Re: Some comments on the ansi test suite

>>>>> "Peter" == Peter Van Eynde <[email protected]> writes:

    Peter> name
    Peter> ...
    Peter>    5. n. (of a character) a string that names the character and that has
    Peter>    length greater than one. (All non-graphic characters are required to
    Peter>    have names unless they have some implementation-defined attribute
    Peter>    which is not null. Whether or not other characters have names is
    Peter>    implementation-dependent.) 

Ok, I missed that entry.  Guess we'll have to come up with some names
for CMUCL.
   
    Peter> as (code-char 65) doesn't have a name I fear we break the standard, not?

(code-char 65) is #\A and is mentioned in the section on char-name as
returning NIL as one possible option.  I assume you meant some other
character?

I note though, that CMUCL uses #\linefeed for ASCII LF.  I think it
should probably be #\Newline instead, since CMUCL is still Unix only.
I guess that would change if CMUCL ran on a Mac where the newline
character is CR.

    >> o What is ALLTEST-LEGACY- 1045, 1061, and 1073 testing?  These fail
    >> for me on CMUCL/Sparc, but presumably work on x86.  So what is being
    >> tested?  That it computes those functions correctly?

    Peter> These are part of the less useful history of ansi-test. It came from an
    Peter> implementation that seems to have had its share of numeric bugs. So
    Peter> there are a lot of simple tests like this. Of course they fall on their
    Peter> face the moment you change fp format... 

What do you propose?  Leave them?  Remove them?  Change them to be be
an approximately equal type of test instead of equality?

    >> o ALLTEST-LEGACY-1666 examines the printed value of
    >> most-positive-double-float.  What is the purpose of the test?  That
    >> it can print it out?  That it prints it out correctly?

    Peter> I suppose both. The value of tests like that is pretty low. However
    Peter> think about the 
    Peter> * (/ 1 (/ 3 2) 0)

    Peter> 1/0

    Peter> bug in cmucl...

This is fixed in CVS, as you probably know.

    Peter> In short: please commit fixes (with comments as to why) if you have the
    Peter> time and inclination.

Ok, but I think I'll send my patches here first, since I don't
understand all the implications.

Ray


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Stuff, things, and much much more.
http://thinkgeek.com/sf