x86_64 is used as a baseline: the "x86_64" entry, whatever status it has,
is transformed into "all", and then the other entries in ARCHITECTURES
either dropped or rearranged appropriately.
* tests/basic-types/array-uint8-uchar-compat.vala is failing for me
on x86_64. This test, which appeared in 0.42.1, also fails with
the previous release, 0.42.0, when "copied" there.
Noteworthy: on x86 that test is OK for both 0.42.0 and 0.42.1.
* tests/control-flow/bug736774-2.vala still fails on all arches.
* Bump version using the exact copy of the patch for 0.30.1 since
there are no offset mismatches.
* Add !x86_gcc2 to secondary arch.
* Replace hard-coded 0.32 by ${portVersion%.*} to reduce future diffs.
* Use vala_x86_common instead of vala_common_x86 and ajust provides
and requires. (No other package is using vala so this fix is safe.)
* Iron INSTALL() to get rid of the empty data/aclocal/ in vala since
{vala,vapigen}.m4 are in vala{,_x86}_common.
* Add "-e" to sed as some OSes have strict/picky implementations.
The "$secondaryArchSuffix" suffix was missing for vala in PROVIDES
and for vala_common in PROVIDES_common. This is required on x86_gcc2
because vala_x86 requires vala_common_x86.