[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Reopening this bug



Bdale Garbee <bdale@gag.com> writes:

> 
> Aha.  Smoking gun.  I don't know exactly what the issue is/was, but a while
> back (first of the year or so) I spent an evening on IRC with someone figuring
> out that the problem was something about the compiler, maybe an optimizer

Ahh, I got this from Raul, thought he had talked to you:

Raul Miller <moth@debian.org> writes:

> Thanks to Joel, I think I have a solution for this bug (just use
<math.h>
> instead of <nan.h>), but I don't know how to test it.


I told him about faure and lully.

Incidentally, I run the Alpha autobuilder, so it was built on my
system :-)  If this turns out to be the problem, I would suggest
applying a patch in your package so that this same problem doesn't
recur everytime the package gets rebuilt on alpha.

Thanks,
John

> problem, I really don't recall, when the package was built for alpha by the
> auto-builder.  They built a version that worked, and uploaded it... my guess
> is that my uploading a new version triggered the same but in the auto
> builders?  Sorry, I don't remember who was helping chase this, it's been a
> while.
> 
> Can you try recompiling from the source package on your system and seeing if
> the problem persists?  I'll CC debian-alpha on this in case someone else wants
> to jump in with some insight.
> 
> I'm on my way out of town (to Toronto) for the weekend, but will be happy to 
> help chase this down starting Monday night, if need be.
> 
> Bdale

-- 
John Goerzen   Linux, Unix consulting & programming   jgoerzen@complete.org |
Developer, Debian GNU/Linux (Free powerful OS upgrade)       www.debian.org |
----------------------------------------------------------------------------+
The 250,983rd digit of pi is 8.


Reply to: