Simon Burge <simonb@telstra.com.au> writes:
> The only rebuttal against this I can think if is that it reduces the
> amount of code we can share between ports (say sparc & sparc64, pmax
> and newsmips, m68k*).  Given the nature of the programs we are talking
> about, this may not be a sound argument.  If you can offer even the
> flimsiest of arguments against that one (and mrg doesn't counter-strike)
> then I'll drop bootprog_name[] ;-)

i dunno, who's to say that it should start with "NetBSD/${MACHINE}"?

Also Who's to say that putting the version in a script which
contatenates text from one of its arguments is better than putting it
in the source?  e.g. on the Alpha, instead of having one nice string
embedded in exactly one source file, half of the string would now come
from text hard-coded in newvers.sh, half would come from a makefile,
and the array -- which doesn't have a declaration in a well known
place, btw -- still has to be _used_ in the source.  It just doesn't
seem to make much sense to me...


If you wanted to go whole hog and stuff _all_ of the strings, name,
revision, maker, and date into single string, well, that'd be useful;
it'd save code space for multiple printfs, and a couple of NULs.  8-)

In that case, there should just be a single bootprog_version[], and
options should exist to omit the various parts (e.g. revision, maker,
date) information so that for places where it needs to be squeezed
the MI script can still be used.

Also, if you wanted to put a declaration for bootprog_version[]
someplace (and make it const!) then that'd be useful too.


cgd