Re: [PATCH 5.5.670] pseudohash, one more try

On Thu, 02 Mar 2000 05:31:36 EST, John Tobey wrote:
>Okay.  I've put the compile-time check back in, this time in
>Perl_newASSIGNOP.  pp_aassign walks the op tree when an AV/AVHV is
>found and the op flag indicates a % expression exists somewhere on the
>left side.  Simple hash variables (%foo) are discounted because they
>cannot be pseudohashes.  I believe that false positives (op tree
>walked but no avhv) would be rare.

Haven't tried it, but this certainly looks much better than the
earlier ones.

Thanks.


Sarathy
gsar@ActiveState.com

• [PATCH 5.5.660] improved semantics for assign to pseudohash by John Tobey
• Re: [PATCH 5.5.660] improved semantics for assign to pseudohash by Michael G Schwern
• Re: [PATCH 5.5.660] improved semantics for assign to pseudohash by John Tobey
• Re: [PATCH 5.5.660] improved semantics for assign to pseudohash by Michael G Schwern
• Re: [PATCH 5.5.660] improved semantics for assign to pseudohash by Nick Ing-Simmons
• [PATCH 5.5.660] better avhv patch by John Tobey
• Re: [PATCH 5.5.660] better avhv patch by Gurusamy Sarathy
• [PATCH 5.5.670] pseudohash, one more try by John Tobey
• Re: [PATCH 5.5.670] pseudohash, one more try by Gurusamy Sarathy
• Re: [PATCH 5.5.660] better avhv patch by John Tobey
• Re: [PATCH 5.5.660] improved semantics for assign to pseudohash by Ilya Zakharevich
• Re: [PATCH 5.5.660] improved semantics for assign to pseudohash by Gurusamy Sarathy
• Re: [PATCH 5.5.660] improved semantics for assign to pseudohash by Ilya Zakharevich
• Re: [PATCH 5.5.660] better avhv patch by M.J.T. Guy
• Re: [PATCH 5.5.660] better avhv patch by Joshua N Pritikin
• Re: [PATCH 5.5.660] better avhv patch by Gurusamy Sarathy
• Smoke [5.8.1] 20803 PASS solaris 2.6 (sun4/1 cpu) by Alan.Burlison