diff options
| author | Paul Eggleton <paul.eggleton@linux.intel.com> | 2012-06-18 18:55:56 +0100 |
|---|---|---|
| committer | Richard Purdie <richard.purdie@linuxfoundation.org> | 2012-06-21 12:37:23 +0100 |
| commit | 1e50225df2d6b5f10c5648703282235a11b1efb3 (patch) | |
| tree | a57c480b35b62134a2019dfc50a3d6446e43d7ec /meta/classes/archive-configured-source.bbclass | |
| parent | 9bf6b60e1599cf5dd87089d42584583cdfd6807a (diff) | |
| download | openembedded-core-1e50225df2d6b5f10c5648703282235a11b1efb3.tar.gz openembedded-core-1e50225df2d6b5f10c5648703282235a11b1efb3.tar.bz2 openembedded-core-1e50225df2d6b5f10c5648703282235a11b1efb3.zip | |
conf/bitbake.conf: fix reparsing after -p is used
The bitbake wrapper script is set up such that the -p (--parse-only)
command line option is not executed under pseudo, and it sets the
PSEUDO_BUILD variable to indicate whether or not pseudo is being used.
Since PSEUDO_BUILD is allowed through into the environment via
BB_ENV_EXTRAWHITE it influences the data hash and thus if you run
"bitbake -p" and then run bitbake again to actually build something, the
change to PSEUDO_BUILD causes the cache from the -p execution not to be
used. This is fixed simply by adding PSEUDO_BUILD to
BB_HASHCONFIG_WHITELIST in bitbake.conf so that it doesn't influence the
data hash.
Fixes [YOCTO #2600].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Diffstat (limited to 'meta/classes/archive-configured-source.bbclass')
0 files changed, 0 insertions, 0 deletions