| Age | Commit message (Collapse) | Author |
|---|
|
Due to some logic within opkg, a package with the name matching a
dependency will always win over a package with that name in RPROVIDES -
even if there is an RCONFLICTS (which is silently ignored), higher feed
priority and version. The end result is that buildtools gets perl
installed instead of the nativesdk-buildtools-perl-dummy package and
that perl (with missing dependencies) gets used in preference to the
host one, which is precisely what we were trying to avoid.
This is almost certainly a bug in opkg, especially as the other
package's dependencies aren't properly installed under these
circumstances either. However, specifying RREPLACES works around this,
and with no apparent side-effects is probably the safest solution for
now.
At the same time I noticed that in prepending to SDK_PACKAGE_ARCHS we
were actually ending up with a low priority for the dummy package feed
rather than a high one, so change to append it instead. This has no
effect on the packages that get installed at the moment, but should be
done in case the package manager behaviour changes to factor in the feed
priority in future.
Fixes [YOCTO #9469].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Since python expansion of bb data variables is disable in order to
provide a standard interface usage for expand variables this variables
aren't expanded now so change to call d.expand('${VAR}').
This API expansion change was causing to install perl in nativesdk.
Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Recent changes to this recipe caused automated PR increments
to break, regressing package feeds. The only way to recover
is to bump PR, so do this centrally to fix anyone affected.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Turns out I did a silly thing in OE-Core revision
9b1831cf4a2940dca1d23f14dff460ff5a50a520 and forgot to remove the
explicit setting of PACKAGE_ARCH outside of the anonymous python
function; the original bug was apparently fixed but the functionality of
allarch.bbclass was being disabled because it was able to see that
PACKAGE_ARCH was not set to "all" - which was what I was trying to
ensure.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This recipe produces an empty dummy package (in order to satisfy
dependencies on perl so we don't have perl within buildtools-tarball).
Because we were inheriting nativesdk here the recipe was being rebuilt,
but having forced PACKAGE_ARCH to a particular value the packages for
each architecture were stepping on eachother. Since the packages are
empty they can in fact be allarch (even though they won't actually go
into the "all" package feed). It turns out that nheriting nativesdk
wasn't actually necessary either, so drop that.
Fixes [YOCTO #8509].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Since the split out of git-perltools, some git tools (such as "git am",
"git send-email" and "git-submodule") have no longer been part of the
buildtools. We need these, so add them back in.
However, adding git-perltools to buildtools triggers perl itself being
brought into buildtools as well, and we don't want that; but we also
don't want to have to hack the git recipe or indeed anything else that
starts depending on perl. Thus, add a dummy package which gets installed
in its place, in a separate package architecture that is only enabled
for buildtools to ensure it doesn't start appearing in place of
nativesdk-perl anywhere else.
Fixes [YOCTO #7033].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
Paul Eggleton
Aníbal Limón
Richard Purdie