unofficial mirror of bug-guile@gnu.org 
 help / color / mirror / Atom feed
* string-ci* oddity
@ 2008-12-19 14:55 Bill Schottstaedt
  2009-01-02  2:36 ` Thien-Thi Nguyen
  0 siblings, 1 reply; 3+ messages in thread
From: Bill Schottstaedt @ 2008-12-19 14:55 UTC (permalink / raw)
  To: bug-guile

Why do both Guile and Gauche give this result in string-ci<?
(and the other string-ci functions similarly):

guile> (char-ci<? #\a #\_)
#t
guile> (string-ci<? "a" "_")
#f

gosh> (char-ci<? #\a #\_)
#t
gosh> (string-ci<? "a" "_")
#f

The odd chars are ASCII 91 to 96:

guile> (string-ci<? "a" "[")
#f
guile> (char-ci<? #\a #\[)
#t






^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: string-ci* oddity
  2008-12-19 14:55 string-ci* oddity Bill Schottstaedt
@ 2009-01-02  2:36 ` Thien-Thi Nguyen
  2009-02-02  0:00   ` Thien-Thi Nguyen
  0 siblings, 1 reply; 3+ messages in thread
From: Thien-Thi Nguyen @ 2009-01-02  2:36 UTC (permalink / raw)
  To: bug-guile

() "Bill Schottstaedt" <bil@ccrma.Stanford.EDU>
() Fri, 19 Dec 2008 06:55:21 -0800

   Why do both Guile and Gauche give this result in string-ci<?
   (and the other string-ci functions similarly):

   The odd chars are ASCII 91 to 96:

ASCII 91-96 lie between the two ranges A-Z and a-z.
One procedure smashes case up and the other down.
Smashing happens unconditionally.

(define (my-char-ci<? p q)
  (< (down p)) (down q))

Another more complicated (but arguably more correct) approach
would be to determine if one/both of the args are not smashable,
and entirely avoid smashing in that case.  Something like:

(define (smashable? c)
  (or (<= #\a c #\z) (<= #\A c #\Z)))

(define (my-char-ci<? p q)
  (if (and (smashable? p) (smashable? q))
      (< (down p) (down q))
      (< p q)))

Pseudoscheme (numeric operators don't actually take chars),
but you get the idea.

thi





^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: string-ci* oddity
  2009-01-02  2:36 ` Thien-Thi Nguyen
@ 2009-02-02  0:00   ` Thien-Thi Nguyen
  0 siblings, 0 replies; 3+ messages in thread
From: Thien-Thi Nguyen @ 2009-02-02  0:00 UTC (permalink / raw)
  To: bug-guile

FYI, below is the code that all ci comparison
funcs will be using in the next Guile 1.4.x release.

This means, for example:
(char-ci<? #\a #\_) => #f.

thi


___________________________________________________________________
// libguile/chars.c

#define ISLOWER(c)  (islower (c) ? (1 + c - 'a') : 0)
#define ISUPPER(c)  (isupper (c) ? (1 + c - 'A') : 0)

static int
ccmp_ci (int x, int y)
{
  int lx, ux = 0, ly, uy = 0, rv;

  rv = ((! ((lx = ISLOWER (x)) || (ux = ISUPPER (x))))
        ||
        (! ((ly = ISLOWER (y)) || (uy = ISUPPER (y)))))
    /* One or both do not satisfy `isalpha'; subtract directly.  */
    ? (x - y)
    /* Both satisfy `isalpha'; subtract in one domain or another.  */
    : (lx
       ? (lx - (ly
                ? ly
                : uy))
       : (ux - (uy
                ? uy
                : ly)));
  return !rv
    ? 0
    : (GOOD (rv)
       ?  1
       : -1);
}





^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2009-02-02  0:00 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-12-19 14:55 string-ci* oddity Bill Schottstaedt
2009-01-02  2:36 ` Thien-Thi Nguyen
2009-02-02  0:00   ` Thien-Thi Nguyen

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).