On Thu, Dec 30, 2004 at 03:21:02PM -0800, Yitzchak Scott-Thoennes wrote: > On Mon, Dec 20, 2004 at 10:36:23AM -0800, Yitzchak Scott-Thoennes wrote: > > On Mon, Dec 20, 2004 at 03:57:48PM +0100, Rafael Garcia-Suarez wrote: > > > H.Merijn Brand wrote: > > > > > > > > "Give this a whirl" now leads to you guys having tested it? > > > > I can aply, but without any testing or feedback, I feel kinda dangerous > > > > > > Please don't, I just glanced at it yet and I'm not yet sure it works :p > > > > Can we agree in principle on the tests? (That the error should apply > > to win32, but that even where there is no real lstat, perl should keep > > track of whether stat or lstat was used.) > > Ping. > > Has anyone on win32 had the chance to try the patch in: > http://nntp.perl.org/group/perl.perl5.porters/96987 I note that several other places (properly, IMO) use PerlLIO_lstat without checking HAS_LSTAT: pp_hot.c in pp_readline, to not return unmatched wildcards from glob(); doio.c in my_lstat, used by -l filetest, and in apply, to forbid root from unlinking directories without -U; and util.c in getcwd_sv.Thread Previous | Thread Next