[ILUG] new gcccpuopt available

[Kieran.Tully AT acm.org]

On 04/06/05, Brian Foster <blf at blf.utvinternet.ie> wrote:

>  |  > there is the curious belief that expr(1) does FP
>  |  > (floating point) math, or at least FP comparisons.
>  |
>  | No there isn't.
>  | sed is used to manipulate gcc_version so that
>  | FP ordering is respected by the lexicographic
>  | comparison.
> 
> YES there is.  in v0.3, there is a comment before
> the sed(1) saying:
> 
>    #manipulate version so can do floating point comparisons
> 
> ergo, there is what I said:  a "curious belief".
> the comment proves that at one time, someone did
> think something — presumably expr(1) — did FP.
> 
> does this matter?  Yes!  a casual reader of the
> script could easily be mislead.  comments matter.

I'm not going to speculate on anyone's beliefs (curious or otherwise :).
So I won't speculate on how you are interpreting the comment.

But to me that comment says 
 "we need to munge gcc_version to achieve the effect of a FP comparison".  

Nowhere is it suggested that expr on its own can do FP comparisons.
Quite the contary; sed is only used because the author is aware of
the expr limitation, and without the sed, I can't see the need for the
comment in question. 

-- 
Kieran Tully, Software Developer and Tenor
Reply to Kieran.Tully AT acm.org
http://kieran.tul.ly  http://www.cs.tcd.ie/~tullyka