RE: [PATCH] Remove log2 define

Tamar Christina <[email protected]>
Newsgroups gmane.comp.lib.newlib
Message-ID <DB6PR0802MB23096BFD5A17E10A7F3DA156FF020@DB6PR0802MB2309.eurprd08.prod.outlook.com>
> -----Original Message-----
> From: [email protected] <[email protected]> On
> Behalf Of Corinna Vinschen
> Sent: Monday, August 27, 2018 11:45
> To: Wilco Dijkstra <[email protected]>
> Cc: [email protected]; nd <[email protected]>
> Subject: Re: [PATCH] Remove log2 define
> 
> On Aug 24 16:04, Wilco Dijkstra wrote:
> > Corinna Vinschen wrote:
> >
> > >> Since newlib now has a fast implementation of log2,
> > >
> > > Only if !__OBSOLETE_MATH
> >
> > Is there any reason keep that old code, particularly for the rewritten
> > double precision functions? We now have a log2 that is both fast,
> > accurate and correct, which isn't true for the obsolete version.
> 
> If you want to make it default for all targets, I guess you can go ahead and
> send patches to remove the obsolete code.
> 

Should that action actually block this one? We've had cases where this optimization isn't safe
when the user compiles this header and their code with fast-math.  So I'd say there's an actual
bug here as well.

Regards,
Tamar.

> Corinna
> 
> --
> Corinna Vinschen
> Cygwin Maintainer
> Red Hat
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.