_remove operator -- taskhash issue

_remove operator -- taskhash issue

[Mark Hatle]

I noticed something today on morty.  (So this might be fixed in master.)

I've got a variable that is defined as (-e output):

# $PACKAGECONFIG [20 operations]
#   _append[pn-qemu-native]
/home/mhatle/git/lpd/wr9-Base/build-test/conf/local.conf:384
#     " sdl"
#   _append[pn-nativesdk-qemu]
/home/mhatle/git/lpd/wr9-Base/build-test/conf/local.conf:385
#     " sdl"
#   set
/home/mhatle/git/lpd/wr9-Base/layers/oe-core/meta/conf/documentation.conf:317
#     [doc] "This variable provides a means of enabling or disabling features of
a recipe on a per-recipe basis."
#   set
/home/mhatle/git/lpd/wr9-Base/layers/oe-core/meta/recipes-graphics/mesa/mesa.inc:30
#     [_defaultval] "gbm egl gles dri ${MESA_CRYPTO}
${@bb.utils.contains('DISTRO_FEATURES', 'x11', 'x11', '',
d)}           ${@bb.utils.contains('DISTRO_FEATURES', 'wayland', 'wayland', '',
d)}           "
# pre-expansion value:
#   "gbm egl gles dri ${MESA_CRYPTO}
${@bb.utils.contains('DISTRO_FEATURES', 'x11', 'x11', '', d)}           ${
@bb.utils.contains('DISTRO_FEATURES', 'wayland', 'wayland', '', d)}           "
PACKAGECONFIG="gbm egl gles dri openssl                 x11
       "

(note lots of white space above).

If I add:

PACKAGECONFIG_remove_override = "foobar"

where the override is _NOT_ selected I now get, same as above --except-- for the
following:

#   _remove[override] /home/mhatle/git/lpd/wr9-Base/build-test/conf/local.conf:1
#     "foobar"
# pre-expansion value:
#   "gbm egl gles dri ${MESA_CRYPTO} ${@bb.utils.contains('DISTRO_FEATURES',
'x11', 'x11', '', d)} ${@bb.utils.contains('DISTRO_FEATURES', 'wayland',
'wayland', '', d)}"
PACKAGECONFIG="gbm egl gles dri openssl x11"


Note, the white space is gone, even though the evaluated contents of the
variable is not affected.  The problem with this is that it's changing the task
hash for my recipe's .bbappend, even though the override is not triggered.

--Mark

Re: _remove operator -- taskhash issue

[Richard Purdie]

On Wed, 2017-03-15 at 23:14 -0500, Mark Hatle wrote:
> I noticed something today on morty.  (So this might be fixed in
> master.)

It is indeed fixed in master:

http://git.yoctoproject.org/cgit.cgi/poky/commit/bitbake/?id=45df694a9f472ac2f684aadac4d864c3dfdc48a7

I've backported that fix to morty since it seems like a reasonably
thing to fix there.

I also wrote a test case for it and added it to bitbake-selftest.

Cheers,

Richard