Re: improve INSTALL contents (was: Core-utils 7.2; building only 'su')

[Keith Marshall]

On Friday 15 May 2009 03:45:21 Alfred M. Szmidt wrote:
> How about this?  I took into account Ralf's comments as well.

We're talking a generic "one-size-fits-all" INSTALL here, right?

You *cannot* unequivocally assert this...

>    If you wish to install the package into a staging directory
>    (e.g. for packaging or testing purposes) then you can set the
>    DESTDIR variable, @samp{make address@hidden install}.

On the one hand, DESTDIR is not even *required*, (by GCS), to be 
supported by any given project; it is a "strongly recommended" 
option, but it is by no means a mandatory requirement.  This does 
*not* fit, for any project which *chooses* not to implement DESTDIR 
support.  s/can/may/ and you come closer to reality, but it still 
needs qualification, to indicate that DESTDIR isn't a universally 
supported feature.

The other hand, we have already addressed; there are situations in 
which, even if DESTDIR support *is* provided, it is so conceptually 
broken as to be utterly useless.  A prime example of this is the 
case of native building of projects on MS-Windows, (which, like it 
or not, remains the most popular and widely deployed desktop OS on 
the planet), and where prefacing an arbitrary DESTDIR to a properly 
defined $prefix, (which begins with a drive designator), results in 
an invalid path name, causing the `make install' to blow up in the 
user's face.

-- 

Regards,
Keith.