gcl-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gcl-devel] Re: Proposed GCL changes re *FEATURES*, READ, and #+


From: Camm Maguire
Subject: [Gcl-devel] Re: Proposed GCL changes re *FEATURES*, READ, and #+
Date: 24 Nov 2003 11:51:36 -0500
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2

Greetings, and thanks for your report!

I'm cc'ing this to our ansi expert.  Paul, do you have any comments?

Actually, we have modified the reader for Matt Kaufmann to be more
like Allegro, so providing we know what the right thing to do is, a
change is quite feasible.

Take care,


Robert Boyer <address@hidden> writes:

> Here are two very closely related suggested changes to gcl for ANSI
> compliance.  Both changes are probably just a few minutes work for someone
> who fully understands READ's implementation, if such a person exists.
> 
> -------------------------------------------------------------------------------
> 
> Proposed change: The #+ read macro should do its READ with the current
> package being the KEYWORD package.
> 
> Justification: "While reading the test, the current package is the KEYWORD
> package."  From HyperSpec\Body\sec_2-4-8-17.html.
> 
> Proof that this is not being done correctly at the moment:
> 
> >*features*
> 
> (:ANSI-CL :COMMON-LISP :COMPILER :NUMLIB :CLOS-CONDITIONS
>     :PCL-STRUCTURES :PORTABLE-COMMONLOOPS :PCL :TURBO-CLOSURE-ENV-SIZE
>     :TURBO-CLOSURE :SDEBUG :DEFPACKAGE TRUNCATE_USE_C CLX-LITTLE-ENDIAN
>     BSD MC68020 BSD386 SGC IEEE-FLOATING-POINT UNIX GMP GCL AKCL COMMON
>     KCL)
> 
> >#+ gcl 23
> 
> 23
> 
> >
> 
> Also, I suspect that the #+ reader for gcl isn't actually just checking for
> symbol-names as opposed to symbols.  So this change should also enforce that
> read, when processing #+, is checking for symbols, not symbol-names.  I think
> this is another current non-ANSI compliance for gcl, since :gcl is not on
> *features* currently.  Consider:
> 
> >#+ :gcl 23
> 
> 23
> 
> -------------------------------------------------------------------------------
> 
> Proposed change: For any symbol s currently on *features*, the keyword symbol
> :s probably ought to be added to *features*.
> 
> Justification.  Why break old code unnecessarily?  Lots of users, including
> me, depend in old code upon features, in my case AKCL, being on features.
> 
> Bob
> 
> P. S.  My hats off to anyone with the balls to touch READ.  Even by 1970,
> Jonl White or others had named it in Maclisp the "hirsute" reader for its
> hairiness.
> 
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah




reply via email to

[Prev in Thread] Current Thread [Next in Thread]