empty dirs in glib-2.0
Hi,
Compiling the Angstrom Linux, in the file openembedded-core/meta/recipes- core/glib-2.0/glib-2.0_2.30.2.bb (a85f283ffae5407b23532ce2262142ee2d986af2) there are the followings:
do_install_append() {
# remove some unpackaged files
rm -f ${D}${libdir}/gdbus-2.0/codegen/*.pyc
rm -f ${D}${libdir}/gdbus-2.0/codegen/*.pyo
# and empty dirs
rmdir ${D}${libdir}/gio/modules/
rmdir ${D}${libdir}/gio/
}I have problems with the "empty dirs" section: the directory is not empty and therefore the installation fails here.
I replaced it with "rm -rf ..." and now it worx but i do not sure it is a good solution.
Why these directories are expected to be empty? What does it mean if they are not?
Thanx in advance
Gyorgy Kovesdi
Quoted 15 lines · Kövesdi György
Hi,
Compiling the Angstrom Linux, in the file openembedded-core/meta/recipes- core/glib-2.0/glib-2.0_2.30.2.bb (
a85f283ffae5407b23532ce2262142ee2d986af2) there are the followings:
do_install_append(){# remove some unpackaged files rm -f ${D}${libdir}/gdbus-2.0/codegen/*.pyc rm -f ${D}${libdir}/gdbus-2.0/codegen/*.pyo # and empty dirs rmdir ${D}${libdir}/gio/modules/ rmdir ${D}${libdir}/gio/ }
Quoted 6 lines
I have problems with the "empty dirs" section: the directory is not empty and therefore the installation fails here.
I replaced it with "rm -rf ..." and now it worx but i do not sure it is a good solution.
Why these directories are expected to be empty? What does it mean if they are not?
rmdir was used specifically so we'd know if the directory wasn't empty. What files are in there?
Are there any .bbappends to the glib recipe in any of your other layers?
Cheers,
Richard
Quoted 2 lines
rmdir was used specifically so we'd know if the directory wasn't empty. What files are in there?
...../usr/lib/gio/modules:
total 32K
drwxr-xr-x 2 kgy users 4.0K Feb 27 11:26 . drwxr-xr-x 3 kgy users 4.0K Feb 27 11:26 .. -rwxr-xr-x 1 kgy users 2.0K Feb 27 11:26 libgiofam.la -rwxr-xr-x 1 kgy users 20K Feb 27 11:26 libgiofam.so
It is reproducible now, even in a brand new checkout.
Regards
Gyorgy Kovesdi
Quoted 1 line
It is reproducible now, even in a brand new checkout.
The revision of the glib-2.0_2.30.2.bb file is 4642ae91cee2978956a61cf15df7ea62e3eb1726
Quoted 12 lines · Gyorgy Kovesdi
Quoted 2 lines
rmdir was used specifically so we'd know if the directory wasn't empty. What files are in there?
...../usr/lib/gio/modules:
total 32K
drwxr-xr-x 2 kgy users 4.0K Feb 27 11:26 .
drwxr-xr-x 3 kgy users 4.0K Feb 27 11:26 ..(quoted diff) +0
-2-rwxr-xr-x 1 kgy users 2.0K Feb 27 11:26 libgiofam.la -rwxr-xr-x 1 kgy users 20K Feb 27 11:26 libgiofam.soIt is reproducible now, even in a brand new checkout.
Quoted 8 lines
Regards
Gyorgy Kovesdi_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
I can confirm I also just ran into this issue on an Angstrom Master checkout.
Same problem!!!
Quoted 29 lines · Jack Mitchell
Quoted 12 lines · Gyorgy Kovesdi
Quoted 2 lines
rmdir was used specifically so we'd know if the directory wasn't empty. What files are in there?
...../usr/lib/gio/modules:
total 32K
drwxr-xr-x 2 kgy users 4.0K Feb 27 11:26 .
drwxr-xr-x 3 kgy users 4.0K Feb 27 11:26 ..(quoted diff) +0
-2-rwxr-xr-x 1 kgy users 2.0K Feb 27 11:26 libgiofam.la -rwxr-xr-x 1 kgy users 20K Feb 27 11:26 libgiofam.soIt is reproducible now, even in a brand new checkout.
Quoted 8 lines
Regards
Gyorgy Kovesdi_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-coreI can confirm I also just ran into this issue on an Angstrom Master checkout.
_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Signature · 3 lines
Yi Qingliang [email protected] https://niqingliang2003.wordpress.com
Quoted 28 lines · Ni Qingliang
Same problem!!!
Quoted 25 lines · Jack Mitchell
Quoted 12 lines · Gyorgy Kovesdi
Quoted 2 lines
rmdir was used specifically so we'd know if the directory wasn't empty. What files are in there?
...../usr/lib/gio/modules:
total 32K
drwxr-xr-x 2 kgy users 4.0K Feb 27 11:26 .
drwxr-xr-x 3 kgy users 4.0K Feb 27 11:26 ..(quoted diff) +0
-2-rwxr-xr-x 1 kgy users 2.0K Feb 27 11:26 libgiofam.la -rwxr-xr-x 1 kgy users 20K Feb 27 11:26 libgiofam.soIt is reproducible now, even in a brand new checkout.
Quoted 8 lines
Regards
Gyorgy Kovesdi_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-coreI can confirm I also just ran into this issue on an Angstrom Master checkout.
You can fix the issue adding --disable-fam like in glib-2.0/glib.inc
-EXTRA_OECONF_virtclass-native = "--disable-dtrace --disable-systemtap" +EXTRA_OECONF_virtclass-native = "--disable-dtrace --disable-systemtap --disable-fam"
The issue has been ackowledged, more tests are going on. Fix will follow soon.
Regards
Andrea
Quoted 15 lines
Quoted 4 lines
_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core--
Yi Qingliang
[email protected]
https://niqingliang2003.wordpress.com_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Quoted 22 lines · Ni Qingliang
Same problem!!!
Quoted 19 lines · Jack Mitchell
Quoted 10 lines · Gyorgy Kovesdi
Quoted 2 lines
rmdir was used specifically so we'd know if the directory wasn't empty. What files are in there?
...../usr/lib/gio/modules:
total 32K
drwxr-xr-x 2 kgy users 4.0K Feb 27 11:26 .
drwxr-xr-x 3 kgy users 4.0K Feb 27 11:26 ..(quoted diff) +0
-2-rwxr-xr-x 1 kgy users 2.0K Feb 27 11:26 libgiofam.la -rwxr-xr-x 1 kgy users 20K Feb 27 11:26 libgiofam.soQuoted 5 lines
It is reproducible now, even in a brand new checkout.
Regards
Gyorgy KovesdiI can confirm I also just ran into this issue on an Angstrom Master checkout.
We need to add --disable-fam to EXTRA_OECONF. Since a lot of people are hitting this, I've pushed such a patch directly.
Cheers,
Richard