[r6rs-discuss] Case sensitivity

[Joe Marshall]

On Mon, Feb 23, 2009 at 4:37 PM, John Cowan <cowan at ccil.org> wrote:
> Joe Marshall scripsit:
>
>> The biggest issue I had was that when PLT switched, I had to go
>> and edit a bunch of legacy code.  Unfortunately, calls to READ in
>> PLT Scheme used the current dynamic value of the case-sensitivity switch,
>> rather than the lexical value apparent at the call site, so I had to go
>> find and edit all the calls to READ as well.
>
> That's a very interesting point: it's not obvious to me that READ should
> behave so.  Just because my code is written in UTF-8, doesn't mean that
> by default I want all my text files to be opened as UTF-8.

The other option, where READ happens to use whatever dynamic value
the switch was last set to is clearly wrong.  READ is going to need a
case argument.  (And so is WRITE)

> Presumably by the time macro expansion is being done, all
> external representations have already been internalized, so
> (assuming down-folding) any external IF is already |if|.

Suppose I have a non-hygienic macro that introduces an identifier
named `Foo' and you wish to use it within a case-insensitive file.
If your external code has already been down-folded, it will never
be able to access the binding.



-- 
~jrm