This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Move new totalorder from libm to libc
- From: Florian Weimer <fw at deneb dot enyo dot de>
- To: "Gabriel F. T. Gomes" <gabriel at inconstante dot net dot br>
- Cc: <libc-alpha at sourceware dot org>
- Date: Tue, 17 Dec 2019 20:53:40 +0100
- Subject: Re: Move new totalorder from libm to libc
- References: <87tv62daqi.fsf@mid.deneb.enyo.de> <20191216140714.558e1313@tereshkova> <87o8w8192v.fsf@mid.deneb.enyo.de> <20191217115527.22e08b95@tereshkova>
* Gabriel F. T. Gomes:
> Do you think that the runtime crash (when the wrong version of totalorder
> gets selected) is worse than the runtime difference caused by wrong
> versions being selected for exp, log, pow (as Andreas mentioned)?
If I said there would be crash, I think I was wrong. The function
selected is the pointer-less variant, and that would just produce
wrong results and not crash. So it's as hard to debug as the long
double mismatch.
However, it should matter far less because it's an obscure function.