [15:26] <iain> yup, something is majorly wrong with compiling wine [15:27] <TD> ooh, heap arena b0rkage, lovely [15:27] <TD> iain: could you report it to wine-devel? you don't need to be subscribed [15:27] <iain> TD: you seeing it too now? [15:28] Action: iain will try to remember [15:28] <TD> nope [15:28] <TD> i'm not sure glu is built on my system though [15:29] <TD> i'm not even sure it's needed. perhaps hacking it out of the makefile will work? [15:29] <TD> Makefile.in in dlls/ should do the trick [15:29] pb_ (~pb...@ds...) left irc: "bbl" [15:29] <TD> apps tend to ship with glu32.dll [15:30] <iain> I disabled with --without-gl [15:30] <iain> and then it broke majorly in ddraw [15:30] <iain> with the same problems and more [15:30] <TD> that's wierd indeed [15:31] <Company> don'T use cvs of gstr^H^H^H^Hwine then [15:32] <TD> i guess a tarball release might work better, but i doubt it. binaries are best :) [15:32] <TD> if you like i can simply upload my build [15:32] <TD> it should work ok on your system, if you have glibc 2.3 [15:32] <iain> nah, its no big deal [15:33] <iain> I'll try a binary [15:33] <TD> yeah, i meant upload my binary install [15:34] <iain> no, its fine, I'm not that concerned about it [15:34] <iain> if I get a ps2 tomorrow, I not need games on windows anyway :D [15:35] <TD> heh [15:36] <TD> waaa [15:36] <TD> i switched off exec-shield and guess what [15:36] <TD> now it works [15:36] <TD> i thought we had worked around that ok. i guess not :( [15:41] <iain> "An unexpected error occurred" [15:41] <iain> surely all errors are unexpected [15:42] <TD> heh [15:42] <TD> what says that? [15:42] <iain> just this commandos setup program [15:45] <mathrick> iain: "Document cannot be closed while it's being open" - Corel Photo Paint ;) [15:45] <iain> haahaa [15:50] <mathrick> man, this mms protocol is hilarious [15:50] <mathrick> "The player version name must start with the value "NSPlayer", if any other name of player is sent, the [15:50] <mathrick> server will automatically send a default movie called `Upgrade Your Player', regardless of any [15:50] <mathrick> filename you may request. This is a 15 second movie telling you how to upgrade." [15:50] <thomasvs> Company: did you at some point disable building of wingo cothreads ? [15:50] <thomasvs> Company: if so, can I remove source code and libs/ext/cothreads for now ? [15:54] <Company> someone disabled it, though it wasn't me [15:54] <Company> but yeah, you can remove that [15:55] <Company> iain: the "can not linK" error on videotestsrc ! osssink is expected for some people ;) [16:00] foser (d0...@22...) joined #gstreamer. [16:00] hyriand (~hy...@no...) left irc: Read error: 54 (Connection reset by peer) [16:00] hyriand (~hy...@no...) joined #gstreamer. [16:10] apoc_ (~ap...@dy...) joined #gstreamer. [16:23] <mathrick> mathrick@megumi:~/Dev/wine/wine$ wine [16:23] <mathrick> Segmentation fault [16:23] <mathrick> :( [16:24] <Company> wee, perfect win32 emulation [16:25] <mathrick> yep, too perfect :\ [16:26] apoc__ (~ap...@dy...) left irc: Read error: 110 (Connection timed out) [16:27] <mathrick> TD: any suggestions? [16:29] <Company> note: never advertise the fact that you're a $program developer outside of $program channels [16:29] <Company> merry christmas [16:29] Company (~Co...@p5...) left irc: Remote closed the connection [16:35] foser (d0...@22...) left irc: "[ I want to believe ]" [16:44] pb_ (~pb@2002:5160:45ef:0:240:95ff:fe30:aa84) joined #gstreamer. [16:45] Action: sxpert tried "make distcheck" in gst-plugins... failed somewhere in spider... [16:47] <iain> and wheeee wine-snapshot has same errors :/ [17:01] <thomasvs> urgh [17:01] <thomasvs> can someone run gst-launch sinesrc ! fakesink for me ? [17:02] <thomasvs> tell me if it segfaults :) [17:10] <sxpert> yep, same here [17:10] <sxpert> segfaults [17:11] Action: swentelke home [17:11] <swentelke> happy christmas everyone :) [17:12] <swentelke> and thomasvs, one of these days ssx3 :p [17:12] swentelke (~swentel@D576894C.kabel.telenet.be) left irc: "toedeloe" [17:21] <iain> hmmmm [17:22] <iain> why does my oneton element now get caps with no channels, then caps with 1 channel twice, then caps with 2 channels? [17:26] <mathrick> wtf? I even did ./autogen.sh, and ffmpeg still builds, but install rules are empty :\ [17:27] <mathrick> thomasvs: any ideas what it could be? [17:33] <thomasvs> mathrick: no, no clue, unless someone changes something. it Just Works here. [17:33] <mathrick> huh? [17:33] <mathrick> megumi:/home/mathrick/Dev/gstreamer/gst-plugins# gst-launch-0.7 videotestsrc ! xvimagesink [17:33] <mathrick> [17:33] <mathrick> GStreamer-ERROR **: No default scheduler name - do you have a registry ? [17:33] <mathrick> aborting... [17:33] <thomasvs> mathrick: run gst-register [17:33] <mathrick> i ran gst-register 10 seconds earlier [17:33] <thomasvs> gst-register-0.7 [17:33] <mathrick> -0.7 [17:33] <thomasvs> hm [17:34] <thomasvs> you sure ? [17:34] <mathrick> loaded 84 plugins with 145 features [17:34] <mathrick> but: mathrick@megumi:~/Dev/gstreamer/gst-plugins$ ls /usr/local/lib/gstreamer-0.7/ | wc -l [17:34] <mathrick> 210 [17:34] Action: iain wants someone to explain the new caps stuff to him [17:34] <thomasvs> iain: can you run sinesrc ! fakesink for me ? [17:35] <mathrick> thomasvs: it segfaulted before i re(built/installed) [17:36] <mathrick> thomasvs: I don't get it, it's clean co from cvs, only ffmpeg swallowed by hand [17:37] <mathrick> thomasvs: moreover, i did rm /usr/local/lib/gstreamer-0.7 prior to make install [17:37] <mathrick> thomasvs: I checked - libgstffmpeg.so is there now, it doesn't get registered tho [17:37] <thomasvs> mathrick: are you sure you have the right build-dir options to configure ? [17:38] sxpert (~sx...@sx...) left irc: Read error: 104 (Connection reset by peer) [17:38] <mathrick> thomasvs: ./autogen.sh --disable-plugin-builddir [17:38] <mathrick> thomasvs: I may try my usual way (./autogen.sh && ./configure) [17:39] <mathrick> thomasvs: and yes, everything just installs into usual /usr/local [17:39] <thomasvs> mathrick: ./autogen.sh --disable-plugin-builddir is not the same as ./autogen.sh -- --disable-plugin-builddir [17:40] BBB (~rb...@st...) joined #gstreamer. [17:41] <mathrick> thomasvs: I did it as Company advised, before I've been always doing ./autogen.sh -- --disable-plugin-builddir [17:41] <TD> mathrick: not sure, but it's a new problem, to do with the nptl detection. what distro? [17:41] <mathrick> TD: Debian Sid [17:41] <BBB> thomasvs: will you be working on GStreamer in januari? [17:42] <TD> mathrick: what glibc? do you have nptl? [17:42] <TD> can you get a backtrace? [17:42] <mathrick> TD: ii libc6 2.3.2.ds1-10 GNU C Library: Shared libraries and Timezone [17:43] <mathrick> TD: hmm, I'll try [17:43] <mathrick> TD: Program received signal SIGTRAP, Trace/breakpoint trap. [17:43] <mathrick> [Switching to Thread 16384 (LWP 25199)] [17:43] <mathrick> 0x40000c00 in ?? () from /lib/ld-linux.so.2 [17:43] <mathrick> (gdb) bt [17:43] <mathrick> #0 0x40000c00 in ?? () from /lib/ld-linux.so.2 [17:43] <mathrick> TD: not much :\ [17:44] <BBB> is this when you start it in gdb? [17:44] <BBB> or when you open the core in gdb? [17:44] <mathrick> BBB: start in gdb [17:44] <BBB> odd [17:45] <TD> mathrick: useful huh. try "c" [17:45] <TD> mathrick: i think that's the first break before it starts [17:45] <mathrick> TD: (gdb) c [17:45] <mathrick> Continuing. [17:45] <mathrick> /usr/local/bin/wine-kthread: error while loading shared libraries: libc.so.6: failed to map segment from shared object: Cannot allocate memory [17:45] <mathrick> [17:45] <mathrick> Program exited with code 0177. [17:46] <TD> wow. b0rked or what [17:46] <BBB> wow... [17:46] <TD> heh. try this: [17:46] <BBB> that's the coolest error I've ever seen [17:46] <TD> ldd `which wine-kthread` [17:46] <TD> be amazed at how voodoo wine can be.... [17:46] <mathrick> mathrick@megumi:~/Dev/wine/wine$ ldd `which wine-kthread` [17:46] <mathrick> not a dynamic executable [17:46] <mathrick> huh? [17:47] <TD> oh, outside of the wine directory [17:47] <TD> i guess you have . in path [17:47] <TD> ldd /usr/local/bin/wine-kthread works too [17:47] <mathrick> mathrick@megumi:~/Dev/wine/wine$ echo $PATH [17:47] <mathrick> /usr/local/bin:/usr/bin:/bin:/usr/bin/X11:/usr/games [17:47] <TD> does it work? [17:47] <mathrick> mathrick@megumi:~/Dev/wine/wine$ ldd /usr/local/bin/wine-kthread [17:47] <mathrick> not a dynamic executable [17:48] <TD> so what is it? [17:48] <TD> file /usr/local/bin/wine-kthread [17:48] <mathrick> mathrick@megumi:~/Dev/wine/wine$ file /usr/local/bin/wine-kthread [17:48] <mathrick> /usr/local/bin/wine-kthread: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.2.0, dynamically linked (uses shared libs), not stripped [17:48] <TD> haha [17:48] <TD> well, ok, it breaks on my system too. [17:48] <TD> [mike@littlegreen wine]$ ldd /opt/wine/bin/wine-kthread [17:48] <TD> ldd: /lib/ld-linux.so.2 exited with unknown exit code (139) [17:48] <TD> i dunno what we do, but glibc is like wines mortal enemy :) [17:48] <mathrick> wow, cool, it's Bug Day today [17:49] Action: TD shrugs [17:49] <TD> wine is riddled with hacks to work around glibc or kernel breakage :( 2003 was a bad year for us [17:49] <TD> we're still picking up the pieces [17:49] BBB (~rb...@st...) left irc: "Client exiting" [17:49] <TD> at least starcraft works for me [17:50] <mathrick> yeah, linux is a horror when it comes to binary compatibility [17:50] <TD> it's not so bad. it's more that wine has some very unusual requirements. it's a very unusual program [17:50] <mathrick> TD: so, what can I do to run it? [17:51] <TD> i'm not sure.... try stracing it? if you get a segfault it means ntdll didn't even initialize so the fault is very early on in the program [17:52] <mathrick> TD: wanna get strace log on priv? [17:52] <TD> go for it [17:52] Action: iain makes one-to-n more efficient [17:53] Nick change: mathrick -> mathrick|Xmas [17:53] <mathrick|Xmas> bbl, Xmas party is begging :) [17:54] <TD> later :) [17:55] <thomasvs> bargh [17:55] <thomasvs> I *hate* having htese new caps and stuff failing all over [17:55] <iain> me too [17:56] <iain> but what is worse is I have no idea what is happening with it, nor what is meant to be happening [18:02] Action: iain hopes ds will be around soonish [18:05] <thomasvs> yeah, well [18:05] <thomasvs> AFAIK, ds is away for five days [18:05] <iain> bugger# [18:05] <thomasvs> and if simple stuff like sinesrc ! fakesink starts breaking, and I'm tracing it all the way through and I don't know what he intended to do, it only leaves me scared for what else has gone wrong [18:08] <iain> yup, segfaults here too [18:10] <iain> maybe I'll spend my time over the holidays working on my crossfader widget [18:17] <TD> perhaps you guys should adopt a regressions policy [18:17] <TD> ds seemed to think it'd be stupid, but whatever [18:19] <TD> ie, no commits may cause regressions. needs a test suite though [18:20] <thomasvs> yeah, I know [18:20] <thomasvs> I think I'm going to be working on that, because this is stupid [18:26] <iain> how about a "no commits before the commiter goes on holiday for a week"? [18:26] <iain> policy [18:26] <TD> you could try and peer review of patches [18:27] <TD> only one person can commit, everybody must mail their patches to him [18:33] <thomasvs> that needs enough reviewers [18:33] <thomasvs> which we don't have [18:34] <TD> why? wine makes do with only one, and it gets ~4mb of patches a month [18:34] <TD> ok, he works full time. it's doable though. for many years he didn't [18:35] <thomasvs> right. he works full time. reviewing patches. [18:35] <thomasvs> we don't have that :) [18:39] <TD> heh [18:39] <TD> well, like i said, it worked that way even when he had a day job [18:49] <TD> maybe just for major patches? [18:49] <TD> of course it requires an accepted de facto maintainer [18:51] <iain> that needs reviewers with enough knowledge about the code [18:51] <iain> the other way is to send patches to a list [18:51] <iain> and get people to review them before committal [18:51] <iain> Ximian did/does that with Evolution coming up to a big release [18:53] <iain> my problem isn't that the patches broke, its that no-one really knows what the changes are, and how its all supposed to work [18:53] <TD> yeah [19:07] thomasvs (~th...@po...) left irc: Read error: 113 (No route to host) [19:24] ChrisHJW (~chr...@p5...) left irc: Read error: 110 (Connection timed out) [20:12] thomasvs (~th...@17...) joined #gstreamer. [20:18] walters (wa...@ve...) joined #gstreamer. [21:27] walters (wa...@ve...) left irc: "out" [21:59] Company (~Co...@p5...) joined #gstreamer. [22:00] <Misirlou> why hello [22:00] Nick change: Company -> CompVsVideo [22:00] <Misirlou> an epic battle, no doubt [22:02] Action: CompVsVideo got a TV card for christmas [22:18] aldug (~al...@al...) joined #gstreamer. [22:23] walters (wa...@ve...) joined #gstreamer. [22:43] ChrisHJW (~chr...@p5...) joined #gstreamer. [22:49] Zeenix (~zak@203.135.11.148) joined #gstreamer. [22:58] matrixise (~Stef@213.213.207.180) left irc: Remote closed the connection [23:02] mathrick|Xmas (~mathrick@Zietka-18.a-inter.net) left irc: Read error: 54 (Connection reset by peer) [23:04] mathrick (~mathrick@Zietka-18.a-inter.net) joined #gstreamer. [23:36] Miko5881 (~Mik...@68...) joined #gstreamer. [23:57] Company (~Co...@p5...) joined #gstreamer. [23:57] CompVsVideo (~Co...@p5...) left irc: Read error: 104 (Connection reset by peer) [23:59] <Misirlou> thomasvs: So are you working on a test suite? [00:00] --- Thu Dec 25 2003 [00:01] <walters> Misirlou: there already is a test suite...it's just not extensive enough :) [00:02] <Company> media test suite [00:02] <Company> a testsuite testing actual media files [00:02] <Company> he's suppose... [truncated message content] |