[cfarm-users] What are people testing 32-bit builds on?

[Peter Gutmann]

Werner Koch writes:

>I have asked for years that this should be done given that POSIX explictly
>allows this.

I'm not sure how much it'll help given that both the standards and many
environments make it impossible to find out what time_t actually is without
building an executable and running it, which isn't terribly practical when
"running it" means flashing a firmware image and attaching a debugger.  The
systems that use unsigned time_t, and let you know about it, are MQX, Quadros,
Telit, and VxWorks, but for everything else it's almost, if not actually,
impossible to tell what you're getting, although for non-RTOSes you can assume
time_t is signed.

Obviously all of the ones I've mentioned above are non-Posix so I'm not sure
if they care too much about how Posix-compliant their time_t really is.

Peter.