IRIX success

[Dan Jacobowitz <dr...@false.org>]

For reference, the server compiles and correctly loads DSOs on IRIX
6.2 using the stock cc (which looks to me kind of like version 7.1).
__attribute__ was repeatedly redefined in
/usr/local/include/sys/cdefs.h, which at CMU could be just about
anyone's fault.  The other warning was about libap and libos not being
used to resolve symbols.

It indeed starts and serves just fine.

Dan

Re: IRIX success

[Paul Sutton <pa...@c2.net>]

On Sat, 19 Sep 1998, Dan Jacobowitz wrote:
> For reference, the server compiles and correctly loads DSOs on IRIX
> 6.2 using the stock cc (which looks to me kind of like version 7.1).

I just tried this setup.

> __attribute__ was repeatedly redefined in
> /usr/local/include/sys/cdefs.h, which at CMU could be just about

Didn't see this one.

> anyone's fault.  The other warning was about libap and libos not being
> used to resolve symbols.

Yep, got those, from various programs in support only.

In addition I got a bunch of "statement in unreachable" warnings from
regex, on constructs like this

        switch (OP(s)) {
        case OBACK_:            /* the vilest depths */
                i = OPND(s);
                .... various stuff ...
                return(backref(m, sp+len, stop, ss+1, stopst, lev));
                break;

Not surprising then. Plus the old "if (... (r->chuncked = 1) ...)"
warning, and a warning about an argument mismatch in the call of
ap_md5_binary() within ap_md5() which looks bogus.

It then runs fine.

Paul