[Openmcl-devel] Is it safe to make hash tables with non-standard tests?

[Ron Garret]

EQUAL(P) and STRING= do not do the same thing.  STRING= accepts string designators.  EQUAL doesn't.

? (equal "FOO" 'foo)
NIL
? (equalp "FOO" 'foo)
NIL
? (string= "FOO" 'foo)
T

(Well, to be strictly correct, EQUAL(P) accepts string designators, but it does not compare them with string=.)

rg

On Jul 14, 2013, at 5:54 PM, Tom Emerson wrote:

> What CLtL1 says isn't really relevant in 2013. There also isn't a need to create a hash table for #'string= because that functionality is already supported by the standard.
> 
> So the statement in Ron's message that a 'string= hashtable [...] is not in the standard" is false, and there is no reason to use a non-standard approach when the standard covers both this case and a #'string-equal case.
> 
> So while my response doesn't directly answer his exact question, it answers it in a way that obviates the need for his original query.
> 
> 
> 
> On Sun, Jul 14, 2013 at 8:48 PM, Gary Byers <gb at clozure.com> wrote:
> If the question was "can TEST functions other than EQ/EQL/EQUAL/EQUALP
> be used non-portably in CCL, and should that be considered a supported
> extension that should be documented ?", the answer is "yes".  IIRC, in
> CLtL1 EQUALP wasn't a standard test function, and supporting arbitrary
> test functions (and user-provided hashing functions) was one way of
> supporting EQUALP as a test.
> 
> If the question was "is this a non-portable extenstion ?" ... I don't
> think that that was the question.
> 
> If the question was "is there a compelling reason to use a non-standard
> test function (other than typechecking and whatever the performance tradeoffs
> are at the moment), and is such a reason more compelling than portability
> concerns that one may or may not have ?" ... well, that's a good question
> to ask and one that different people can have good reasons to answer differently.
> 
> 
> On Sun, 14 Jul 2013, Tom Emerson wrote:
> 
> You don't need to: using 'equal as your test achieves the same result.
> Welcome to Clozure Common Lisp Version 1.9-r15757 ?(LinuxX8664)!
> 
> ? (defvar foo (make-hash-table :test 'equal))
> #<HASH-TABLE :TEST EQUAL size 0/60 #x3020006F2C3D>
> ? (setf (gethash "abc" foo) 100)
> 100
> ? (gethash "abc" foo)
> 100
> T
> ? (gethash "ABC" foo)
> NIL
> NIL
> 
> If instead you want the test to be string-equal then use 'equalp
> 
> ? (setq foo (make-hash-table :test 'equalp))
> #<HASH-TABLE :TEST EQUALP size 0/60 #x3020006F9D5D>
> ? (setf (gethash "abc" foo) 100)
> 100
> ? (gethash "abc" foo)
> 100
> T
> ? (gethash "ABC" foo)
> 100
> T
> 
> This is portable: same results in SBCL.
> 
> 
> 
> On Sun, Jul 14, 2013 at 3:53 PM, Ron Garret <ron at flownet.com> wrote:
>       I need a string= hash table, which is not in the standard. ?It
> 
>       appears that CCL can actually make one:
> 
>       (make-hash-table :test 'string= :hash-function 'sxhash)
> 
>       This seems to do the Right Thing, but ?it's not documented. ?Is
> 
>       it safe to use this feature?
> 
>       rg
> 
>       _______________________________________________
>       Openmcl-devel mailing list
>       Openmcl-devel at clozure.com
>       http://clozure.com/mailman/listinfo/openmcl-devel
> 
> 
> 
> 
> --
> Tom Emerson
> tremerson at gmail.com
> http://www.dreamersrealm.net/tree
> 
> 
> 
> 
> 
> -- 
> Tom Emerson
> tremerson at gmail.com
> http://www.dreamersrealm.net/tree

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clozure.com/pipermail/openmcl-devel/attachments/20130714/2e7b673c/attachment.htm>