On Jan 29, 2004, at 7:15 AM, Nick Ing-Simmons wrote:
Brad Guillory <[EMAIL PROTECTED]> writes:Last spring someone committed a patch to fix the tests on windows platforms (see Change 18966 by [EMAIL PROTECTED] on 2003/03/14 04:20:51). This broke the tests on my Redhat box. Here is a compromise patch:
--- t/enc_module.t.orig 2004-01-28 11:34:27.000000000 -0600 +++ t/enc_module.t 2004-01-28 11:34:42.000000000 -0600 @@ -54,7 +54,7 @@ $obj = Mod_EUCJP->new; my $i = 0; while(<STDIN>){ - s/\r?\n\z//; + s/\r?\n\z?//;
Unless I am more confused than I think I am an optional zero-width assertion is same as no assertion at all. So that is equivalent to s/\r?\n// Was \z meant to be ^Z (DOS EOF marker)?
is ($cmp[$i++], $_, "encoding vs. STDIN - $i"); }
Also someone may want to put in the docs or in the fail report that LC_ALL must be set during "make test"
Thanks, BMG
The \z is from the original test file. It did not match on my machine so I made it optional. I should note that the s// used to be a chomp which worked fine for me, Change 18966 by [EMAIL PROTECTED] on 2003/03/14 04:20:51 that I mentioned in my original message changed chomp to s/\r?\n\z//.
Looking at the email address that I sent this to I should also mention that this change is for the Encode package against Encode-1.99
Thanks again, BMG
