Front page | perl.perl5.porters |
Postings from April 2003
On Tuesday, April 22, 2003, at 03:50 AM, Enache Adrian wrote: > You just discovered it now ? > > That's an old story > > http://nntp.perl.org/group/perl.perl5.porters/73394 > http://nntp.perl.org/group/perl.perl5.porters/73931 It LOOKS like an old story and very likely related, but it is DEFINITELY DIFFERENT. I did apply your patch in 1.92 and I thought this problem was gone at least in front of me as Encode maintainer since it did go OK on FreeBSD + thread. > > bleedperl t/enc_module.t > 1..3 > ok 1 - encoding vs. STDOUT > ok 2 - encoding vs. STDIN - 1 > ok 3 - encoding vs. STDIN - 2 > > bleedperl -V > [snip] > Characteristics of this binary (from libperl): > Compile-time options: DEBUGGING MULTIPLICITY USE_ITHREADS > USE_64_BIT_INT USE_LARGE_FILES PERL_IMPLICIT_CONTEXT > Locally applied patches: > DEVEL18374 > Built under freebsd > Compiled at Apr 7 2003 23:07:10 . Back then relocating "close STDOUT;" to the end of the script solved the problem but this time it did not help so it is different -- though very likely related deeply. > Run my little test from then with FreeBSD 4.8 ( not with 4.4 or Darwin > ) > and see how it works. Your script worked as you said on both OLD and LATEST bleedperl. > I suggested you to put > > close STDOUT unless $^O eq 'freebsd'; And I instead relocated "close STDOUT" AFTER the tests so this will not do any good this time. Dan the (Encode Maintainer|FreeBSD thread novice)Thread Previous | Thread Next