15943 2025-07-25 08:47:05 +0000 devtool upgrade fails for recipe that contains a git submodule 2025-12-03 08:22:07 +0000 1 1 7 Build System, Metadata & Runtime OE-Core Scripts and Tools unspecified x86 Multiple CLOSED FIXED Medium normal 5.3 1 pistora.tobias pistora.tobias richard.purdie ross.burton --- --- No (bug/feature does not impact docs) oldest_to_newest 102514 0 5127 pistora.tobias 2025-07-25 08:47:05 +0000 Created attachment 5127 My bugfix based on yocto-5.0.8 The devtool upgrade command fails while it goes through a git submodule directory. It happens when devtool executes 'git rev-list devtool-base..<submodule commit hash>'. The debug log shows that 'git rev-list' is called from parent directory but with the commit hash of the submodule which does not exist there ("amqp-proxy" is the parent git repo and "contracts" the submodule): -- snipped -- DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git fetch DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git checkout 35b87c6723cf5afbee0cf02c7447eb8c833185af DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git tag -f devtool-base-new DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git submodule update --recursive DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git submodule foreach 'git tag -f devtool-base-new' DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git submodule --quiet foreach 'echo $sm_path' DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git branch -r --contains 35b87c6723cf5afbee0cf02c7447eb8c833185af DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git rev-parse HEAD DEBUG: Running command <build_path>/workspace/sources/amqp-proxy/contracts> git rev-parse HEAD DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git checkout devtool-patched -b devtool DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git branch --list devtool-override-* DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git rev-list devtool-base..35b87c6723cf5afbee0cf02c7447eb8c833185af INFO: Rebasing devtool onto 35b87c6723cf5afbee0cf02c7447eb8c833185af DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git checkout devtool DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git rebase 35b87c6723cf5afbee0cf02c7447eb8c833185af DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git rev-list devtool-base..35b87c6723cf5afbee0cf02c7447eb8c833185af DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git checkout devtool DEBUG: Running command <build_path>/workspace/sources/amqp-proxy/contracts> git checkout devtool-patched -b devtool DEBUG: Running command <build_path>/workspace/sources/amqp-proxy/contracts> git branch --list devtool-override-* DEBUG: Running command <build_path>/workspace/sources/amqp-proxy> git rev-list devtool-base..90727cda36bab25fb106fc100e5a6177467f5456 ERROR: Execution of 'git rev-list devtool-base..90727cda36bab25fb106fc100e5a6177467f5456' failed with exit code 128: fatal: Invalid revision range devtool-base..90727cda36bab25fb106fc100e5a6177467f5456 -- snipped end -- The full command that was executed for this output is 'command: devtool -d upgrade amqp-proxy'. The related recipe uses "gitsm://" for fetching the code. After inspecting the related code in 'poky/scripts/lib/devtool/upgrade.py' I have found a working fix for me. See attached patch, based on yocto 5.0.8. Hope this helps. Conclusion: Is seems that the path argument is missing when running the 'git rev-list' calls from submodule directories. However, because I have found the same (unfixed) code in master branch and have not found a bug report for this issue yet (the affected code was add around a year ago), I am not sure if the implementation is intended and I am somehow terribly wrong with what I am doing... 102533 1 richard.purdie 2025-07-31 14:40:23 +0000 "oe-selftest -r devtool" has a lot of selftests for devtool. Would you be able to write a test case which covers this scenario? If so, at a quick glance, I think your patch seems reasonable and you could send that for review too? 102538 2 pistora.tobias 2025-08-01 08:30:36 +0000 (In reply to Richard Purdie from comment #1) > "oe-selftest -r devtool" has a lot of selftests for devtool. Would you be > able to write a test case which covers this scenario? I think so... After a quick glance inside the test code I got an idea where to look next^^. However, it may take a while but I will get used to it. > If so, at a quick > glance, I think your patch seems reasonable and you could send that for > review too? Sure, I will do that. 103466 3 pistora.tobias 2025-12-03 08:17:03 +0000 Fixed in 4c7b0007327d592b8f5d1f50ebb0325dc483bd87 (https://git.yoctoproject.org/poky/commit/?id=4c7b0007327d592b8f5d1f50ebb0325dc483bd87) 103467 4 pistora.tobias 2025-12-03 08:20:51 +0000 The requested test case can be found in 921faf324658371193b1f0847ebc1912d7fef25d (https://git.yoctoproject.org/poky/commit/?id=921faf324658371193b1f0847ebc1912d7fef25d) 103468 5 pistora.tobias 2025-12-03 08:22:07 +0000 fixed 5127 2025-07-25 08:47:05 +0000 2025-07-25 08:47:05 +0000 My bugfix based on yocto-5.0.8 fix-devtool-bug-that-brakes-upgrade-for-recipes-with-submodule.patch application/mbox 1886 pistora.tobias RnJvbSAzZDJlMTc2YjNhMGJmNzBmNmI4Mjk1MWM1Yzg5ZWVlZTljOWE2YjUyIE1vbiBTZXAgMTcg MDA6MDA6MDAgMjAwMQpGcm9tOiBUb2JpYXMgUGlzdG9yYSA8dC5waXN0b3JhQHNjaHJhY2stc2Vj b25ldC5jb20+CkRhdGU6IFRodSwgMjQgSnVsIDIwMjUgMTA6Mjg6MjAgKzAyMDAKU3ViamVjdDog W1BBVENIXSBmaXggZGV2dG9vbCBidWcgdGhhdCBicmFrZXMgdXBncmFkZSBmb3IgcmVjaXBlcyB3 aXRoIGdpdAogc3VibW9kdWxlcwoKJ2dpdCByZXYtbGlzdCcgY21kcyBoYXZlIGJlZW4gZXhlY3V0 ZWQgaW4gcGFyZW50IGdpdCByZXBvIGRpciB3aGlsZSBwcm9jZXNzaW5nIHN1Ym1vZHVsZXMgaW5z aWRlIHRoZWlyIHN1Yi1kaXJzCi0tLQogc2NyaXB0cy9saWIvZGV2dG9vbC91cGdyYWRlLnB5IHwg NCArKy0tCiAxIGZpbGUgY2hhbmdlZCwgMiBpbnNlcnRpb25zKCspLCAyIGRlbGV0aW9ucygtKQoK ZGlmZiAtLWdpdCBhL3NjcmlwdHMvbGliL2RldnRvb2wvdXBncmFkZS5weSBiL3NjcmlwdHMvbGli L2RldnRvb2wvdXBncmFkZS5weQppbmRleCBmYTViOGVmM2M3Li5kODVkYjBiYjNlIDEwMDY0NAot LS0gYS9zY3JpcHRzL2xpYi9kZXZ0b29sL3VwZ3JhZGUucHkKKysrIGIvc2NyaXB0cy9saWIvZGV2 dG9vbC91cGdyYWRlLnB5CkBAIC0yNzksNyArMjc5LDcgQEAgZGVmIF9leHRyYWN0X25ld19zb3Vy Y2UobmV3cHYsIHNyY3RyZWUsIG5vX3BhdGNoLCBzcmNyZXYsIHNyY2JyYW5jaCwgYnJhbmNoLCBr ZWUKICAgICAgICAgICAgICMgbm90ZSBpcyBzdGlsbCBhdHRhY2hlZCB0byB0aGUgb2xkIGNvbW1p dC4gQXZvaWQgdGhpcyBieSBtYWtpbmcKICAgICAgICAgICAgICMgc3VyZSBhbGwgb2xkIGRldnRv b2wgcmVsYXRlZCBjb21taXRzIGhhdmUgYSBub3RlIGF0dGFjaGVkIHRvIHRoZW0KICAgICAgICAg ICAgICMgKHRoaXMgYXNzdW1lcyBnaXQgY29uZmlnIG5vdGVzLnJld3JpdGVNb2RlIGlzIHNldCB0 byBpZ25vcmUpLgotICAgICAgICAgICAgKHN0ZG91dCwgXykgPSBfX3J1bignZ2l0IHJldi1saXN0 IGRldnRvb2wtYmFzZS4uJXMnICUgdGFyZ2V0X2JyYW5jaCkKKyAgICAgICAgICAgIChzdGRvdXQs IF8pID0gX3J1bignZ2l0IHJldi1saXN0IGRldnRvb2wtYmFzZS4uJXMnICUgdGFyZ2V0X2JyYW5j aCwgY3dkPXBhdGgpCiAgICAgICAgICAgICBmb3IgcmV2IGluIHN0ZG91dC5zcGxpdGxpbmVzKCk6 CiAgICAgICAgICAgICAgICAgaWYgbm90IG9lLnBhdGNoLkdpdEFwcGx5VHJlZS5nZXROb3Rlcyhw YXRoLCByZXYpOgogICAgICAgICAgICAgICAgICAgICBvZS5wYXRjaC5HaXRBcHBseVRyZWUuYWRk Tm90ZShwYXRoLCByZXYsICJkdW1teSIpCkBAIC0yOTcsNyArMjk3LDcgQEAgZGVmIF9leHRyYWN0 X25ld19zb3VyY2UobmV3cHYsIHNyY3RyZWUsIG5vX3BhdGNoLCBzcmNyZXYsIHNyY2JyYW5jaCwg YnJhbmNoLCBrZWUKICAgICAgICAgICAgICAgICAgICAgICAgIGxvZ2dlci53YXJuaW5nKCdDb21t YW5kIFwnJXNcJyBmYWlsZWQ6XG4lcycgJSAoZS5jb21tYW5kLCBlLnN0ZG91dCkpCiAKICAgICAg ICAgICAgICMgUmVtb3ZlIGFueSBkdW1teSBub3RlcyBhZGRlZCBhYm92ZS4KLSAgICAgICAgICAg IChzdGRvdXQsIF8pID0gX19ydW4oJ2dpdCByZXYtbGlzdCBkZXZ0b29sLWJhc2UuLiVzJyAlIHRh cmdldF9icmFuY2gpCisgICAgICAgICAgICAoc3Rkb3V0LCBfKSA9IF9ydW4oJ2dpdCByZXYtbGlz dCBkZXZ0b29sLWJhc2UuLiVzJyAlIHRhcmdldF9icmFuY2gsIGN3ZD1wYXRoKQogICAgICAgICAg ICAgZm9yIHJldiBpbiBzdGRvdXQuc3BsaXRsaW5lcygpOgogICAgICAgICAgICAgICAgIG9lLnBh dGNoLkdpdEFwcGx5VHJlZS5yZW1vdmVOb3RlKHBhdGgsIHJldiwgImR1bW15IikKIAotLSAKMi4z OS41Cgo=