From: Hugh Sasse <hgs@...>
Date: 2006-11-21T00:16:06+09:00
Subject: Re: Ruby 1.[89].\d+ and beyond.

On Mon, 20 Nov 2006, URABE Shyouhei wrote:

> Hugh Sasse wrote:
> > #define RUBY_VERSION_MAJOR 1
> > #define RUBY_VERSION_MINOR 6
> > #define RUBY_VERSION_TEENY 8
        [...]
> Please.  Don't forget RUBY_VERSION_CODE.  Browse the Google Code Search
> (or whatever you please) to remember how it's impossible to extend
> _MINOR and _TEENY beyond 9.
> 
> http://www.google.com/codesearch?q=%22%23if+RUBY_VERSION_CODE%22
> 
> I think RUBY_VERSION_CODE is one of the biggest mistake we made.  But
> whether that was a mistake or not it's so widely used.  Breaking
> compatibility to this C-level API come to no good.

Ouch.  The only solution to that that I can think of would be 
to make the constant float, and add decimal places.  I've forgotten
what happens about coercion in C ... After poking around the net a
bit I think that would work, but it would still be horrid.  It would
mean adding TEENY_TEENY or some such balderdash to the names.

brains hgs 16 %> ./silly_version
comparison workedbrains hgs 17 %>
brains hgs 17 %> cat silly_version.c
#include <stdio.h>

int main(int argc, char *argv[])
{

  if (180 < 180.5)
  {
    printf("comparison worked");
  }else{
    printf("comparison FAILED");
  };
  return(0);
}
brains hgs 18 %>

Such version numbers should still do the right thing in the string case.

        Hugh
> 
>