On Sun, Dec 30, 2018 at 02:10:52AM +0200, Hannu Vuolasaho wrote:
> The xschem-2.8.1 doesn't support that. I'll write patch to upstream so next
> version will have that fixed too.
It's a make variable so you can always pass it through MAKE_ENV or
MAKE_FLAGS, but since that's relevant for the fake/install, FAKE_FLAGS
seems even more suitable.
Can you look into upstreaming the GNUism fixes, too?
> Already fixed in svn. Next release will drop all of those patches.
OK. Leaving the fprintf fix as you did is fine, I think.
> ./configure has --debug option and it will set -g -O0. I added
> .ifdef DEBUG to turn that configure variable on.
A bit static but good enough for me.
> I took my time and tried to make the Makefile more readable.
Much better, thanks.
> Added that too. I left the comment if someone wants to go with 8.6 for
> some reason or it becomes default tcl.
Not sure if that's worth it, at least stuff like the
`port-lib-depends-check' target will detect these when updating or
switching to 8.6 - I'd remove it and just inline XSCHEM_LDFLAGS.
> Fixed this with horrible sed patch. Is there a better way? I also
> added uninstall
> target for the man page but is it needed anywhere?
Yes, see `man -Otag SUBST_VARS bsd.port.mk' and other users in the tree.
> How about totally evil :) Those are Synopsys/Mentor tools.
So should we disable this behaviour? What's the impact?
No comments:
Post a Comment