On Sun, March 6, 2011 22:58, Harrie Hoogeveen wrote: > On Sun, Mar 6, 2011 at 6:02 PM, <[email protected]> wrote: > >> On Sun, March 6, 2011 00:23, Harrie Hoogeveen wrote: >> > On Sat, Mar 5, 2011 at 11:29 PM, <[email protected]> wrote: >> > >> >> On Sat, March 5, 2011 22:18, Harrie Hoogeveen wrote: >> >> > Hello, >> >> > >> >> > I have build version 1.4.2-svn, I have attached the ebuild. >> >> > It seems to be working fine, but no change for the video device >> >> selection: >> >> > Starting... >> >> > *** Using vloopback plugin for fs playback, agreed to use palette >> type >> >> 519 >> >> > ( >> >> > UYVY:Y'CbCr (clamped) ). *** >> >> > Successfully loaded 59 Weed filters >> >> > Loading default keymap from >> >> /home/mivz/.lives-dir/default.keymap...done. >> >> > Started jack audio subsystem. >> >> > >> >> > Checking optional >> >> > >> >> >> dependencies:mplayer...detected...convert...detected...composite...detected...sox...detected >> >> > cdda2wav...detected...jackd...detected...pulse audio...NOT >> >> > >> DETECTED...python...detected...dvgrab...detected...xwininfo...detected... >> >> > >> >> > Window manager reports as "compiz"; number of monitors detected: 1 >> >> > Temp directory is /tmp/livestmp/ >> >> > Compiled with jack support, good ! >> >> > Welcome to LiVES version 1.4.2-svn. >> >> > >> >> > But not much seems to have changed for the vloopback settings. I >> still >> >> > just >> >> > see /dev/video8 in the selection list. When closing it does create >> the >> >> new >> >> > vpp_defaults file. >> >> > Also the option dialog still appears on the second workspace. >> >> > >> >> OK, I checked in another possible fix for this. It may be working OK >> now. > > > To bad, it is still the same. The window pops up at the second desktop. I > also noticed how there are not other popup windows in the settings dialog. > Only at the vj effects parameters. they popup on the correct desktop. >
OK. Then it must be due to the window manager doing the placement. These are two examples of one dialog box being opened in front of another dialog box. So obviously the window manager is trying to be too smart and opening the second dialog on the second screen, the idea being not to obscure the first dialog box. I can try to fix this by setting the screen manually in these cases, if that does not work then you will have to play around with your window manager window placement. > >> >> >> > The OSC commands for loop and pingpong are working perfect. >> >> > >> >> > Can I provide you with any debugging information some how? >> >> > >> >> > Regards, >> >> > Harrie >> >> > >> >> >> >> Yes, please edit >> lives/lives-plugins/plugins/playback/video/vloopback.c >> >> and uncomment the line: >> >> >> >> //fprintf(stderr,"got %s\n",devname); >> >> >> >> Then re-install and start from a terminal, see what output you get. >> >> >> >> >> >> Gabriel. >> >> >> >> >> > That seems to give all the devices: >> > LiVES 1.4.2-svn >> > Copyright 2002-2010 Gabriel Finch ([email protected]) and others. >> > LiVES comes with ABSOLUTELY NO WARRANTY >> > This is free software, and you are welcome to redistribute it >> > under certain conditions; see the file COPYING for details. >> > >> > got /dev/video0 >> > got /dev/video2 >> > got /dev/video4 >> > got /dev/video6 >> > got /dev/video8 >> > >> >> Hmm that is strange. Just to make sure you are running the correct >> version >> of the plugin, I bumped the version number. When you restart LiVES you >> should see a message now about the plugin being updated. Also please >> check >> the version number in Advanced settings, it should be 1.0.2 >> >> Something else I would like you to try, in the vloopback. file, add the >> following line: >> >> while (vdevs[i]!=NULL) { >> snprintf(devstr+slen,30000-slen,"%s|",vdevs[i]); >> fprintf(stderr,"devstr is now %s\n",devstr); // <<<<<<<<<<<< new >> slen+=strlen(vdevs[i])+1; >> free(vdevs[i++]); >> } >> free( vdevs ); >> > > Yes, it says 1.0.2. > The output is: > devstr is now /dev/video0| > devstr is now /dev/video0|/dev/video2| > devstr is now /dev/video0|/dev/video2|/dev/video4| > devstr is now /dev/video0|/dev/video2|/dev/video4|/dev/video6| > devstr is now /dev/video0|/dev/video2|/dev/video4|/dev/video6|/dev/video8| > > But also all the devices are there now I see :D ??? > I really hope you changed something else also... > See note at bottom! > > >> > But every time I hit the video dev select list it also gives a >> warning: >> > (lives:23495): GLib-GObject-WARNING **: gsignal.c:3079: signal name >> > `depressed' is invalid for instance `0x7f4c2001a560' >> > >> >> Yes, I see this too. It seems to be a bug in gtk+ or maybe gtk+ and >> compiz. >> >> >> > Another minor thing I got, if I run make with -j12, which will make it >> use >> > many cpu's at the same time during compile it will give an error when >> I >> > execute lives: lives: error while loading shared libraries: >> > libweed_slice.so.0: cannot open shared object file: No such file or >> > directory >> > When I just run make, on one cpu, every thing will be fine. >> > >> >> Not sure what the problem is there - when I run with -j4 it works fine. >> What happens when you type >> ldd /usr/bin/lives-exe | grep libweed >> > > As expected: libweed_slice.so.0 => not found > But when I put -j1 as default make option it is: > libweed_slice.so.0 => /usr/lib/libweed_slice.so.0 (0x00007f9095470000) > And that is right. If I put -j12, the file is not there. > but if I put -j1 it is... > this through the use of the ebuild file and the default make opts in > /etc/make.conf on gentoo. > > NOTE: > So I have noticing some unexpected changes, within 5 min. Which are > definitely not yours. I work from the ebuild file which checks out svn on > every compile. I use the ebuild unpack and merge commands to make changes > in > between. It's like every even time I compile it works and even odd time I > compile it breaks. I just compiled it with -j12 successfully on a even > compile, in the odd compile it broke again, and the one after it worked > again. I feel like a round robin algorithm is pulling my chain. Could it > be > the sourceforge svn service is load balanced and not synced in time? No. > Or > should I look in some caching issues in the portage/ebuild compile chain? > Still, the window placement on the advanced vloopback settings is not > working for sure. The rest is fixed. > > Regards, > Harrie > Well, since lives compiles without errors, it must be finding the lib at compile time. So something must be deleting this library after compilation (and presumably after make install, as that would install the library anyway). So I would check what the ebuild is doing. If the library is actually there in /usr/lib, it may just be that you need to run ldconfig as root after doing the make install (you should do this anyway as part of the ebuild). Regards, Gabriel. ------------------------------------------------------------------------------ What You Don't Know About Data Connectivity CAN Hurt You This paper provides an overview of data connectivity, details its effect on application quality, and explores various alternative solutions. http://p.sf.net/sfu/progress-d2d _______________________________________________ Lives-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/lives-users
