subst.t: Test something I nearly broke
authorFather Chrysostomos <sprout@cpan.org>
Mon, 26 Nov 2012 17:40:34 +0000 (09:40 -0800)
committerFather Chrysostomos <sprout@cpan.org>
Tue, 27 Nov 2012 15:05:04 +0000 (07:05 -0800)
commite1b145039d15f6202e21cd23abb95117153eb816
tree4c260a5c1cd7d12c44c73bf54abc005976d0b425
parentd78f32f607952d58a998c5b7554572320dc57b2a
subst.t: Test something I nearly broke

The saving and restoring of $@ when utf8 tables were looked up was
causing pp_subst’s string pointers to go stale.  The existing mechanism
of copying it temporarily to another scalar was already very fragile,
and the new COW mechanism broke it (but I fixed it in the same commit
that introduced the new COW mechanism).
t/re/subst.t