* [9fans] const declarators in structs
@ 2000-11-16 10:35 Stephen Parker
0 siblings, 0 replies; 2+ messages in thread
From: Stephen Parker @ 2000-11-16 10:35 UTC (permalink / raw)
To: '9fans@cse.psu.edu'
i spent a while this morning pondering what an error message from 8c meant.
i tracked it down to a struct of the form:
struct S {
int const *p;
};
the compiler suggested that unnamed objects in structs should be of struct
type.
(i know that const isn't really a plan9 thing, but this was a bit of code
from a small device that i was trying to port.)
stephen
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [9fans] const declarators in structs
@ 2000-11-16 21:23 Richard Miller
0 siblings, 0 replies; 2+ messages in thread
From: Richard Miller @ 2000-11-16 21:23 UTC (permalink / raw)
To: 9fans
stephen.parker@pitechnology.com reports a misleading C compiler error message from:
struct S {
int const *p;
};
The underlying cause is that the compiler expects const/volatile keywords before
other simple type names in a structure/union element declarator (but nowhere else) --
in the same context "const int *p" is considered acceptable.
Is this a deliberate choice or an oversight? If the latter, a simple correction
is to change the grammar as follows:
/sys/src/cmd/cc/cc.y:216 c cc.y:216
< etlist
---
> tlist
/sys/src/cmd/cc/cc.y:221 c cc.y:221
< | edecl etlist
---
> | edecl tlist
The definitions of etlist, etypes and tnlist then become redundant and can be
pruned from cc.y as well.
-- Richard Miller
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2000-11-16 21:23 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2000-11-16 10:35 [9fans] const declarators in structs Stephen Parker
2000-11-16 21:23 Richard Miller
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).