On Thu, 13 Jun 2019 03:46:27 -0700, hv wrote: > On Wed, 12 Jun 2019 17:22:33 -0700, tonyc wrote: > > The attached fixes it (applies on top of the 131867 fix). > > Can you explain why the similar code for *+ and *- is not similarly > wrong? That might be worth a comment in the code too: it looks like > this bug started as essentially a copy of that code. There's special handling in gv_fetchpvn_flags(): if (len == 1 && stash == PL_defstash) { maybe_multimagic_gv(gv, name, sv_type); } that calls maybe_multimagic_gv() even if the glob already exists for single character names. Loading Tie::Hash::NamedCapture when only @{^CAPTURE} is being referenced is a minor performance hit, maybe that's what the handling for @+ is trying to avoid. > It would also be worth swapping the '+' and '-' arguments in the two > require_tie_mod_s() calls, as Hauke also noted - if anyone ever sees > it, it'll be a confusing-enough error message without being confusing > for the wrong reason. Attached, applies after the other two. Tony --- via perlbug: queue: perl5 status: open https://rt.perl.org/Ticket/Display.html?id=134193Thread Previous