Re: [Sbcl-devel] THE declaration, TYPEP, FUNCTION type specifier

[Nikodemus S. <nik...@ra...>]

2009/3/27 Gábor Melis <me...@re...>:

> This is a regression from "1.0.19.3: more careful PROGV and SET".

Here's a proposed fix. Add STRICT argument to %%TYPEP and use
%FUN-TYPE + CSUBTYPEP to check the type if STRICT is not specified
(STRICT is the default, of course.) Functions which pass that test
(and only function which pass that test) should agree with what the
compiler believes is the type of the variable.

Types turn my brain to mush pretty quickly, so double-checking my
thinking would be appreciated.

Cheers,

 -- Nikodemus