Re: [Mingw-cross-env-list] Switch from i386 to i686?

[Volker Grabsch]

Mark Brand <address@hidden> schrieb:
> >As nobody reported a demand for pre-i686 compatibility, I updated
> >the project's target triplet to i686-pc-mingw32.
> >http://hg.savannah.gnu.org/hgweb/mingw-cross-env/rev/7529d9e111fe
> 
> There was an official sounding recommendation from the Mingw-users list 
> on 24/9 to use simply "mingw32" as the prefix.

No, it was stated explicitly as "personal taste":

    | I compile my own MinGW cross-compiler from source; I invoke
    | it as `mingw32-gcc ...'.  Some may choose to add the `i386'
    | (or equivalent `i?86') CPU designator and `pc' VENDOR identifier,
    | [...]

However, using "mingw32" as _OS identifier_ was stated as official
view, and I fully agree:

> >The official view of the MinGW project, (and this
> >is an authoritative observation), is that the OS identifier should
> >be specified *exactly* as `mingw32'. 

The remaining question is what the CPU and the VENDOR identifiers
should be.

I can't find any official MinGW view on them, so I adjusted the
CPU identifier to the real need (i686) and the VENDOR identifier
to the only VENDOR that makes sense (pc).

I don't think it's a good idea to leave them out, as they are
strongly recommended in the GNU manuals (Autoconf, maybe also GCC).
Maybe we could leave out the VENDOR indentifier, but I doubt we
could really leave out the CPU indentifier.

Maybe it would be interesting to know what the 64-bit MinGW people
(http://mingw-w64.sourceforge.net/) use as OS, CPU and VENDOR identifier.

> Does including "i686" in the prefix 
> actually cause i686 optimizations to be used? I used to assume it did, 
> but now I wonder.

I'm pretty sure it does. However, I haven't tested it yet.


Greets,

    Volker

-- 
Volker Grabsch
---<<(())>>---
Administrator
NotJustHosting GbR