* Added MythArchive plugin for archiving recorded shows * Menus are now drawn by MythUI using OpenGL. This option can be enabled/disabled in the Appearance settings. * Improved internal DVD player - now supporting menus and other missing features * Added MHEG content implementation (Interactive TV in UK) * Added Hotplug support for removable media in Media Monitor and MythGallery * Added support for the HDHomeRun encoding device * Added support for basic FreeBox recorders * Added support for H.264 (aka MPEG-4 AVC) TS decoding * Added an MPEG1/MPEG2/MPEG4-AVC IP network recorder * Added internal UPnP server support for TV and Music * Added experimental second commercial detector * New socket class for backend communications * OSD image cache which improves channel changing speed * Fixed program transition while Watching LiveTV * Added beginnings of firewire capture support for MacOS * Support for DVB radio channels and guide data collected via EIT for them * Added mouse support in menus, including gestures
http://allen.brooker.gb.net/misc/kitten-0day.jpg
Holy crap... You're kidding me... This release will actually be delayed even further. It was planned for 8am EST in Gentoo however I had a HD crash on my MythTV box and the entire drive needs to be replaced and all the ebuilds are gone that I wrote for 0.20. So, I need a new HD... Need to build it back up and get MythTV on there before I'll be able to commit the ebuild... So sit tight kids. Jakub, enjoy the flood of bugzilla reports since this will probably take a week or so.
Created attachment 96736 [details] media-tv/mythtv/mythtv-0.20.ebuild For anyone that wants to try this... Here is an ebuild i am currently testing on four systems. Notes: * Based on ebuild mythtv 0.19_p10505 * if you set use flag xvmc and VIDEO_CARDS="openchrome" it will assume you use the openchrome X11 driver and use the confugure option "--enable-xvmc-pro" * newer checking for nvidia-driver * removed UTF8 check for mysql, it was causing my build to fail, but not sure if this is really needed? simply uncomment if needs to be there
*** Bug 147214 has been marked as a duplicate of this bug. ***
Hey Karl, do you also have ebuilds for the plugins as well?
(In reply to comment #5) > Hey Karl, do you also have ebuilds for the plugins as well? > Here ya go Tim. There are really no changes, only rename the ebuild. FYI there was also themes checked into portage http://packages.gentoo.org/ebuilds/?mythtv-themes-0.20 and mythtv: http://packages.gentoo.org/ebuilds/?mythtv-0.20_p11154 Doug, I don't really agree with what you did on xvmc-pro because the via Unichrome and UnichromePRO chipsets are different and require different configs (at least in the 0.19 mythtv, and for xine-lib) I suppose the fact that opencrhome driver is not in gentoo yet is a different bug tho.
(In reply to comment #6) It would be nice if the ebuild supported both unichrome and unichrome-pro as per: http://mythtv.org/wiki/index.php/XvMC#Compiling If compiling manually for either nvidia, I810 or unichrome, use: ./configure --enable-xvmc If you are compiling for unichrome-pro, use: ./configure --enable-xvmc --enable-xvmc-pro --enable-xvmc-vld
Created attachment 96836 [details] Error when patch is applied Some feedback, When trying to emerge this ebuild, the following error occurs. (see attachment)
Karl, You can't just add a video_cards USE flag locally to an ebuild. So whether you agree with me or not, it doesn't matter. The fact is that your solution is broken from the eyes of QA. Bas, Which version of patch do you have... 2.5.9 or 2.5.9-r1. I believe there's an issue with the tarball that's causing this.
(In reply to comment #9) > Bas, > > Which version of patch do you have... 2.5.9 or 2.5.9-r1. I believe there's an > issue with the tarball that's causing this. > I have the same issue as Bas, emerge fails with the patch error. My patch version is 2.5.9. The error from emerge is 'can't find file to patch at input line 5,' but manually applying the patch fails as well: patch /var/tmp/portage/mythtv-0.20_p11154/work/mythtv-0.20/libs/libmythtv/programinfo.cpp /var/tmp/portage/mythtv-0.20_p11154/work/mythtv-0.20-rev11154.patch patching file /var/tmp/portage/mythtv-0.20_p11154/work/mythtv-0.20/libs/libmythtv/programinfo.cpp Hunk #1 FAILED at 1663. Hunk #2 FAILED at 1674. 2 out of 2 hunks FAILED -- saving rejects to file /var/tmp/portage/mythtv-0.20_p11154/work/mythtv-0.20/libs/libmythtv/programinfo.cpp.rej
(In reply to comment #9) <snip> > > Which version of patch do you have... 2.5.9 or 2.5.9-r1. I believe there's an > issue with the tarball that's causing this. I had 2.5.9, but after updating to -r1 i am able to patch it, and am now building mythtv. Thanks for the tip.
Got the Error >>> Unpacking source... >>> Unpacking mythtv-0.20.tar.bz2 to /var/tmp/portage/mythtv-0.20_p11163/work >>> Unpacking mythtv-0.20_svn11163.patch.bz2 to /var/tmp/portage/mythtv-0.20_p11163/work * Cannot find $EPATCH_SOURCE! Value for $EPATCH_SOURCE is: * * /var/tmp/portage/mythtv-0.20_p11163/work/mythtv-0.20-rev11163.patch * ( mythtv-0.20-rev11163.patch ) !!! ERROR: media-tv/mythtv-0.20_p11163 failed. when emerging mythtv had to change LINE 105 in mythtv-0.20_p11163.ebuild from epatch "${WORKDIR}"/${PN}-${MY_PV}_rev${PATCHREV}.patch to epatch "${WORKDIR}"/${PN}-${MY_PV}_svn${PATCHREV}.patch to fix that
Emerging worked without a problem for me, and mythtv more or less works (it's still pretty horrible but certainly better than the previous versions).
What does "it's still pretty horrible" mean? I can't imagine any part of mythtv being horrible, and my experience has been quite the opposite.
Basically, setting up mythtv was extremely painful for me with mysql5, non-US region (with tvguide containing funny characters) and problematic or missing tv guide. But that's offtopic for this bug, I will write a mail to the mythtv mailing list.
Petr, did you really need to use mysql-5 for some reason?
Yes, I'm running other projects on the box and those require mysql-5.
David: Horribly poor code and poor coding practices. Terrible code assumptions made which cause MythTV to generally only work on the developer machines and require patches for each version that's released. Poor attention to memory management (try running it under valgrind...). Poor attention to C++ design standards (don't create a 300 line constructor method). Poor API specificiations causing libraries do cross depend, resulting in a brittle build environment. Hacks layered on hacks to fix an issue. Random crashes and segfaults (most people run Myth under a cron task)
Doug: Wow, I didn't realize it was so badly coded. I guess I am luckly that mythtv has been working so well for me so far. I'll have to browse through the code one of these days.
Speaking of bugs... The frontpage at http://www.mythtv.org/ now mentions that a bug got checked in before release, which prevents the main cause of mythtv, the scheduled recording, to function. Some patch is posted. Just to point that out to the ppl working on this ebuild. Still have to defend MythTV from Dougs comments. my current installation is built from live subversion repository somewhere around february. Since that day, I haven't had a single hickup on my mythbox. It had somewhere around 180 days uptime before I rebooted it a week ago to install a new set of fans in the case. Poor code, bad manners, I don't know, it might all be true. But the box works flawlessly and actually delivers what it's there for. I can't think of any alternative that outperforms mythtv, while avoiding to "look under the hood".
Not quite true, the problem was with mythweb, not mythtv itself. I'll stop using this bug as a discussion forum now. ;-)
(In reply to comment #20) > Speaking of bugs... The frontpage at http://www.mythtv.org/ now mentions that a > bug got checked in before release, which prevents the main cause of mythtv, the > scheduled recording, to function. Some patch is posted. Just to point that out > to the ppl working on this ebuild. Right. The 0.20 ebuilds actually use subversion snapshots from the release-0-20-fixes branch. So you're not actually installing the 0.20a release, you're installing a later version.
(In reply to comment #18) > David: Horribly poor code and poor coding practices. Terrible code assumptions > made which cause MythTV to generally only work on the developer machines and > require patches for each version that's released. Poor attention to memory > management (try running it under valgrind...). Poor attention to C++ design > standards (don't create a 300 line constructor method). Poor API > specificiations causing libraries do cross depend, resulting in a brittle build > environment. Hacks layered on hacks to fix an issue. Random crashes and > segfaults (most people run Myth under a cron task) > So in other words, like 90% of open source projects out there. Just kidding. Seriously, though, I'll be the first to acknowlege well designed code, good standards, and best practices, but the lack of these will likely not prevent me from using the product (api notwithstanding). I am, however, seeing a problem with the TvRec function not being compiled with mpeg support on a box with two PVR250's. Ebuild is 0.20_p11163
Aaron: You'll have to provide some more details.
(In reply to comment #23) > I am, however, seeing a problem with the TvRec function not being compiled with > mpeg support on a box with two PVR250's. Ebuild is 0.20_p11163 Make sure you used the "ivtv" USE flag. If you did, post emerge --info and emerge -pv mythtv
0.20 is bumped, so closing bugs. If you have other mythtv issues, create a new bug. Thanks.
