On Thu, Nov 05, 2015 at 12:18:23PM +0000, Stuart Henderson wrote:
On 2015/11/05 11:42, Tati Chevron wrote:
>Is there some reason to not simply use the packages though? You'll end up
>with a lot less junk installed on your system than building it yourself.
I need to patch the mutt source to improve handling of keyboards with
keys beyond F12. For example, at the moment adding bindings for F15-F24
fails.
Fair enough. Things will go a bit smoother if you pkg_add the build
dependencies though (libidn, lynx, docbook-xsl, qdbm)
Have you tried to get this diff upstream at all? It shouldn't be too
controversial and would help other users and make things easier for you.
Not yet, but you're right, I probably should do.
On 2015-11-04, Tati Chevron <[email protected]> wrote:
But what's changed to cause this new behaviour?
cp(1) changed from using chflags to chflagsat to preserve nanosecond
precision in timestamps. I think most ports developers stopped using
systrace when the support to run "make fake" as non-root was added,
so as you're finding out it's more likely to break now (it also slows
down builds a fair bit, and can cause the build to behave differently
than without systrace e.g. autoconf checks could behave differently,
or bad codepaths could be taken in programs when system calls fail
that they aren't expecting).
So is the use of systrace as described in the FAQ 15.3.3, now depreciated?
Or will these breakages still be fixed as they're discovered?
I've just noticed that the list server is silently dropping some of my
mail... Odd.
--
Tati Chevron
Perl and FORTRAN specialist.
SWABSIT development and migration department.
http://www.swabsit.com
