Path: utzoo!utgpu!water!watmath!clyde!ima!haddock!karl
From: karl@haddock.ISC.COM (Karl Heuer)
Newsgroups: comp.lang.c
Subject: struct alignment (was: While talking about useful additions, ...)
Message-ID: <2398@haddock.ISC.COM>
Date: 27 Jan 88 02:36:36 GMT
References: <253@vsi1.UUCP> <127@ghostwheel.UUCP> <2376@haddock.ISC.COM> <7178@brl-smoke.ARPA>
Reply-To: karl@haddock.ima.isc.com (Karl Heuer)
Organization: Interactive Systems, Boston
Lines: 18

In article <7178@brl-smoke.ARPA> gwyn@brl.arpa (Doug Gwyn (VLD/VMB) <gwyn>) writes:
>In article <2376@haddock.ISC.COM> karl@haddock.ima.isc.com (Karl Heuer) writes:
>>Thus, struct{char} and struct{int} may have different alignment constraints.
>
>I used to think that, too, but a few weeks ago after discussion with
>Dennis Ritchie I realized that all struct pointers have to "smell" the
>same, which implies that they might as well have the same alignment.

The way I see it, you have two choices on a word-addressible machine: either
word-align all structures (and use word pointers for struct pointers), or use
the minimal required (VAX-like) alignment (and use byte pointers for all
structs, even those that are known to be word-aligned.)  C vendors on such
machines will have to weigh the time-space tradeoff, but there's no reason to
penalize the folks who use byte-addressible machines.  On a VAX, where all
pointers "smell" the same anyway, there should be no problem with struct{char}
having char alignment.  (Of course, the user can't portably depend on it.)

Karl W. Z. Heuer (ima!haddock!karl or karl@haddock.isc.com), The Walking Lint