Follow the upstream recommendations for packagers and switch to
multi-packages:
devel/gettext -> devel/gettext,-runtime
devel/gettext-tools -> devel/gettext,-tools
(new) devel/gettext,-textstyle
the command line. This resulted in shared libraries with an .interp section,
confusing plist generation tools.
Discussed with espie@, guenther@ and kettenis@; initial patch from espie@.
ok espie@
Keep using Java 8, but add necessary patches and update jffi and
jruby-launcher dependencies so that Java 11 could be used. The JRuby
team still recommends the use of Java 8 for JRuby, though that will
be changing in the future.
Keep using Java 8, but add necessary patches and update jffi and
jruby-launcher dependencies so that Java 11 could be used. The JRuby
team still recommends the use of Java 8 for JRuby, though that will
be changing in the future.
lang/python port module. I've not yet come up with a port that
would not need this and one can always set MODPY_TESTDEP to "no"
to prevent the module from touching TEST_DEPENDS.
Idea from afresh1 who pointed out the cpan module already does this.
aja "I support this move."
OK sthen@
- there was never a lang/python3/3.6 (@pkgpath marker for this is present in
current 3.6 plists), python3/3.x was only used for 3.2, so remove these.
- they should probably have originally used "@pkgpath lang/python/3.5" to
trigger updates from 3.5 when 3.5 was removed, but no point adding those
to 3.6 now that 3.7 is the default version. so add @pkgpath lang/python/3.5
markers in 3.7 PLISTs, fixing the upgrade path if old versions were present
(there was previously nothing to trigger 3.5 being removed/updated so
stray packages would be left around).
problem reported by Laurie Tratt
Do the same thing we do with the flang driver and force linking with the
static libraries instead.
Reported off-list by j <AT> bitminer <DOT> ca -- thanks!
like the rest of the ports tree. This also allows removing a bunch of
manual setting of PATH="${PORTPATH}" HOME="${PORTHOME}" done in various
ports etc. This also makes sure CFLAGS is passed through (not everything
honours it but it does improve at least some ports).
Remove NO_CCACHE from www/honk that was added because the above problem
resulted in ccache variables not being passed through correctly breaking
the cc calls in this.
ok kmos@
"ld: error: XX: has non-ABS relocation R_386_GOTOFF against symbol 'XX'"
is seen on other ports (telephony/asterisk, math/py-scipy) when built
with gcc 8 and znotext does not help there, so is unlikely to help here
either.
