Sorry about that...forgot to reply to all.  Working with 4 languages at
once tends to turn my brain to mush.

-----Original Message-----
From: Bob Showalter [mailto:[EMAIL PROTECTED]]
Sent: Monday, June 24, 2002 2:41 PM
To: 'Andrew Stone'
Subject: RE: Binary File Pattern Matching


Send this to the list. I really don't want to keep private threads going.
Thanks.

> -----Original Message-----
> From: Andrew Stone [mailto:[EMAIL PROTECTED]]
> Sent: Monday, June 24, 2002 2:26 PM
> To: Bob Showalter
> Subject: RE: Binary File Pattern Matching
> 
> 
> Bob,
> 
> Thanks, and for some more info, the file is all integer based. So
> I know that the ASCII string scenario doesn't come into play.  These
> files are all generated by FORTRAN where everything (even the 
> characters)
> is stored as INTEGER*4.  That should help me then, correct?
> 
> thanks for the help,
> andy
> 
> -----Original Message-----
> From: Bob Showalter [mailto:[EMAIL PROTECTED]]
> Sent: Monday, June 24, 2002 2:19 PM
> To: 'Andrew Stone'; [EMAIL PROTECTED]
> Subject: RE: Binary File Pattern Matching
> 
> 
> > -----Original Message-----
> > From: Andrew Stone [mailto:[EMAIL PROTECTED]]
> > Sent: Monday, June 24, 2002 1:29 PM
> > To: [EMAIL PROTECTED]
> > Subject: Binary File Pattern Matching
> > 
> > 
> > Hello all,
> > 
> > I'm having a problem with (I suppose) pattern matching.
> > Here's the task: To take binary data from windows(little-endian)
> > and convert to Unix(big-endian).  
> 
> Well, Perl's pack() and unpack() functions can help with that.
> 
> > If the file is all numeric, I don't
> > have a problem, but it isn't.  I need to be able to recognize that
> > one of the four bytes is non-numeric data (spaces, alpha, 
> > punctuation..).
> > If one of the four is non-numeric then I don't want to do the byte
> > swapping.  Right now if there is character data it gets 
> > swapped so 'TEST'
> > becomes 'TSET'.  I've been searching the web, but haven't 
> > found anything
> > that works for me.
> 
> This really can't be done. The sequence 'TEST' in the file may be
> an ASCII string with the word test. Or, it may be the long integer
> 1413829460 packed in network (big-endian) order. Or something else
> altogether. There is no way to say which it is without knowing
> something about the structure of the file.
> 
>    $ perl -le 'print pack "N", 1413829460'
>    TEST
> 

-- 
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to