Hello Glenn,

I would like to override CC.PREFIX.ARCHIVE and CC.SUFFIX.ARCHIVE since
I have a number of existing libraries which do not use the convention
lib$(NAME).a but rather $(NAME).lib, and I would prefer not to mix
conventions. The probe info isn't wrong as such, since the compiler can
accept lib$(NAME).a - it just doesn't match what I need.

I am cross-compiling, and did have a struggle to get probe to run
successfully, but those problems weren't connected to this one with
library naming conventions.

Alexander

On Thu, 11 Nov 2010, Glenn Fowler wrote:

> 
> the probe info load happens early in the nmake startup
> during the makerules load which hapens before and user
> --global or --file (including the default $(MAKEFILES)) are read
> 
> as you noted you can override individual CC.* on the command line
> including specifying a complete probe info file via CC.PROBE=path
> but the command line workaround is not general
> 
> background: the nmake probe information is generated by
>       probe -g C make <path-to-cc>
>       probe -g C make <path-to-cc>
> and listed by
>       probe -l C make <path-to-cc>
> the path name is
>       probe -k C make <path-to-cc>
> <path-to-cc> is picked up from $(CC) using the typical rules
> for resolving CC (command-line environment makefile rules)
> 
> usually CC.* overrides are an indication that the probe info is incorrect
> why do you need the override?
> is the probe info incorrect?
> are you cross-compiling?
> 
> On Wed, 10 Nov 2010 13:31:10 +0100 (CET) Alexander Thorp wrote:
> > I would like to override the values for CC.PREFIX.ARCHIVE and
> > CC.SUFFIX.ARCHIVE set in the probe information file, without having
> > to edit the probe information file itself. Is this possible?
> 
> > Setting the values on the command line works:
> 
> > $ nmake CC.PREFIX.ARCHIVE="" CC.SUFFIX.ARCHIVE=.lib
> 
> > However, setting these values in my Makefile does not work,
> > presumably because Makerules.mk has already been read at this point.
> 
> > Alexander
> 
_______________________________________________
ast-users mailing list
[email protected]
https://mailman.research.att.com/mailman/listinfo/ast-users

Reply via email to