the whole stuff in vdr-xineliboutput-9999 is paste twice in the ebuild and please call cvs_src_unpack from src_unpack, else cvs checkout runs two times >>> Unpacking source... * Fetching CVS module vdr-xineliboutput into /usr/portage/distfiles/cvs-src ... * Running cvs -q -f -z1 -d ":pserver:anonymous:@xineliboutput.cvs.sourceforge.net:/cvsroot/xineliboutput" login Logging in to :pserver:anonymous@xineliboutput.cvs.sourceforge.net:2401/cvsroot/xineliboutput * Running cvs -q -f -z1 -d ":pserver:anonymous@xineliboutput.cvs.sourceforge.net:/cvsroot/xineliboutput" update -dP vdr-xineliboutput * Copying vdr-xineliboutput from /usr/portage/distfiles/cvs-src ... * CVS module vdr-xineliboutput is now in /var/tmp/portage/media-plugins/vdr-xineliboutput-9999/work >>> Source unpacked in /var/tmp/portage/media-plugins/vdr-xineliboutput-9999/work >>> Preparing source in /var/tmp/portage/media-plugins/vdr-xineliboutput-9999/work/vdr-xineliboutput ... * Fetching CVS module vdr-xineliboutput into /usr/portage/distfiles/cvs-src ... * Running cvs -q -f -z1 -d ":pserver:anonymous:@xineliboutput.cvs.sourceforge.net:/cvsroot/xineliboutput" login Logging in to :pserver:anonymous@xineliboutput.cvs.sourceforge.net:2401/cvsroot/xineliboutput * Running cvs -q -f -z1 -d ":pserver:anonymous@xineliboutput.cvs.sourceforge.net:/cvsroot/xineliboutput" update -dP vdr-xineliboutput * Copying vdr-xineliboutput from /usr/portage/distfiles/cvs-src ... * CVS module vdr-xineliboutput is now in /var/tmp/portage/media-plugins/vdr-xineliboutput-9999/work also, this version of xineliboutput from cvs works with vdr-1.7.9 Reproducible: Always
vdr-1.7.9 was taken from overlay vdr-devel
Created attachment 212329 [details] vdr-xineliboutput-9999.ebuild (In reply to comment #0) > the whole stuff in vdr-xineliboutput-9999 is paste twice in the ebuild I wonder how this could slip through repoman. > please call cvs_src_unpack from src_unpack, else cvs checkout runs two > times Or better just remove cvs_src_unpack from src_prepare. I think the ebuild needs a few additional dies. Just curious what is the reson why "GENTOO_VDR_CONDITIONAL=yes" defined before the eapi? Why is the "die" commented out after the check for use vdr and xine, if it can not build without one of them the ebuild should die. Note: The attached ebuild is completely untested as I don't use xineliboutput.
Created attachment 212356 [details] vdr-xineliboutput-9999.ebuild Wrong usage of ${PN} fixed.
This ebuild was fully my fault. I wrote it to give the cvs version of xineliboutput a try. It wasn't my intention to commit it to the tree. Sry for the noise. g, mueli
Created attachment 212517 [details, diff] vdr-xineliboutput-9999.diff (In reply to comment #4) > This ebuild was fully my fault. I wrote it to give the cvs version of > xineliboutput a try. It wasn't my intention to commit it to the tree. Sry for > the noise. > g, mueli halcy0n already fixed the duplication bug. So I attach a diff for the missing changes. All the MY_* stuff is not needed here. Also the ebuild should use "|| die for all the helper functions". cvs_src_unpack should not be necessary in src_prepare as it is already called in src_unpack from the cvs eclass.
The duplication bug simply "happened". That's something which can happen and which should be detected by repoman. It's strange that he didn't. The MY_P, `|| die` and the cvs_src_unpack issues are bugs which are fully my fault. I really had no intention to bring this ebuild into tree - it was a fast hack and find the way into tree through IRC. I also recommend to mask the ebuild. People running ~ARCH shouldn't automatically emerge it. Only if you know what you are doing it is advisable to use such a snapshot ebuild. g, mueli
(In reply to comment #6) > The duplication bug simply "happened". That's something which can happen and > which should be detected by repoman. It's strange that he didn't. > The MY_P, `|| die` and the cvs_src_unpack issues are bugs which are fully my > fault. I really had no intention to bring this ebuild into tree - it was a fast > hack and find the way into tree through IRC. I also recommend to mask the > ebuild. People running ~ARCH shouldn't automatically emerge it. Only if you > know what you are doing it is advisable to use such a snapshot ebuild. > g, mueli No problem, mistakes can happen to everyone. I just found this gem on the bug tracker and tried to help. I do not even use it myself. The keyword issue is another bug. Live ebuilds should always be masked or have the keywords removed completely.
Created attachment 212716 [details, diff] vdr-xineliboutput-1.0.4_p20091118.diff
Created attachment 212718 [details, diff] vdr-xineliboutput-9999.diff
Somebody who actually uses this ebuild should test the above patches and if they are okay somebody should commit it. I am off now.
thx for your work but don't use the die function with "use <keyword>" if i don't use the keyword, it breaks
(In reply to comment #11) > thx for your work > but don't use the die function with "use <keyword>" > if i don't use the keyword, it breaks No these functions need to die, else if something breaks it is possible that the ebuild silently fails to install important stuff. Replace the following: use fbcon && dobin vdr-fbfe || die "doins failed" with this: if use fbcon ; then dobin vdr-fbfe || die "doins failed" fi and similar for: use X && dobin vdr-sxfe || die "doins failed" If it still fails then there is an error with the build process.
i don't set the useflag for fbcon with the diff for vdr-xineliboutput-9999, it fails ... >>> Install vdr-xineliboutput-9999 into /var/tmp/portage/media-plugins/vdr-xineliboutput-9999/image/ category media-plugins * Installing locales SO_VERSION=1.0.90-cvs * * ERROR: media-plugins/vdr-xineliboutput-9999 failed. * Call stack: * ebuild.sh, line 49: Called src_install * environment, line 3320: Called die * The specific snippet of code: * use fbcon && dobin vdr-fbfe || die "doins failed"; * The die message: * doins failed * * If you need support, post the topmost build error, and the call stack if relevant. * A complete build log is located at '/var/tmp/portage/media-plugins/vdr-xineliboutput-9999/temp/build.log'. * The ebuild environment file is located at '/var/tmp/portage/media-plugins/vdr-xineliboutput-9999/temp/environment'. * with the diff and changing this in the ebuild: ... if use fbcon ; then dobin vdr-fbfe || die "doins failed" fi if use X ; then dobin vdr-sxfe || die "doins failed" fi ... all is good and there is support for vdpau in cvs, but i can't test this, because i don't owning a nvidia-card
also this is missing "VDR_CONFD_FILE=${FILESDIR}/confd-1.0.0_pre6" in the ebuild, to install vdr.xineliboutput in /etc/conf.d
