Here's an updated patch that I hope will be a little more clear. I've dropped the change to git apply since that wasn't behaving as I expected, then added a selftest on llvm that demonstrates the issue I'm trying to resolve. Without the change in patch.py included here, the selftest that I'm proposing will fail.
One might argue, "Well, if you change the default PATCHTOOL, you could also just change your patch SRC_URIs and drop the striplevel parameter." While that may be true, the same failure can also be demonstrated using devtool on llvm without modifying PATCHTOOL at all, since by default devtool also uses GitApplyTree. So if you don't consider "modifying a recipe's PATCHTOOL without also modify its SRC_URI patches" to be a valid workflow, devtool is another common workflow where this behavior could be encountered. Hope that helps, Kyle Kyle Russell (1): patch.py: don't apply striplevel to git am command .../llvm/files/0001-Test-new-file.patch | 17 +++++++++++++++++ .../recipes-test/llvm/llvm_%.bbappend | 2 ++ meta/lib/oe/patch.py | 2 +- meta/lib/oeqa/selftest/cases/bbtests.py | 18 ++++++++++++++++++ 4 files changed, 38 insertions(+), 1 deletion(-) create mode 100644 meta-selftest/recipes-test/llvm/files/0001-Test-new-file.patch create mode 100644 meta-selftest/recipes-test/llvm/llvm_%.bbappend -- 2.17.1
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#137625): https://lists.openembedded.org/g/openembedded-core/message/137625 Mute This Topic: https://lists.openembedded.org/mt/73366011/21656 Group Owner: [email protected] Unsubscribe: https://lists.openembedded.org/g/openembedded-core/leave/8023207/1426099254/xyzzy [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
