[perl #123029] 5.20.x regression in "state" under PERL_NO_COW

On Tue Oct 21 16:19:36 2014, sprout wrote:
> I’ll write a separate maint patch on a branch.

Please review (and maybe even vote on backporting) the sprout/maint-5.20-123029 branch, which contains two fixes:

1) Don’t swipe short PADTMP buffers when COW is disabled.  That was
   never intended to happen, and it is a regression, in that it
   causes a significant slowdown.
2) Check the flags properly and don’t steal buffers from state vars.

-- 

Father Chrysostomos


---
via perlbug:  queue: perl5 status: open
https://rt.perl.org/Ticket/Display.html?id=123029

• [perl #123029] 5.20.x regression in "state" under PERL_NO_COW by Father Chrysostomos via RT
• Re: [perl #123029] 5.20.x regression in "state" under PERL_NO_COW by James E Keenan
• Re: [perl #123029] 5.20.x regression in "state" under PERL_NO_COW by Aaron Crane
• [perl #123029] 5.20.x regression in "state" under PERL_NO_COW by Father Chrysostomos via RT
• [perl #123029] 5.20.x regression in "state" under PERL_NO_COW by Father Chrysostomos via RT
• [perl #123029] 5.20.x regression in "state" under PERL_NO_COW by Aaron Crane via RT
• [perl #123029] 5.20.x regression in "state" under PERL_NO_COW by Aaron Crane via RT
• [perl #123029] 5.20.x regression in "state" under PERL_NO_COW by Aaron Crane