Re: performance patch in LDIF.pm

2006-11-11 Thread Graham Barr
On Nov 11, 2006, at 2:37 AM, Chris Ridd wrote: Good point. The LDIF spec's got this text, BTW: --- discarded. Implementations SHOULD NOT fold lines in the middle of a multi-byte UTF-8 character. --- Does _wrap do that correctly? Hm, yes and no. _wrap does no special t

Re: performance patch in LDIF.pm

2006-11-11 Thread Chris Ridd
On 10/11/06 1:32, Graham Barr <[EMAIL PROTECTED]> wrote: > Except it is wrong. It does not honor the len argument > > ($x = _wrap("a" x 200, 40)) =~ s/ /./g; > print "$x-\n"; > > > aa > .

Re: performance patch in LDIF.pm

2006-11-09 Thread Graham Barr
On Nov 6, 2006, at 12:24 PM, Chris Ridd wrote: old _wrap() - sub _wrap { if($_[1] > 40) { my $pos = $_[1]; while($pos < length($_[0])) { substr($_[0],$pos,0) = "\n "; $pos += $_[1]+1; } } $_[0]; } new _wrap()

Re: performance patch in LDIF.pm

2006-11-06 Thread Chris Ridd
On 24/10/06 10:20, Robert van Hulsteijn <[EMAIL PROTECTED]> wrote: > Distribution: perl-ldap-0.33 > Perl: 5.8 > OS: Windows 2000 > > Hi, > I found that writing an LDIF record with very long attribute values is > extremely slow. I could narrow this problem down to the function _wrap() > in the LDI