Re: [MacPerl-WebCGI] well split!! (Solved)

At 8:13 PM -0500 3/7/01, [email protected] wrote:
>>I think you may have got it working by fixing the symptom rather than the
>>problem.
>>
>>It seems that your while(<DATA>) part is still reading the whole file in in
>>one go... this is most likely because you have the wrong line endings on the
>>file for that OS. you are then "fixing" this by using the split(/\r/) part
>>which splits the whole file content into lines at  \r (this is what you
>>wanted the while(<DATA>) to do in the first place).
>
>I'm not doubting you, but doesn't the first condition "\r" read the 
>data file and dump the first segment (i.e., start-of-first-record to 
>first-return) into

[snip]

>	Apache/1.3.3 Cobalt (Unix) (Red Hat/Linux) PHP/3.0.7
>
>Now, my data file is configured to have a <ret> at the end of each 
>record. I realize that a <ret> means different things to different 
>environments. All I know is that if I look at the ASCII values of 
>the data text file with a HEX editor, each record ends with 013 
>(DEC) (00D (HEX)) (015 (OCT) which is a CR on the Mac.
>
>With that said, should I be using a different EOR delimiter? After 
>all, I can make it anything I want.
>

When you "configure the data file to have a <ret> at the end of each 
record", how are you doing this? By using "\r"? If so, that's the 
source of the whole problem you've been having with

   while (<DATA>)

Either

a) use "\n", which in Perl is whatever is used for the OS's native 
end of record/end of line, which could be \r\n, \n, \r in the usual 
meaning of those guys. (See Adam Whitney's references.)

or

b)  set $/ locally to "\r" before you get to "while (<DATA>)".


  {
local $/ = "\r";
while (<DATA>) {

    # ...
    }
  }

HTH

1;

-- 

   - Bruce

__bruce_van_allen__santa_cruz_ca__