[Gcl-devel] Re: 2.7.0 make-package bug

[Camm Maguire]

Greetings! OK, this is in -- thanks!

Robert Boyer <address@hidden> writes:

> I hope that you'll have time for the following matter that we discussed in
> August while you're in the vicinity.  I'll be it takes you less than 5
> minutes.
> 
> Bob
> 
> 
> To: Robert Boyer <address@hidden>
> Cc: address@hidden, address@hidden
> Subject: Re: case insensitivity issue for package names
> References: <address@hidden>
> From: Camm Maguire <address@hidden>
> Date: 09 Aug 2005 01:45:28 -0400
> In-Reply-To: <address@hidden>
> Message-ID: <address@hidden>
> Lines: 52
> User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2
> MIME-Version: 1.0
> Content-Type: text/plain; charset=us-ascii
> X-SpamAssassin-Status: No, hits=-2.6 required=5.0
> X-UTCS-Spam-Status: No, hits=-232 required=180
> 
> Greetings, and thank you so much for this proposed fix!
> 
> It has been on our ansi todo list for some time to implement readtable
> case.  If this issue is truly orthogonal to the one below, I think
> your suggestion is fine -- otherwise perhaps it should be part of a
> more systematic case sensitization?  These comments are from someone
> who has not studied the issue yet at all :-).
> 
> Take care,
> 
> Robert Boyer <address@hidden> writes:
> 
> > Matt notes that package lookup in all versions of GCL ignores case, and that
> > is wrong, e.g.,
> > 
> >   (make-package "foo" :use nil)
> >  followed by
> >   (find-package "FOO") 
> >  yields  
> >   #<"foo" package>
> > 
> > I believe this bug is easily fixed by changing the one file o/package.d in
> > the following ways:
> > 
> > 1.  Replace all occurrences of member_string_equal with member_string_eq.
> > 
> > 2.  In the function designate_package, replace the occurrence of 
> > string_equal
> >     with string_eq.
> > 
> > The use of the identifier "string_eq" is less than perfect aesthetically.  
> > It
> > would better be "string=", but probably that is not a legal C identifier.
> > I'm just following what's already defined as string_eq in o/string.d.
> > 
> > Of course, the topic of packages is a horrendously complicated matter and a
> > fix might be a lot more complicated than what I have proposed.  But I 
> > suspect
> > that since we're not going very close to INTERN, USE, or SHADOW, the fix
> > might be this easy.
> > 
> > Bob
> > 
> > P. S.  After you have made the fix above, but before you try to rebuild GCL,
> > you need to fix another bug in the GCL sources in the following way.  In the
> > files "unixport/init_ansi_gcl.lisp" and "unixport/init_ansi_gcl.lisp.in",
> > replace (in-package "lisp") with (in-package "LISP").
> > 
> > 
> > 
> 
> -- 
> Camm Maguire                                          address@hidden
> ==========================================================================
> "The earth is but one country, and mankind its citizens."  --  Baha'u'llah
> 
> 
> 
> 

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