Re: deferred evaulation of function calls

[Robert P. J. Day]

On Thu, 27 May 2004, Paul D. Smith wrote:

> %% "Robert P. J. Day" <address@hidden> writes:
> 
>   >> If you modify the contents of a directory "behind make's back" then
>   >> things like wildcard, etc. may not work properly since they operate on
>   >> the cached version not the real directory contents.
> 
>   rpjd> ah, so what this tells me is that, if i want to match a wildcard
>   rpjd> pattern and i use the wildcard() function, there really is no
>   rpjd> effective difference if i do that at the top of my makefile with
>   rpjd> either of:
> 
> Hm.  Really, it's not that determinant... it depends on when the caching
> is done.  Unfortunately.  The directory cache is a big time savings but
> it also leads to a lot of non-intuitive behavior.  I'm more and more
> ambivalent about it.
> 
> You can always use:
> 
>     files := $(shell echo *o)
> 
> if you want to be sure you're getting what's _really_ in the directory.

it wasn't so much that i desperately *needed* that feature, i was just 
trying to clarify the immediate vs. deferred nature of function calls.
so, as i understand it, function calls *do* distinguish between immediate 
and deferred invocation -- i just happened to pick a bad example involving 
wildcard() that didn't demonstrate how it worked.

rday