hi,

since this is my patch (the htons call), I am the one to blame...

I am not sure why this code is undefined -- it is probably the new -Wsequence-point warning introduced in gcc 3.0, although there _is_ a sequence point between pre-incrementation and a function call (unless htons is now a macro, not a function call)?

How about gcc 3.1?

Krzysztof

Burton Samograd wrote: ...

I'm compiling pd test29 with gcc 3 and i ran into the following warning (treated as error)

s_inter.c: line 564 gives a minor warning of "operation might be undefined" or something like that...i didn't write the exact warning text down. Changing the line from:

server.sin_port = htons((unsigned short)(++portno));