More recipe reporting fixes
cover More recipe reporting fixes
This patchset contains more recipe reporting fixes, some of which are based on the review comments to the previous patchset (which got partially merged into master). Please see specific commits for details.
Version 2 of the patch improves the commit message for stess recipe patch, and adds the use of BPN and BP to the same recipe. It also fixes build failures for fotowall and distcc recipes (incorrectly set S), and xf86-video-omapfb (incorrect commit id).
The following changes since commit 58197f345b203f55de202f8c91129e3d381359ed:
bitbake: toastergui: layerBtn Fix build trigger mechanism (2015-08-06 16:04:50 -0500)
are available in the git repository at:
git://git.yoctoproject.org/poky-contrib akanavin/recipe-reporting-fixes
http://git.yoctoproject.org/cgit.cgi/poky-contrib/log/?h=akanavin/recipe-reporting-fixes
Alexander Kanavin (7):
fotowall: update SRC_URI to use github instead of qt-apps.org
distcc: update SRC_URI to not use Google Code
gptfdisk: update to 1.0.0
oprofile: reorder SRC_URI lines to fix upstream version check
xf86-video-omapfb: switch SRC_URI to debian git repo
stress: update checksums to match upstream tarball, use BPN and BP in
SRC_URI
recipeutils.py: allow all characters in regexes used to parse version
strings
meta/lib/oe/recipeutils.py | 4 +--
meta/recipes-devtools/distcc/distcc_3.1.bb | 7 ++---
.../fdisk/{gptfdisk_git.bb => gptfdisk_1.0.0.bb} | 8 ++---
meta/recipes-extended/stress/stress_1.0.4.bb | 8 ++---
...apfb-from-from-crashing-when-pixelclock-o.patch | 35 ++++++++++++++++++++++
.../xorg-driver/xf86-video-omapfb_git.bb | 5 ++--
meta/recipes-kernel/oprofile/oprofile.inc | 3 +-
meta/recipes-kernel/oprofile/oprofile_1.0.0.bb | 2 --
meta/recipes-qt/qt-apps/fotowall_0.9.bb | 8 ++---
9 files changed, 55 insertions(+), 25 deletions(-)
rename meta/recipes-devtools/fdisk/{gptfdisk_git.bb => gptfdisk_1.0.0.bb} (75%)
create mode 100644 meta/recipes-graphics/xorg-driver/xf86-video-omapfb/0001-Prevents-omapfb-from-from-crashing-when-pixelclock-o.patchSignature · 1 line
2.1.4
PATCH 1/7 fotowall: update SRC_URI to use github instead of qt-apps.org
GitHub is where upstream development is happening, so we are more likely to find out about new releases from there
--- meta/recipes-qt/qt-apps/fotowall_0.9.bb | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-)
meta/recipes-qt/qt-apps/fotowall_0.9.bb +3 -5
@@ -10,14 +10,12 @@ SECTION = "x11/apps" PR = "r4" -SRC_URI = "http://qt-apps.org/CONTENT/content-files/71316-Fotowall-0.9.tar.bz2 \ +SRCREV = "06d5a4142c599604d9a9fd64727b6945eb8cd3da" +SRC_URI = "git://github.com/enricoros/fotowall.git \ file://ExportWizard-depends-on-ui_wizard.patch \ " -SRC_URI[md5sum] = "142ef697332e0777c6d22c5bc96cc438" -SRC_URI[sha256sum] = "e4d0c005d2cb1d7c09438bfc3098eadebc08946e4fbc0655b7fc8b046de3810d" - -S = "${WORKDIR}/Fotowall-${PV}" +S = "${WORKDIR}/git" inherit qt4x11
Signature · 1 line
2.1.4
PATCH 2/7 distcc: update SRC_URI to not use Google Code
In reply to Alexander KanavinGoogle Code is shutting down soon and distcc has moved to github.
--- meta/recipes-devtools/distcc/distcc_3.1.bb | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-)
meta/recipes-devtools/distcc/distcc_3.1.bb +3 -4
@@ -17,16 +17,15 @@ PACKAGECONFIG[popt] = "--without-included-popt,--with-included-popt,popt" RRECOMMENDS_${PN} = "avahi-daemon" -SRC_URI = "http://distcc.googlecode.com/files/${BPN}-${PV}.tar.bz2 \ +SRC_URI = "git://github.com/distcc/distcc.git \ file://separatebuilddir.patch \ file://0001-zeroconf-Include-fcntl.h.patch \ file://default \ file://distccmon-gnome.desktop \ file://distcc \ file://distcc.service" - -SRC_URI[md5sum] = "a1a9d3853df7133669fffec2a9aab9f3" -SRC_URI[sha256sum] = "f55dbafd76bed3ce57e1bbcdab1329227808890d90f4c724fcd2d53f934ddd89" +SRCREV = "9a172d3d20042754a30c0cd9a5d63fd1a8120c31" +S = "${WORKDIR}/git" inherit autotools pkgconfig update-rc.d useradd systemd
Signature · 1 line
2.1.4
PATCH 3/7 gptfdisk: update to 1.0.0
In reply to Alexander KanavinAlso, use tarballs instead of git for fetching source code. git repository does not have version tags, so upstream version check cannot be performed.
---
.../recipes-devtools/fdisk/{gptfdisk_git.bb => gptfdisk_1.0.0.bb} | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
rename meta/recipes-devtools/fdisk/{gptfdisk_git.bb => gptfdisk_1.0.0.bb} (75%)meta/recipes-devtools/fdisk/gptfdisk_1.0.0.bb renamed from meta/recipes-devtools/fdisk/gptfdisk_git.bb +3 -5
@@ -5,11 +5,9 @@ LIC_FILES_CHKSUM = "file://COPYING;md5=59530bdf33659b29e73d4adb9f9f6552" DEPENDS = "util-linux popt ncurses" -PV = "0.8.10+git${SRCPV}" -SRCREV = "a920398fa393f9d6301b32b191bc01e086ab8bc8" -SRC_URI = "git://git.code.sf.net/p/gptfdisk/code" - -S = "${WORKDIR}/git" +SRC_URI = "${SOURCEFORGE_MIRROR}/${BPN}/${PV}/${BP}.tar.gz" +SRC_URI[md5sum] = "2061f917af084215898d4fea04d8388f" +SRC_URI[sha256sum] = "5b66956743a799fc0471cdb032665c1391e82f9c5b3f1d7d726d29fe2ba01d6c" do_install() { install -d ${D}${sbindir}
Signature · 1 line
2.1.4
PATCH 4/7 oprofile: reorder SRC_URI lines to fix upstream version check
In reply to Alexander KanavinThe source tarball should be listed first.
--- meta/recipes-kernel/oprofile/oprofile.inc | 3 ++- meta/recipes-kernel/oprofile/oprofile_1.0.0.bb | 2 -- 2 files changed, 2 insertions(+), 3 deletions(-)
meta/recipes-kernel/oprofile/oprofile.inc +2 -1
@@ -18,7 +18,8 @@ FILES_${PN} = "${bindir} ${libdir}/${BPN}/lib*${SOLIBS} ${datadir}/${BPN}" FILES_${PN}-dev += "${libdir}/${BPN}/lib*${SOLIBSDEV} ${libdir}/${BPN}/lib*.la" FILES_${PN}-staticdev += "${libdir}/${BPN}/lib*.a" -SRC_URI = "file://filemode-fix.patch \ +SRC_URI = "${SOURCEFORGE_MIRROR}/${BPN}/${BPN}-${PV}.tar.gz \ + file://filemode-fix.patch \ file://acinclude.m4 \ file://automake-foreign.patch \ file://oprofile-cross-compile-tests.patch \
meta/recipes-kernel/oprofile/oprofile_1.0.0.bb +0 -2
@@ -3,8 +3,6 @@ require oprofile.inc DEPENDS += "virtual/kernel" DEPENDS_append_powerpc64 = " libpfm4" -SRC_URI += "${SOURCEFORGE_MIRROR}/${BPN}/${BPN}-${PV}.tar.gz" - SRC_URI[md5sum] = "ba0b340e5c421a93959776c836ed35b3" SRC_URI[sha256sum] = "847110b4ecdcf8c8353cd38f94c1b704aad4bfcd9453e38b88d112cfb7e3c45a"
Signature · 1 line
2.1.4
PATCH 5/7 xf86-video-omapfb: switch SRC_URI to debian git repo
In reply to Alexander KanavinThe previous upstream is no longer available. Debian git repository is one commit behind the original repository, so that commit is added as a patch.
--- ...apfb-from-from-crashing-when-pixelclock-o.patch | 35 ++++++++++++++++++++++ .../xorg-driver/xf86-video-omapfb_git.bb | 5 ++-- 2 files changed, 38 insertions(+), 2 deletions(-) create mode 100644 meta/recipes-graphics/xorg-driver/xf86-video-omapfb/0001-Prevents-omapfb-from-from-crashing-when-pixelclock-o.patch
meta/recipes-graphics/xorg-driver/xf86-video-omapfb/0001-Prevents-omapfb-from-from-crashing-when-pixelclock-o.patch new file +35 -0
@@ -0,0 +1,35 @@ +From 28c006c94e57ea71df11ec4fff79d7ffcfc4860f Mon Sep 17 00:00:00 2001 +From: Far McKon <[email protected]> +Date: Tue, 3 May 2011 20:59:04 +0300 +Subject: [PATCH] Prevents omapfb from from crashing when pixelclock of 0 is + sent to it. + +Due to a Linux design bug it is easy to get a pixelclock set to zero +when changing displays at runtime. +--- + src/omapfb-output.c | 9 +++++++-- + 1 file changed, 7 insertions(+), 2 deletions(-) + +diff --git a/src/omapfb-output.c b/src/omapfb-output.c +index f8b4db3..4d59265 100644 +--- a/src/omapfb-output.c ++++ b/src/omapfb-output.c +@@ -125,8 +125,13 @@ OMAPFBOutputGetModes(xf86OutputPtr output) + /* Only populate the native (current) mode */ + mode = calloc(1, sizeof(DisplayModeRec)); + mode->type |= M_T_PREFERRED; +- mode->Clock = PICOS2KHZ(ofb->state_info.pixclock); +- mode->SynthClock = PICOS2KHZ(ofb->state_info.pixclock); ++ if (ofb->state_info.pixclock == 0) { ++ mode->Clock = 0; ++ mode->SynthClock = 0; ++ } else { ++ mode->Clock = PICOS2KHZ(ofb->state_info.pixclock); ++ mode->SynthClock = PICOS2KHZ(ofb->state_info.pixclock); ++ } + mode->HDisplay = ofb->state_info.xres; + mode->HSyncStart = mode->HDisplay + + ofb->state_info.right_margin; +-- +2.1.4 +
meta/recipes-graphics/xorg-driver/xf86-video-omapfb_git.bb +3 -2
@@ -9,11 +9,12 @@ LICENSE = "MIT-X & GPLv2+" LIC_FILES_CHKSUM = "file://COPYING;md5=63e2cbac53863f60e2f43343fb34367f" DEPENDS += "virtual/libx11" -SRCREV = "28c006c94e57ea71df11ec4fff79d7ffcfc4860f" +SRCREV = "33e36c12dde336edbdd34626dd8adfcaebc8fbb8" PR = "${INC_PR}.7" PV = "0.1.1+gitr${SRCPV}" -SRC_URI = "git://git.pingu.fi/xf86-video-omapfb;protocol=http \ +SRC_URI = "git://anonscm.debian.org/collab-maint/xf86-video-omapfb.git \ + file://0001-Prevents-omapfb-from-from-crashing-when-pixelclock-o.patch \ file://0001-Revert-Set-a-large-CRTC-upper-limit-to-not-prune-lar.patch \ file://0002-Revert-Set-virtual-size-when-configuring-framebuffer.patch \ file://0003-force-plain-mode.patch \
Signature · 1 line
2.1.4
PATCH 6/7 stress: update checksums to match upstream tarball, use BPN and BP in SRC_URI
In reply to Alexander KanavinThis is an update to an earlier commit:
http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=5f9d379fac16cfe6ca155077e6279242e72a92a9
After switching the tarball source from Fedora to upstream it was found that the two tarballs are not identical: Fedora tarball has a copy of GPLv3 in the COPYING file, but author's tarball has GPLv2. Otherwise the tarballs are the same.
--- meta/recipes-extended/stress/stress_1.0.4.bb | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
meta/recipes-extended/stress/stress_1.0.4.bb +4 -4
@@ -2,13 +2,13 @@ DESCRIPTION = "Deliberately simple workload generator for POSIX systems. It \ imposes a configurable amount of CPU, memory, I/O, and disk stress on the system." HOMEPAGE = "http://people.seas.harvard.edu/~apw/stress/" LICENSE = "GPLv2" -LIC_FILES_CHKSUM = "file://COPYING;md5=d32239bcb673463ab874e80d47fae504" +LIC_FILES_CHKSUM = "file://COPYING;md5=b234ee4d69f5fce4486a80fdaf4a4263" -SRC_URI = "http://people.seas.harvard.edu/~apw/stress/stress-${PV}.tar.gz \ +SRC_URI = "http://people.seas.harvard.edu/~apw/${BPN}/${BP}.tar.gz \ file://texinfo.patch \ " -SRC_URI[md5sum] = "a607afa695a511765b40993a64c6e2f4" -SRC_URI[sha256sum] = "369c997f65e8426ae8b318d4fdc8e6f07a311cfa77cc4b25dace465c582163c0" +SRC_URI[md5sum] = "890a4236dd1656792f3ef9a190cf99ef" +SRC_URI[sha256sum] = "057e4fc2a7706411e1014bf172e4f94b63a12f18412378fca8684ca92408825b" inherit autotools
Signature · 1 line
2.1.4
PATCH 7/7 recipeutils.py: allow all characters in regexes used to parse version strings
In reply to Alexander KanavinPreviously only numeric characters were matches and anything else was discarded, so 4.0-rc3, 2005e, 1.0.2a and similar versions got truncated.
--- meta/lib/oe/recipeutils.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
meta/lib/oe/recipeutils.py +2 -2
@@ -638,7 +638,7 @@ def get_recipe_pv_without_srcpv(pv, uri_type): sfx = '' if uri_type == 'git': - git_regex = re.compile("(?P<pfx>(v|))(?P<ver>((\d+[\.\-_]*)+))(?P<sfx>(\+|)(git|)(r|)(AUTOINC|)(\+|))(?P<rev>.*)") + git_regex = re.compile("(?P<pfx>v?)(?P<ver>[^\+]*)((?P<sfx>\+(git)?r?(AUTOINC\+))(?P<rev>.*))?") m = git_regex.match(pv) if m:
@@ -646,7 +646,7 @@ def get_recipe_pv_without_srcpv(pv, uri_type): pfx = m.group('pfx') sfx = m.group('sfx') else: - regex = re.compile("(?P<pfx>(v|r|))(?P<ver>((\d+[\.\-_]*)+))") + regex = re.compile("(?P<pfx>(v|r)?)(?P<ver>.*)") m = regex.match(pv) if m: pv = m.group('ver')
Signature · 1 line
2.1.4
Re: [PATCH 3/7] gptfdisk: update to 1.0.0
In reply to Alexander KanavinQuoted 24 lines · Alexander Kanavin
Also, use tarballs instead of git for fetching source code. git repository does not have version tags, so upstream version check cannot be performed.
Signed-off-by Alexander Kanavin <[email protected]>---
.../recipes-devtools/fdisk/{gptfdisk_git.bb => gptfdisk_1.0.0.bb} | 8 +++-----1 file changed, 3 insertions(+), 5 deletions(-)rename meta/recipes-devtools/fdisk/{gptfdisk_git.bb => gptfdisk_1.0.0.bb} (75%)
meta/recipes-devtools/fdisk/gptfdisk_1.0.0.bb renamed from meta/recipes-devtools/fdisk/gptfdisk_git.bb +0
-0@@ -5,11 +5,9 @@ LIC_FILES_CHKSUM = "file://COPYING;md5=59530bdf33659b29e73d4adb9f9f6552"DEPENDS = "util-linux popt ncurses"
Quoted 1 line
-PV = "0.8.10+git${SRCPV}”
its always preferable to have PV contain SRCPV for recipes using SCM based fetch mechanism. Don’t drop it
Quoted 18 lines
meta/recipes-devtools/fdisk/gptfdisk_1.0.0.bb renamed from meta/recipes-devtools/fdisk/gptfdisk_git.bb +3
-4-SRCREV = "a920398fa393f9d6301b32b191bc01e086ab8bc8" -SRC_URI = "git://git.code.sf.net/p/gptfdisk/code" - -S = "${WORKDIR}/git" +SRC_URI = "${SOURCEFORGE_MIRROR}/${BPN}/${PV}/${BP}.tar.gz" +SRC_URI[md5sum] = "2061f917af084215898d4fea04d8388f" +SRC_URI[sha256sum] = "5b66956743a799fc0471cdb032665c1391e82f9c5b3f1d7d726d29fe2ba01d6c"do_install() { install -d ${D}${sbindir}--
2.1.4--
_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [PATCH 2/7] distcc: update SRC_URI to not use Google Code
In reply to Alexander KanavinQuoted 20 lines · Alexander Kanavin
Google Code is shutting down soon and distcc has moved to github.
Signed-off-by Alexander Kanavin <[email protected]>---
meta/recipes-devtools/distcc/distcc_3.1.bb | 7 +++----1 file changed, 3 insertions(+), 4 deletions(-)meta/recipes-devtools/distcc/distcc_3.1.bb +0
-0@@ -17,16 +17,15 @@ PACKAGECONFIG[popt] = "--without-included-popt,--with-included-popt,popt"RRECOMMENDS_${PN} = "avahi-daemon"
meta/recipes-devtools/distcc/distcc_3.1.bb +1
-1-SRC_URI = "http://distcc.googlecode.com/files/${BPN}-${PV}.tar.bz2 \ +SRC_URI = "git://github.com/distcc/distcc.git \
meta/recipes-devtools/distcc/distcc_3.1.bb PACKAGECONFIG[popt] = "--without-included-popt,--with-included-popt,popt"
Quoted 10 lines
meta/recipes-devtools/distcc/distcc_3.1.bb +1
-3file://separatebuilddir.patch \ file://0001-zeroconf-Include-fcntl.h.patch \ file://default \ file://distccmon-gnome.desktop \ file://distcc \ file://distcc.service" - -SRC_URI[md5sum] = "a1a9d3853df7133669fffec2a9aab9f3" -SRC_URI[sha256sum] = "f55dbafd76bed3ce57e1bbcdab1329227808890d90f4c724fcd2d53f934ddd89" +SRCREV = “9a172d3d20042754a30c0cd9a5d63fd1a8120c31"
it will be good to add PV = “3.1+git${SRCPV}” here as well. You might rename the recipe to distcc_git.bb then as well.
Quoted 12 lines
+S = "${WORKDIR}/git"
inherit autotools pkgconfig update-rc.d useradd systemd
--
2.1.4--
_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [PATCH 3/7] gptfdisk: update to 1.0.0
In reply to Khem RajQuoted 21 lines · Khem Raj
Quoted 21 lines · Alexander Kanavin
Also, use tarballs instead of git for fetching source code. git repository does not have version tags, so upstream version check cannot be performed.
Signed-off-by Alexander Kanavin <[email protected]>---
.../recipes-devtools/fdisk/{gptfdisk_git.bb => gptfdisk_1.0.0.bb} | 8 +++-----1 file changed, 3 insertions(+), 5 deletions(-)rename meta/recipes-devtools/fdisk/{gptfdisk_git.bb => gptfdisk_1.0.0.bb} (75%)
meta/recipes-devtools/fdisk/gptfdisk_1.0.0.bb renamed from meta/recipes-devtools/fdisk/gptfdisk_git.bb +0
-0@@ -5,11 +5,9 @@ LIC_FILES_CHKSUM = "file://COPYING;md5=59530bdf33659b29e73d4adb9f9f6552"
Quoted 5 lines
Quoted 3 lines
DEPENDS = "util-linux popt ncurses"
-PV = "0.8.10+git${SRCPV}”
its always preferable to have PV contain SRCPV for recipes using SCM based fetch mechanism. Don’t drop it
There has been discussion about this in other threads and the conclusion reached is that where there is a specific version being selected, having PV of that specific version is a good thing. With "1.0.0+gitXXX" its unclear if its really 1.0.0 or not.
Cheers,
Richard
Improving SRCPV? (was: Re: [PATCH 3/7] gptfdisk: update to 1.0.0)
Quoted 6 lines · Richard Purdie
Quoted 1 line
its always preferable to have PV contain SRCPV for recipes using SCM based fetch mechanism. Don’t drop it
There has been discussion about this in other threads and the conclusion reached is that where there is a specific version being selected, having PV of that specific version is a good thing. With "1.0.0+gitXXX" its unclear if its really 1.0.0 or not.
Yes. When SCM is used to fetch a pristine tagged release, adding SRCPV just adds unnecessary clutter. My understanding is that SRCPV is only there to enforce package version going forwards when changing one git snapshot for a later git snapshot, when both are based on the same tagged release.
Alex