Patch review
cover Patch review
Please review this final set of patches for the 3.1.8 release and have comments back by end of day Friday.
Passed a-full on autobuilder:
https://autobuilder.yoctoproject.org/typhoon/#/builders/83/builds/2172
The following changes since commit cc49591d84d241d90e3dccb3e174ddfd737de311:
meta/lib/oe/rootfs.py: Fix typo "Restoreing" -> "Restoring" (2021-05-14 07:16:38 -1000)
are available in the Git repository at:
git://git.openembedded.org/openembedded-core-contrib stable/dunfell-nut
http://cgit.openembedded.org/openembedded-core-contrib/log/?h=stable/dunfell-nut
Richard Purdie (2):
Revert "cml1.bbclass: Return sorted list of cfg files"
sstate: Handle manifest 'corruption' issue
Stefan Ghinea (1):
boost: fix do_fetch failure
meta/classes/cml1.bbclass | 2 +- meta/classes/sstate.bbclass | 16 +++++++++++++++- meta/recipes-support/boost/boost-1.72.0.inc | 2 +- 3 files changed, 17 insertions(+), 3 deletions(-)
Signature · 1 line
2.25.1
PATCH 1/3 Revert "cml1.bbclass: Return sorted list of cfg files"
From: Richard Purdie <[email protected]>
This reverts commit dee41e92f0efac7e453597bed4b4c02f867e3aa9.
This patch breaks cases where some config files make changes to earlier ones, ordering is important. The reproducibility issue in busybox was elsewhere.
--- meta/classes/cml1.bbclass | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
meta/classes/cml1.bbclass +1 -1
@@ -6,7 +6,7 @@ def find_cfgs(d): if s.endswith('.cfg'): sources_list.append(s) - return sorted(sources_list) + return sources_list cml1_do_configure() { set -e
Signature · 1 line
2.25.1
PATCH 2/3 boost: fix do_fetch failure
In reply to Steve SakomanFrom: Stefan Ghinea <[email protected]>
Bintray service has been discontinued causing boost do_fetch to fail:
WARNING: boost-1.72.0-r0 do_fetch: Failed to fetch URL
https://dl.bintray.com/boostorg/release/1.76.0/source/boost_1_72_0.tar.bz2,
attempting MIRRORS if available
--- meta/recipes-support/boost/boost-1.72.0.inc | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
meta/recipes-support/boost/boost-1.72.0.inc +1 -1
@@ -11,7 +11,7 @@ BOOST_VER = "${@"_".join(d.getVar("PV").split("."))}" BOOST_MAJ = "${@"_".join(d.getVar("PV").split(".")[0:2])}" BOOST_P = "boost_${BOOST_VER}" -SRC_URI = "https://dl.bintray.com/boostorg/release/${PV}/source/${BOOST_P}.tar.bz2" +SRC_URI = "https://boostorg.jfrog.io/artifactory/main/release/${PV}/source/${BOOST_P}.tar.bz2" SRC_URI[md5sum] = "cb40943d2a2cb8ce08d42bc48b0f84f0" SRC_URI[sha256sum] = "59c9b274bc451cf91a9ba1dd2c7fdcaf5d60b1b3aa83f2c9fa143417cc660722"
Signature · 1 line
2.25.1
PATCH 3/3 sstate: Handle manifest 'corruption' issue
In reply to Steve SakomanFrom: Richard Purdie <[email protected]>
Under certain build patterns, warnings about missing manifests can appear. These are real issues where the manifest was removed and shouldn't have been.
Martin Jansa was able to find a reproducer of:
MACHINE=qemux86 bitbake zlib-native
echo 'PR = "r1"' >> meta/recipes-core/zlib/zlib_1.2.11.bb MACHINE=qemux86-64 bitbake zlib-native
MACHINE=qemux86 bitbake zlib-native
<the zlib-native manifest is now removed along with the sysroot-components contents>
The code maintains a per machine list of stamps but a per PACAGE_ARCH list of stamp/manifest/workdir mappings. The latter is only appended to for speed with the assumption that once stamps are gone, the code wouldn't trigger.
The code only ever appends to the mapping list (for speed/efficency under lock) meaning that multiple entries can result where the stamp/workdir differs due to version changes but the manifest remains the same.
By switching MACHINE part way through the build, the older stamp is referenced and the manifest is incorrectly removed as it matches an now obsolete entry in the mapping file.
There are two possible fixes, one is to rewrite the mapping file every time which means adding regexs, iterating and generally complicating that code. The second option is to only use the last mapping entry in the file for a given manifest and ignore any earlier ones. This patch implments the latter.
Also drop the stale entries if we are rewriting it.
--- meta/classes/sstate.bbclass | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-)
meta/classes/sstate.bbclass +15 -1
@@ -319,6 +319,8 @@ def sstate_install(ss, d): if os.path.exists(i): with open(i, "r") as f: manifests = f.readlines() + # We append new entries, we don't remove older entries which may have the same + # manifest name but different versions from stamp/workdir. See below. if filedata not in manifests: with open(i, "a+") as f: f.write(filedata)
@@ -1175,11 +1177,21 @@ python sstate_eventhandler2() { i = d.expand("${SSTATE_MANIFESTS}/index-" + a) if not os.path.exists(i): continue + manseen = set() + ignore = [] with open(i, "r") as f: lines = f.readlines() - for l in lines: + for l in reversed(lines): try: (stamp, manifest, workdir) = l.split() + # The index may have multiple entries for the same manifest as the code above only appends + # new entries and there may be an entry with matching manifest but differing version in stamp/workdir. + # The last entry in the list is the valid one, any earlier entries with matching manifests + # should be ignored. + if manifest in manseen: + ignore.append(l) + continue + manseen.add(manifest) if stamp not in stamps and stamp not in preservestamps and stamp in machineindex: toremove.append(l) if stamp not in seen:
@@ -1210,6 +1222,8 @@ python sstate_eventhandler2() { with open(i, "w") as f: for l in lines: + if l in ignore: + continue f.write(l) machineindex |= set(stamps) with open(mi, "w") as f:
Signature · 1 line
2.25.1