Error executing a python function in exec_python_func()

Question

By building the Yocto image I encountered this issue during populating Yocto SDK. Do somebody have any idea where and how to fix it?

ERROR: Error executing a python function in exec_python_func() autogenerated:

The stack trace of python calls that resulted in this exception/failure was:
File: 'exec_python_func() autogenerated', lineno: 2, function: <module>
     0001:
 *** 0002:extend_recipe_sysroot(d)
     0003:
File: '/yocto/warrior/sources/poky/meta/classes/staging.bbclass', lineno: 557, function: extend_recipe_sysroot
     0553:                    dest = newmanifest[l]
     0554:                    if l.endswith("/"):
     0555:                        staging_copydir(l, targetdir, dest, seendirs)
     0556:                        continue
 *** 0557:                    staging_copyfile(l, targetdir, dest, postinsts, seendirs)
     0558:
     0559:    bb.note("Installed into sysroot: %s" % str(msg_adding))
     0560:    bb.note("Skipping as already exists in sysroot: %s" % str(msg_exists))
     0561:
File: '/yocto/warrior/sources/poky/meta/classes/staging.bbclass', lineno: 151, function: staging_copyfile
     0147:        os.symlink(linkto, dest)
     0148:        #bb.warn(c)
     0149:    else:
     0150:        try:
 *** 0151:            os.link(c, dest)
     0152:        except OSError as err:
     0153:            if err.errno == errno.EXDEV:
     0154:                bb.utils.copyfile(c, dest)
     0155:            else:
Exception: FileExistsError: [Errno 17] File exists: '/yocto/warrior/build-xwayland/tmp/sysroots-components/aarch64/tensorflow/usr/include/tensorflow/contrib/lite/string_util.h' -> '/yocto/warrior/build-xwayland/tmp/work/imx8qmmek-poky-linux/fsl-image-qt5/1.0-r0/recipe-sysroot/usr/include/tensorflow/contrib/lite/string_util.h'

ERROR: [Errno 17] File exists: '/yocto/warrior/build-xwayland/tmp/sysroots-components/aarch64/tensorflow/usr/include/tensorflow/contrib/lite/string_util.h' -> '/media/nxa18908/work/yocto/warrior/build-xwayland/tmp/work/imx8qmmek-poky-linux/fsl-image-qt5/1.0-r0/recipe-sysroot/usr/include/tensorflow/contrib/lite/string_util.h'
DEBUG: Python function extend_recipe_sysroot finished
ERROR: Function failed: extend_recipe_sysroot

It does not like duplicate files

ERROR: fsl-image-qt5-1.0-r0 do_sdk_depends: The file /usr/include/tensorflow/contrib/lite/string_util.h is installed by both tensorflow and tensorflow-lite, aborting

Delldella

Are you running a clean build? If not, remove build/tmp/* or issue bitbake image-name -c cleanall and try again. – Cali

Yes, clean build, the error occurs during SDK generation bitbake *image* -c populate_sdk_ext – Delldella

It does not like duplicate ERROR: fsl-image-qt5-1.0-r0 do_sdk_depends: The file /usr/include/tensorflow/contrib/lite/string_util.h is installed by both tensorflow and tensorflow-lite, aborting – Delldella

you cannot add two different packages that install the same file in rootfs in the same image – Commons

there are two recipes tensorflow-lite and tensorflow sharing codebase - any flag on override? – Delldella

openembedded provides numerous tensorflow , *-native, *-lite , *-for-poets , *-estimator etc... so assume there is a way to generate correct ext-SDK with no conflicts with the use of multiple tensorflow recipes – Delldella

Can you tell which layers do you use and which version of Yocto? – Commons

Is your image being generated flawlessly? Are you installing both tensorflow and tensorflow-lite on it? Or is this just a matter of generating the eSDK? – Cali

Yes, image is generated flawlessly, and SDK as well. and it's just the matter of generating extended-SDK. – Delldella

Maybe you can get more help on Yocto mailing-list? – Commons

Answer 1

Removing/deleting tmp-glibc in bitbake_build and then building with bitbake is successful

Lave