Re: Small patch for SCO systems

>> Dont remove -O for gcc 2.95.3 on SCO OpenServer.
>
>
> Are you sure? We don't have much trust in those GCC versions... I no
> longer remember the exact conditions when this GCC optimization bug was
> discovered but I think it was a generic X86 bug, not tied to a specific
> OS platform..
Pretty sure. We've done a great deal of bugfixing on that branch
and I compile about 2G of open source stuff with -O6 and
-fomit-frame-pointer and have no problems.

> The -O flag does not make that big difference for Squid. If in any doubt
> on the compiler quality it's recommended not to use it..
If you want to err on the side of caution I'm all for that,
and it wouldnt upset me at all if you want to keep that test
and not apply that bit of the patch.

>
>> Add check fot setenv as not all systems have it.
>
>
> Already fixed by using putenv instead.
Ok great. Thanks.

>> Include syslog.h when checking for the syslog function.
>
>
> good I suppose.
Yeah without it, it fails to detect that syslog is available,
so the default RunCache fails as squid coredumps immediately
on startup becuase it uses -s.

>> * src/squid.h: Compare SQUID_MAXFD to FD_SETSIZE, not
>> DEFAULT_FD_SETSIZE, as it is the former that is used during
>> compilation, not the latter.
>
>
> FD_SETSIZE is NOT available at this point in the header file. Only after
> the system headers have been included at which point it's too late to
> override the system defaults.
>
> DEFAULT_FD_SETSIZE is the FD_SETSIZE detected by configure, and is
> available here via autoconf.h.
>
> The purpose of this check is to see if Squid needs to override
> FD_SETSIZE by defining it before including the system headers, or if the
> system default FD_SETSIZE is sufficient.
Ok then the check should be more complete and be compared to
both, becuase the current scheme foils things if you compile
with make CC='gcc -DFD_SETSIZE=12345' for example. I realize
that in theory, that should have been the args used to configure
so that that value is detected at autoconf time, but it does
no harm to check both values. Without this I got compiler
warnings that FD_SETSIZE is redfined.

Kean
Received on Tue Nov 15 2005 - 03:16:34 MST