This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Updating top-level autoconf to 2.59


On Thu, Feb 08, 2007 at 11:20:48PM -0800, Michael Eager wrote:
> 
> >Per autoconf docs, --host is supposed to mean --cross.  --native is
> >the absence of --host.  No room for inconsistencies in options.
> 
> If it worked correctly, I doubt there would be much discussion.
> It seems that configure is always trying to guess whether the
> build is native or host.

Errm, if there is need for guessing beyond native detection by
./config.guess, then the user's command line options are inexplicit,
surely? (And if there's no need for guessing, it can only be helpful if
none is done.) Doubtless, my experience is a subset of what configure
needs to do to aid all users, but FWIW that says:

If neither --host nor --target are specified, we're building native.

If both --target and --host default to native, then specifying one or
both creates a single or double cross. 

--build seems to be a synonym for --host, so its only useful purpose
that I've imagined so far is to specify a native build if configure
fails to detect that to the user's satisfaction. (And that could be done
with --host, if that did not spuriously trigger a cross.)

(If there's a need to build a native compiler as if it were a cross,
then --host and --target would only need to be specified as native.)

More guessing than that wouldn't help this cross compiler builder.
(On PowerPc, AVR, and V850, the required fudges haven't been too bad so
far, IIRC. But then I'm a few tool releases behind.)

'Avagoodweekend,

Erik


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]