|
From: | Ralf Hemmecke |
Subject: | [Axiom-math] Re: [open-axiom-devel] [fricas-devel] Re: [Aldor-l] Type equivalence of domains in Axiom and Aldor |
Date: | Fri, 23 Nov 2007 18:40:04 +0100 |
User-agent: | Thunderbird 2.0.0.9 (X11/20071031) |
On 11/23/2007 06:04 PM, Bill Page wrote:
Thanks, Ralf. I like your version better than mine. It makes more sense than the apparently redundant syntactic change I proposed ... although I wonder why we should define a local function rather than just write it inline. Would you agree however that that fact that Saul's original variant does not compile should be considered a regression bug in the Aldor compiler?
Yes and no. Suppose you want to declare ---BEGIN aaa.as #include "aldor" define Cat: Category == with { foo: String -> Integer; foo: (s: String) -> Integer; } ---END aaa.asThe types of both foos are different. The question now is, would you like Cat to export 2 or only 1 function?
Now there is a very special case... Would you like X and Y X: Record(x: A); Y: Record(y: A); be of the same type?So I cannot really say yes or no to your question and would rather be happy to hear more information from the people who changed the compiler behaviour.
What is the background? And (more importantly) what is the semantics of "Cat" according to the language definition. I cannot say whether it exports one or two functions.
Ralf
[Prev in Thread] | Current Thread | [Next in Thread] |