Re: config.guess includes darwin kernel major.minor.patch

[Ben Elliston]

On Mon, Feb 29, 2016 at 02:06:03PM -0600, Patrick Mahoney wrote:

> Not sure if this is a bug or intended behavior, but on Darwin
> config.guess includes the output of `uname -r`, which seems to be
> the major.minor.patch version of the darwin kernel.

It's not unusual, but there is no convention. It depends on the system
type being guessed.

> This test is too aggressive, since it means a program (execline)
> cannot be compiled against a library (skalibs) that was itself built
> on a different version of darwin, even though everything works just
> fine after removing the test. See the discussion over at skalibs
> mailing list:

It would not be too difficult to match the output of config.guess
against *:Darwin:*:* and, if matched, strip the unwanted version
information using sed.

If I remove the extra information, someone different will
complain. ;-)

Ben

signature.asc
Description: Digital signature