| |
| |
| |
|
Comments:
<0> _ds, hmm i'm afraid i'm gonna give up gxine maintainership in gentoo... the whole seamonkey/firefox/spidermonkey/what'snext confusion is getting my nerves destroyed
<1> :-\
<0> you said that when it links to spidermonkey the nsplugin would be unstable, correct?
<1> The plugin, yes, but that was fixed some time ago.
<0> fixed as in it can still link to spidermonkey and work fine?
<1> Fixed as in the plugin isn't linked to that at all.
<2> q. general question about http streaming and especially the content of the stream at the begining. when a radio is streaming ogg via http, does the http contain some sort of header or directly the ogg data ?
<0> zorglu1, the standard http headers, i suppose
<0> _ds, that would be good then, the main problem for me is now to find a way to maintain the nsplugin working
<2> Flameeyes: but no special mp3/ogg header ? the player takes care itself of parsing the sound properly even if it is in a middle of a song ?
<0> zorglu1, it's up to the container, so i suppose yes
<0> mp3 and ogg containers are designed for streaming too
<2> ok thanks
<1> Ouch. Just checked the plugin and *bang*.
<1> (Or at least *something* caused firefox to crash...)
<1> #0 0x00002b111c68562c in raise () from /lib/libpthread.so.0
<1> #1 0x000000000044ad74 in ?? ()
<1> #2 <signal handler called>
<1> #3 0x00002aaaacb022c0 in NPP_SetWindow (instance=0x245e228, window=0x236f368)
<1> at plugin.c:492
<1> #4 0x00002aaaacb02e36 in Private_SetWindow (instance=0x245e228,
<1> window=0x236f368) at npunix.c:184
<1> That's this line:
<1> (aargh, pasting is playing up)
<1> this->display = ((NPSetWindowCallbackStruct *) window->ws_info)->display;
<1> (got it. Forgot to hold down Shift...)
<1> window->ws_info is NULL.
<0> not good
<1> Hmm. Now it just exits.
<1> "Error: Couldn't find per display information"
<1> ... and what happens with the plugin with that browser?
<0> _ds, don't know, i use kaffeine :P
<3> anyone around for a little xine-lib compilation help?
<0> sigger_, sort of :)
<3> heh, thx. pastebin'ing.
<3> eh, pastebin.com acting funny. here is where the error creeps in libtool-nofpic: link: ERROR: object name conflicts: .libs/libvcd.lax/libcdio.a// mnt/hda1/xine-lib-1.1.1/src/input/vcd/libvcd/../../../../src/input/vcd/libcdio/. libs/libcdio.a
<0> hmm which distribution?
<3> Damn Small Linux
<3> gcc1 with libs
<0> gcc1?
<3> its gcc - the 1 is just a reference to which gcc package
<0> ah
<0> hmm try to regen the whole autotools chain, maybe libtool is acting up
<3> heh, I haven't compiled anything in like 15 years!
<0> just run autogen.sh script
<0> you need autoconf, automake and libtool
<3> oh yeah, I had problems apt-getting autoconf etc.
<0> _ds, can i ask you an off topic question?
<0> vlcshell.cpp:665: error: 'struct NPSetWindowCallbackStruct' has no member named 'display'
<4> can xine play rpza quicktime video?
<0> do you know what that means and how to resolve it?
<0> robert_, rpza?
<4> yes, rpza- did I stutter? :P
<0> robert_, i simply have no idea what rpza is :)
<4> ah
<4> ok
<1> Flameeyes, what?
<0> _ds, that error above
<1> Hmm, odd.
<1> That doesn't match what documentation I can find...
<0> that is building vlc's (yeah i know) against firefox 1.5
<1> ... and I'm sure that I'd see it here :-|
<0> #ifdef MOZ_X11
<0> Display* display;
<0> this seems to be the problem
<0> and now i could ask
<0> how the hell can you run mozilla without x11 ?_?
<1> Use W*nd*ws? ;-)
<0> _ds, it's in a Unix conditional
<0> oh i think i got it, it missed HAVE_MOZILLA_CONFIG_H definition, what an idiotic way to handle includes for them tho
<1> Hmm... which file is that defined in? I have a suspicion that there may be a mismatch between what gxineplugin wants and what's actually used...
<0> _ds, mozilla-config.h
<0> robert_, seems like not (after a bit of googling)
<4> hm, damn
<0> it might work with win32codecs tho
<1> That answer *was* for struct NPSetWindowCallbackStruct, wasn't it...?
<0> _ds, which one?
<1> Good question :-)
<1> gxineplugin works with Firefox 1.0 but not 1.5...
<0> _ds, i suppose then that it works for mozilla but not seamonkey?
<1> Probably. Not tried.
<0> seamonkey uses firefox 1.5 api
<1> Hmm.
<1> So, wherever it's defined in the (exported) Moz or FF headers...
<0> gah don't have the output handy, i only grepped it
<0> i'll wait to see if vlc builds, if it does i'll try gxine with nsplugin enabled on the server i'm working on
<1> gxine/browser-plugin/npapi.h is what'll need to be updated, I should think.
<0> oh it includes the headers?
<1> It... has its own local copy.
<0> maybe it's simpler to use the installed copy when present? although that will make my maintainership pretty difficult :P
<1> It'll make things work with 1.5, and I can test both...
<0> although making vlc build against firefox 1.5 and seamonkey is proving a quite big effort
<0> mainly because they reshuffled stuff around
<0> so i'm basically implementing a third mozilla switch
<0> the first is automatic, if mozilla-config command is found
<0> the second is manual, providing the path to the sdk (used on osx and windows, and with the now obsolete gecko-sdk), the third would be --enable-mozilla=[firefox|seamonkey] and uses pkg-config
<1> Useful elsewhere, hopefully...
<0> you can probably take the pkg-config part out of my patch and have it more or less done :P
<1> I still need to know which file(s) should be #included, though...
<1> No doubt that'll be in the patch too :-)
<0> _ds, uhm no those are the same, it seems
<0> _ds, take a look to this: https://bugs.gentoo.org/show_bug.cgi?id=139071 seems indeed to be a problem related to xine
<1> Hmm... kaffeine-mozilla should have *exactly* the same problem as gxine's plugin :-)
<0> kaffeine-mozilla has quite a lot more problems
<0> i think i also removed it from portage already
<0> not working on 64-bit arches at all
<0> crashing firefox
<0> and so on
<1> X went down in the most interesting way possible :-\
<0> _ds, which one?
<1> 7.0.22
<1> (i.e. current testing)
<0> no i meant which way
<1> Oh. It froze (mostly). The pointer could still be moved.
<0> gah i seen that before
<0> usually it's an application which grabbed the whole input
<1> I've been able to kill it and continue before
<0> on kde it's either amarok or kdesktop doing so
<0> it doesn't answer to ctrl-alt-backspace
<0> so in those cases i take the ibook or change the kvm and just kill those two
<1> Not sure what caused it, though I *think* that it was something between vdr & gxine.
<1> X seemed to be ****ing up CPU time - something was causing intermittent I/O suspension...
<0> _ds, can you open two programs using xv on your system at the same time?
<1> No.
<0> i need to find why
<0> the nvidia driver could, this is probably the main glitch i found with this card
<1> Hmm. packages.debian.org seems to be somewhat broken :-|
<0> s:packages.::; s:.org::
Return to
#xine
or
Go to some related
logs:
#sql
#dns
psl2 wine
abi version mismatch gaim ubuntu
#css
ubuntu sdl a22p
#php
#oe
hamslug
#web
|
|