Some thoughts on egcs 1.0.2

[H.J. Lu]

> 
> 	You are 100% right.  We can't expect any Linux distributor to
> base their product on egcs if it can't correctly compile the C
> library.  I don't think there is any disagreement on this.
> 
> 	However, part of being acceptable to Linux (or any platform)
> means stabililty and high quality.  So we have to walk the tightrope
> on the volatility of changes in 1.0.2, I'm sure you are aware of this.
> 

I am fully aware that. But if there is a real bug which prevents
egcs 1.0.2 from working correctly on linux/glibc 2, we should find
a way to fix it, even if that means big changes. That is part of what
I meant by "whatever it takes."

> 	We've managed to do this well for 1.0.1.  I'm confident we can

It seems that not many people have run math test on glibc 2.1 with
egcs 1.0.1. There is no math test in glibc 2.0, which hides the bug.

> get it right for 1.0.2.  And as you made abundantly clear, we have
> make it our top priority.
> 

Let's hope so.


-- 
H.J. Lu (hjl@gnu.org)