Your testing suggests the problem is not deterministic. Which is not
surprising since SIGINT handling has always been slightly flakey. But there
is absolutely no way commit f4476100 can "restore proper ^C behavior." Just
do `git log -p -1 f4476100` to see that it cannot possibly change the
behavior of fish since all it does is change a comment. In fact, it
shouldn't even change the probability of hitting the problem since it
should not even change the code that is emitted. I have to wonder if there
is something wrong with your testing methodology. Having said that, I can
reliably reproduce the problem starting with commit ab189a7 thru git head
and cannot reproduce it prior to that commit. Both on macOS 10.12 and
Ubuntu 16.10.
On Wed, Jan 25, 2017 at 6:53 PM, David B. Lamkins <da...@lamkins.net> wrote:
> Heh. Fair enough. :) This is a different system than the one on which I
> noticed the problem this morning. Also, I neglected to do a clean build. So
> let me be more rigourous about building and testing, this time...
>
> With
> > ./configure; and make clean all; and sudo make install
> ... followed by starting a fresh shell, commit f4476100 does indeed
> restore proper ^C behavior. Huh?
>
> I did have to run that twice. The first build failed with
> make: *** No rule to make target 'FISH-BUILD-VERSION-FILE', needed by
> 'obj/fish_version.o'. Stop.
>
> OK... I do have local changes to share/functions/fish_vi_cursor.fish . I
> don't think that's related in any way, but for completeness I did a git
> stash and repeated as above. Same outcome.
>
> I switched back to the master branch and repeated the above. Same issue
> with the first build attempt. Success on the second attempt.
>
> With the build from the master branch installed (I didn't bother to stash
> my local fish_vi_cursor.fish, since that didn't affect previous outcomes),
> ^C doesn't work.
>
>
>
> On Wed, Jan 25, 2017 at 06:23:17PM -0800, Kurtis Rader wrote:
> > On Wed, Jan 25, 2017 at 6:04 PM, David B. Lamkins <[1]da...@lamkins.net>
> wrote:
> >
> > Commit f4476100 does correct the problem.
> >
> >
> > Not on my system. And it isn't possible for that commit to fix the
> problem
> > because it only changes a comment. This is the entirety of that change:
> >
> > -/// A class to aid iteration over jobs list. Note this is used from a
> signal
> > handler, so it must be
> > -/// careful to not allocate memory.
> > +/// A class to aid iteration over jobs list
> >
> > --
> > Kurtis Rader
> > Caretaker of the exceptional canines Junior and Hank
> >
> > References:
> >
> > [1] mailto:da...@lamkins.net
>
--
Kurtis Rader
Caretaker of the exceptional canines Junior and Hank
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
Fish-users mailing list
Fish-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/fish-users
