"B. K. Oxley (binkley)" wrote:
>
> From: "Brian Ingerson" <[EMAIL PROTECTED]>
>
> > I've read this mini-thread over a couple times and am frankly
> confused.
> > Could you boil this problem down into a concise paragraph. Preferably
> > with a patch :)
>
> Sure.
>
> Problem #1: messed up line-ending handling. At least in support of
> Cygwin, you need to follow the advice in perlport and expect $CR?$LF
> instead of \n to terminate lines in the __DATA__ section. Otherwise,
> Inline::read_DATA gets confused by \r\n terminators. I sent that patch
> in an earlier message. This still leaves unaddressed the larger
> question of line ending handling in general (e.g., Macintosh uses just
> \r rather than \n or \r\n), and I don't have a good answer for that.
Thanks for the summary. I wasn't sure if the patch was still valid after
all the different assertions. I'll put out a patched version on the
website. Sometime... (busy)
> Problem #2: With the line ending problems, I get different error
> messages depending on which I am using one language (e.g., __Foo__) or
> two (e.g., __Foo__ and __C__) in the __DATA__ handle. The latter is a
> better message, IMHO. I have no patch for this and haven't looked into
> it too much.
I think I lost some functionality when I "upgraded" the read_DATA
routine. This is exactly the reason I need a better test harness :)
I'll look into this in detail sometime after YAPC. If you want to put
together a test harness script or two for validating all the possible
conditions, that would be great. Look at t/*.t for starters. If you just
want to send a tarball of failure scripts, that'd be good too.
BTW, Inline::Files probably suffers the same fate, because IIRC I stole
that regex from Damian.
Damian, I don't want to be the straw that broke the Camel's back, but
maybe you should send a blank email to [EMAIL PROTECTED]
You are an Inline conspirator now :)
Cheerio, Ingy Hood (and my Merry (wo)Men)
--
perl -le 'use Inline C=>q{SV*JAxH(char*x){return newSVpvf
("Just Another %s Hacker",x);}};print JAxH+Perl'