Subject: Re: (mips): egcs on -current miscompiles strtod.c
To: Jonathan Stone <jonathan@DSG.Stanford.EDU>
From: Niklas Hallqvist <>
List: tech-toolchain
Date: 02/06/1999 00:47:59
> Date: Fri, 05 Feb 1999 14:36:22 -0800
> From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
> Using a union instead of the pointer arithmetic fixes strtod for me.
> Thanks very very much for the tip.

Good to hear it helped.

> Do you know if it's safe to enable Haifa with egcs on mips now?
> Trouble with strtod (which looks like the same bug) was the only thing
> I'd found where Haifa broke the snapshot in the netbsd tree as of
> early November (ie, whatever it was before we upgraded to 1.1.1).

I have no idea, sad to say... My mips' SCSI bus has been dead since
september or so :-(  I have finally found some donators of VL BT445S
cards (yeah, VL is crappy, but that is what my Tyne boots from), so my
ARC box will probably live again in a few weeks, but I bet you have
had time to test Haifa again well before that.

This construct was clearly something aggressive schedulers could trip
upon, so I am not surprised to hear Haifa bit the dust right there too.