Age | Commit message (Collapse) | Author |
|
* Refer to usage.adoc instead of a non-existant docs directory.
* Switch http links to https.
* Update the yocto mailing list address after the move to groups.io.
* Minor style fixes.
Signed-off-by: Paul Barker <pbarker@konsulko.com>
|
|
Signed-off-by: Paul Barker <pbarker@konsulko.com>
|
|
Patchtest wasn't applying some patches because of the line endings EDK2 uses.
If patchest used git am --keep-cr the patches would apply cleanly.
Also the Patch class in lib/patchestpatch.py in the patch test repo
reads the patch using f.read() with the newline parameter implicitly set
to None and this causes newlines to be translated from \r\n into \n.
Hence, when applying the patch with git am it fails; even when using
--keep-cr.
I was able to pass the patchtest tests with this change.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This avoids false positives when user points the target branch
with uppercase, i.e. Morty
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
OE-core does not use tags to set release points so use branches instead.
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
This change required that the mbox or mbox directory must be indicated
through a new command line argument (-m) so the extra parameters can
be provided after double dashes.
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
The main goal is to have a cleaner initscript, just focused on the
patchtest execution where the image takes all the config pieces (exports,
git configs and a simple patchtest runner used by the initscript).
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
purpose
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
created
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Two actions are to update branches depending the situation: in case
user specify a branch, fetch and checkout is done, otherwise just pull
(fetch & merge).
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
There is no need to check for the .git dir presence because the function
'clone' already does it.
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Allows to dynamically gets target branch from a mbox and indicate
patchtest the latter so it can do its job in the proper branch.
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
executed
The intention is to print out the command to be executed, but the echo isn't set
correctly and the command actually runs, causing the following error:
...
You can also run generated qemu images with a command like 'runqemu qemux86'
sh: line 3: test-mboxes: command not found
...
Fixing the echo command so it just displays what's about to be executed.
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Rather than placing the string into a single varibale, use a shell function
that generates it.
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
This avoids several possible problems:
* patch contamination
* incompatible metadata & build folder
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
This is 2ece64bf8cb8841e2e1db87d6c426f3f1637e8a7 but as bbappend because the latter
is not at pyro, the branch use to create guest.
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
The patchtest execution should be done by a non-root user as Bitbake tools are
used in the process. In order to avoid any error, we should switch from a root
user to a non-root system user after setting the mount point and right before
executing any patchtest script.
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
|
|
Because it uses Tinfoil/Bitbake, the patchtest execution should be done by a
non-root user; creating a new system user on initscripts_1.0.bbappend via useradd
to avoid further issues.
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
|
|
This change prevents the guest machine on running without meemory during
the patchtest execution; in particular avoiding ExpansionErrors when Bitbake
is used.
This change ensures the following error doesn't happen:
ERROR: ExpansionError during parsing /home/patchtest-user/openembedded-core/
meta/recipes-extended/images/core-image-testmaster.bb
...
exception OSError: [Errno 12] Cannot allocate memory
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
For some reason, version 4.10 returns and error whilst
preparing the oe-core environment.
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
As part of the migration, packagegroups must point at Python3 packages.
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
|
|
Adds new recipe defining the wrapt module
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
|
|
Adds new recipe defining the unidiff module locally within the patchtest layer.
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
|
|
Adds new recipe defining the pyparsing module locally within the patchtest
layer.
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
|
|
Adds new recipe defining the lazy-object-proxy module.
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
|
|
Adds new recipe defining the isort python3 module.
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
|
|
Adds new recipe defining the Py3 astroid module.
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
|
|
Adds new recipe defining the pylint module locally within the
patchtest layer.
Signed-off-by: Daniela Plascencia <daniela.plascencia@linux.intel.com>
|
|
corrupted
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
For some reason, after a patchtest execution inside the guest machine
using the oe-core repo inside the share folder, the latter gets corrupted.
The corruption is observed by git when git status is executed inside the
host machine. It may be related to the 9p-virtio filesystem and/or the
machine being shutdown incorrectly, leading a dirty filesystem. One other
important finding that occurs often is that even if patchtest is not executed,
after the guest shutdown, this is observed in the host machine at the openembedded-core folder
shared with guest:
diff --git a/meta/recipes-extended/ghostscript/ghostscript/i686 b/meta/recipes-extended/ghostscript/ghostscript/i686
deleted file mode 120000
index 87aaca5903..0000000000
--- a/meta/recipes-extended/ghostscript/ghostscript/i686
+++ /dev/null
@@ -1 +0,0 @@
-i586
\ No newline at end of file
diff --git a/meta/recipes-extended/ghostscript/ghostscript/i686 b/meta/recipes-extended/ghostscript/ghostscript/i686
new file mode 100644
index 0000000000..87aaca5903
--- /dev/null
+++ b/meta/recipes-extended/ghostscript/ghostscript/i686
@@ -0,0 +1 @@
+i586
\ No newline at end of file
diff --git a/scripts/pybootchartgui/pybootchartgui/main.py b/scripts/pybootchartgui/pybootchartgui/main.py
deleted file mode 120000
index b45ae0a3d2..0000000000
--- a/scripts/pybootchartgui/pybootchartgui/main.py
+++ /dev/null
@@ -1 +0,0 @@
-main.py.in
\ No newline at end of file
diff --git a/scripts/pybootchartgui/pybootchartgui/main.py b/scripts/pybootchartgui/pybootchartgui/main.py
new file mode 100644
index 0000000000..b45ae0a3d2
--- /dev/null
+++ b/scripts/pybootchartgui/pybootchartgui/main.py
These two files are the only softlinks found on the problem, so there is something going on
with the share folder and softlinks, which no root cause has been found.
The following change copies the openembedded-core repo locally and tests are done from the latter. Two other modifications:
increase the number of open files, and remove the explicit umount command which is not required. The former
was observed with several set of mboxes leading the following log:
You can also run generated qemu images with a command like 'runqemu qemux86'
/patchtest/mboxes/6627.4.mbox
/patchtest/patchtest/scripts/test-mboxes: line 97: /patchtest/results/6627.4.mbox.log: Too many open files
[YOCTO #11707]
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|
|
Log from the test suite can contain non-json data, so filter the latter
before creating the report.
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
|