Re: 1.4 build problems (sorry!)

[Alex Shinn]

>>>>> "Peter" == Peter C Norton <address@hidden> writes:

    Peter> On Sat, Feb 23, 2002 at 10:06:20AM +0900, Alex Shinn wrote:
    >> Well, they call it the end-of-record separator, since you're
    >> allowed to make records not depend on newlines.

    Peter> Its IFS for Internal Field Separator, and it comes from
    Peter> shell, but it does more then the shell's $IFS.

Well, actually, it's $INPUT_RECORD_SEPARATOR (when use English is in
effect), and is completely different from IFS.  Basically, records are
the lines you read in, fields are what you split the records into.

There's actually no notion of IFS in Perl - split defaults to awk's
behaviour of splitting on whitespace (same as with the -a option),
otherwise you specify an explicit regexp (as with the -F option).

    >> You're better off with an existing template system (autogen works
    >> with Guile), or if you're going to write your own at least use a
    >> proper parser (try SXML).

    Peter> There wasn't and maybe still isn't an xml parser that I could
    Peter> find for guile.  Is expat+guile being done by anyone? What's
    Peter> the syntax?  Writing a dtd for the crappy format I put
    Peter> together should be a no-brainer.

SXML is a nice native Scheme XML parser that I've used with Guile.  If
you're concerned about speed, I believe SXML actually outperforms the
native C expat and fxml, probably because as implemented they call
malloc for every node.

  http://okmij.org/ftp/Scheme/SXML.html

-- 
Alex