Re: [OE-core] [PATCH 0/3] GStreamer 1.0 recipes
On 06.06.2013 22:18, Saul Wold wrote: On 06/06/2013 10:17 AM, Carlos Rafael Giani wrote: On 06.06.2013 18:53, Saul Wold wrote: On 05/31/2013 12:15 PM, Carlos Rafael Giani wrote: These patches introduce recipes for GStreamer 1.0 . They are copied over from my gstreamer 1.0 layer at https://github.com/dv1/meta-gstreamer1.0 . All files except gstreamer1.0-plugins-package.inc were copied; gst-plugins-package.inc was patched to fix a problem with -meta packages and can be used for both 0.10 and 1.0 now (the only difference is the value of the LIBV variable). So is there an update path from 0.10 to 1.0? Also normally we replace the older version with the newer version by "git mv" in order to preserve history where possible. I am not sure if you can do this with a move and then copy the new versions in. Since the packages are renamed, and I don's see any RPROVIDES/RREPLACES or RCONFLICTS to cover the upgrade path. Replace? Why do you want to replace? 1.0 is designed to be able to coexist with 1.0 in the same rootfs (you cannot use both at the same time in a process; however, this isn't a concern for OE, but for application developers). In essence, simply adding the new files from the patch to recipes-multimedia/gstreamer/ should be sufficient. The old 0.10 ones can be left unchanged. One of the core tenants of OE-Core is to only maintain the latest version of a given recipe and have an upgrade path from existing version to the latest version. What I am not sure about is the programs that rely on gstream if all of them are ready for the 1.0. But back to the orignal question, is the gstreamer 1.0 an upgrade path for 0.10 or a completely new gstreamer? Is there an upgrade path? If so, then the packaging needs to reflect that. Sau! While 1.0 inherited a lot from 0.10, it is not a simple upgrade. It is ABI and API incompatible with 0.10. The featureset is also not 100% the same. This situation is comparable to Qt3 vs. Qt4 vs. Qt5, or Gtk2 vs. Gtk3 etc. This is why it is not a good idea to get rid of the 0.10 recipes - you cannot simply build applications that use 0.10 with 1.0 libraries and expect it to compile. Carlos ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 0/3] GStreamer 1.0 recipes
On Thu, 2013-06-06 at 13:18 -0700, Saul Wold wrote: > But back to the orignal question, is the gstreamer 1.0 an upgrade path > for 0.10 or a completely new gstreamer? Is there an upgrade path? If > so, then the packaging needs to reflect that. I think what you're really asking is "is gstreamer 1.0 ABI-compatible with 0.10?" and the answer to that question is no. Binaries compiled against 0.10 will need recompiling to use 1.0 so RPROVIDES would make no sense, and the two libraries are parallel installable so RCONFLICTS is not necessary either. In any case, using RPROVIDES/RCONFLICTS for libraries is sometimes slightly tricky in the presence of the debian autonamer, and is generally something that ought to be left to those distros that care about it rather than trying to solve the problem in oe-core. But all this is largely orthogonal to the question of whether or not we want to keep both versions in OE-Core. There does exist some precedent for that: we have both gtk2 and gtk3, and we also have multiple versions of (at least) linux, pseudo and gcc already. p. ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 0/3] GStreamer 1.0 recipes
On 06/06/2013 10:17 AM, Carlos Rafael Giani wrote: On 06.06.2013 18:53, Saul Wold wrote: On 05/31/2013 12:15 PM, Carlos Rafael Giani wrote: These patches introduce recipes for GStreamer 1.0 . They are copied over from my gstreamer 1.0 layer at https://github.com/dv1/meta-gstreamer1.0 . All files except gstreamer1.0-plugins-package.inc were copied; gst-plugins-package.inc was patched to fix a problem with -meta packages and can be used for both 0.10 and 1.0 now (the only difference is the value of the LIBV variable). So is there an update path from 0.10 to 1.0? Also normally we replace the older version with the newer version by "git mv" in order to preserve history where possible. I am not sure if you can do this with a move and then copy the new versions in. Since the packages are renamed, and I don's see any RPROVIDES/RREPLACES or RCONFLICTS to cover the upgrade path. Replace? Why do you want to replace? 1.0 is designed to be able to coexist with 1.0 in the same rootfs (you cannot use both at the same time in a process; however, this isn't a concern for OE, but for application developers). In essence, simply adding the new files from the patch to recipes-multimedia/gstreamer/ should be sufficient. The old 0.10 ones can be left unchanged. One of the core tenants of OE-Core is to only maintain the latest version of a given recipe and have an upgrade path from existing version to the latest version. What I am not sure about is the programs that rely on gstream if all of them are ready for the 1.0. But back to the orignal question, is the gstreamer 1.0 an upgrade path for 0.10 or a completely new gstreamer? Is there an upgrade path? If so, then the packaging needs to reflect that. Sau! ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 0/3] GStreamer 1.0 recipes
On 06.06.2013 18:53, Saul Wold wrote: On 05/31/2013 12:15 PM, Carlos Rafael Giani wrote: These patches introduce recipes for GStreamer 1.0 . They are copied over from my gstreamer 1.0 layer at https://github.com/dv1/meta-gstreamer1.0 . All files except gstreamer1.0-plugins-package.inc were copied; gst-plugins-package.inc was patched to fix a problem with -meta packages and can be used for both 0.10 and 1.0 now (the only difference is the value of the LIBV variable). So is there an update path from 0.10 to 1.0? Also normally we replace the older version with the newer version by "git mv" in order to preserve history where possible. I am not sure if you can do this with a move and then copy the new versions in. Since the packages are renamed, and I don's see any RPROVIDES/RREPLACES or RCONFLICTS to cover the upgrade path. Replace? Why do you want to replace? 1.0 is designed to be able to coexist with 1.0 in the same rootfs (you cannot use both at the same time in a process; however, this isn't a concern for OE, but for application developers). In essence, simply adding the new files from the patch to recipes-multimedia/gstreamer/ should be sufficient. The old 0.10 ones can be left unchanged. ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 0/3] GStreamer 1.0 recipes
On 05/31/2013 12:15 PM, Carlos Rafael Giani wrote: These patches introduce recipes for GStreamer 1.0 . They are copied over from my gstreamer 1.0 layer at https://github.com/dv1/meta-gstreamer1.0 . All files except gstreamer1.0-plugins-package.inc were copied; gst-plugins-package.inc was patched to fix a problem with -meta packages and can be used for both 0.10 and 1.0 now (the only difference is the value of the LIBV variable). So is there an update path from 0.10 to 1.0? Also normally we replace the older version with the newer version by "git mv" in order to preserve history where possible. I am not sure if you can do this with a move and then copy the new versions in. Since the packages are renamed, and I don's see any RPROVIDES/RREPLACES or RCONFLICTS to cover the upgrade path. Some open questions: * Is it OK to add _git versions of the recipes to OE core? Yes, it's OK to have both a versioned and _git recipe in OE-Core * The gstreamer1.0-omx recipe sets itself to be machine specific if this is explicitely requested. This is necessary, since gst-omx has machine specific codepaths (currently for the Raspberry Pi). By default, it is not machine specific, and uses the Bellagio OpenMAX implementation that is supported by OE core. Is it OK to include options for machine specific support which can be set by BSP layers? Not sure about this one. Sau! Carlos Rafael Giani (3): gstreamer: fixed -meta package rdepends gstreamer: moved LIBV out of gst-plugins-package.inc file gstreamer: added GStreamer 1.0 recipes .../gstreamer/gst-plugins-package.inc | 28 -- meta/recipes-multimedia/gstreamer/gst-plugins.inc | 1 + .../gstreamer/gstreamer1.0-libav.inc | 34 +++ ...-Disable-yasm-for-libav-when-disable-yasm.patch | 33 +++ .../gstreamer1.0-libav/libav_e500mc.patch | 21 + .../gstreamer/gstreamer1.0-libav_1.0.7.bb | 27 ++ .../gstreamer/gstreamer1.0-libav_git.bb| 27 ++ .../gstreamer/gstreamer1.0-omx.inc | 42 + .../0001-omx-fixed-type-error-in-printf-call.patch | 30 ++ .../gstreamer/gstreamer1.0-omx_1.0.0.bb| 12 +++ .../gstreamer/gstreamer1.0-omx_git.bb | 22 + .../gstreamer/gstreamer1.0-plugins-bad.inc | 105 + .../gstreamer/gstreamer1.0-plugins-bad_1.0.7.bb| 12 +++ .../gstreamer/gstreamer1.0-plugins-bad_git.bb | 26 + .../gstreamer/gstreamer1.0-plugins-base.inc| 42 + .../gstreamer/gstreamer1.0-plugins-base_1.0.7.bb | 11 +++ .../gstreamer/gstreamer1.0-plugins-base_git.bb | 18 .../gstreamer/gstreamer1.0-plugins-good.inc| 55 +++ .../gstreamer/gstreamer1.0-plugins-good_1.0.7.bb | 11 +++ .../gstreamer/gstreamer1.0-plugins-good_git.bb | 17 .../gstreamer/gstreamer1.0-plugins-ugly.inc| 33 +++ .../gstreamer/gstreamer1.0-plugins-ugly_1.0.7.bb | 10 ++ .../gstreamer/gstreamer1.0-plugins-ugly_git.bb | 16 .../gstreamer/gstreamer1.0-plugins.inc | 57 +++ meta/recipes-multimedia/gstreamer/gstreamer1.0.inc | 23 + .../0001-Fix-crash-with-gst-inspect.patch | 28 ++ .../gstreamer/gstreamer1.0_1.0.7.bb| 13 +++ .../gstreamer/gstreamer1.0_git.bb | 17 28 files changed, 765 insertions(+), 6 deletions(-) create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-libav.inc create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-libav/0001-Disable-yasm-for-libav-when-disable-yasm.patch create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-libav/libav_e500mc.patch create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-libav_1.0.7.bb create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-libav_git.bb create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-omx.inc create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-omx/0001-omx-fixed-type-error-in-printf-call.patch create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-omx_1.0.0.bb create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-omx_git.bb create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-plugins-bad.inc create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-plugins-bad_1.0.7.bb create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-plugins-bad_git.bb create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-plugins-base.inc create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-plugins-base_1.0.7.bb create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-plugins-base_git.bb create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-plugins-good.inc create mode 100644 meta/recipes-multimedia/gstreamer/gstreamer1.0-plugins-good_1.