Front page | perl.perl5.porters |
Postings from December 2012
On Wed, Dec 19, 2012 at 10:06:04PM +0100, Leon Timmermans wrote: > On Wed, Dec 19, 2012 at 2:55 PM, Father Chrysostomos via RT > <perlbug-comment@perl.org> wrote: > > That logic probably needs to be revisited then. Is the s/// target a > > COW before the substitution? (You don???t need to CC me.) > > No, it's a SvLEN == 0 string, with set magic as fallback to ensure the > SvPVX keeps pointing into the memory map. The fallback however is > wildly inefficient as it needs to copy the copy, hence the warning. > It's there to prevent data-loss and segfaults, not to be used > intentionally. SvLEN == 0 has a historical established meaning of "the core doesn't own this buffer". If COW were disabled for all PVs with SvLEN == 0, does this solve the issues in more than just this module? Nicholas ClarkThread Previous | Thread Next