[comp.sys.pyramid] isgraph in ucb - on MIS, where is it

dickson@escob1.UUCP (Dave Dickson) (04/13/91)

I am trying to compile ksh88f on a Pyramid MIS (OsX 5.0d) in the
UCB universe.  It seems that isgraph() is missing from the
pyramid libraries.  Does anyone have any ideas, is this an
over-site on Pyramid's part, or do I have to write one and merge
it into the library?
-- 
David G. Dickson
Ohio Bell Telephone Co. (614-223-8134)
uunet!escob1!dickson

keyvan@pyra.co.uk (Keyvan Shirnia Pre Sales) (04/14/91)

In article <1344@escob1.UUCP> dickson@escob1.UUCP (Dave Dickson) writes:
>I am trying to compile ksh88f on a Pyramid MIS (OsX 5.0d) in the
>UCB universe.  It seems that isgraph() is missing from the
>pyramid libraries.  Does anyone have any ideas, is this an
>over-site on Pyramid's part, or do I have to write one and merge
>it into the library?

isgraph() is defined in the att C library. You do not have to rewrite the
code for isgraph(). All you need to do is to either corss link into the att
universe (ouch! :-) or extract the routines isgraph.o and version.o from 
/.attlib/libc.a and put them in your own library.

NOTE: There have been quite a lot of discussions regarding cross-linking,
WHICH IS *NOT* SUPPORTED BY PYRAMID. However, my *personal* opinion is that if
you are only porting the software then you have no choice but to cross-link. 
And again, in my opinion there is no reason why this shouldn't work. You just 
have to be careful.


Best regards,
Keyvan

      -m------- Keyvan Shirnia		       Tel : +44 252 373035
    ---mmm----- Pre Sales Support 
  -----mmmmm--- Pyramid Technology Ltd.        keyvan@pyra.co.uk
-------mmmmmmm- Farnborough, England.

guy@auspex.auspex.com (Guy Harris) (04/16/91)

>isgraph() is defined in the att C library.

C *library*?  It doesn't seem to be defined in *AT&T's* C library, at
least as of System V Release 3.1; it's a macro, defined in <ctype.h>.

Are you claiming that there really is an "isgraph.o" in
"/.attlib/libc.a", which contains code for an "isgraph()" *function*? 
If so, why is that?

It's also defined in BSD's <ctype.h>, as of 4.3BSD (although, I infer, it's
not defined in the BSD-environment C include files in OSx).  This is,
amusingly enough, similar to the "memory.h" problem - BSD's been picking
things up from S5, and S5's been picking up things from BSD, so neither
one of them is very pure any more....