u_int32_t vs uint32_t

[Mike A. Harris]

Alan Coopersmith wrote:
> Torrey Lyons wrote:
> 
>> Are there any supported platforms where <stdint.h> does not work?
> 
> 
> Yes, Solaris 9 and earlier.
> 
> What are our supported platforms anyway?  I've been thinking
> off and on for a while that we should put a page in the wiki
> with the requirements for a supported platform so that developers
> can know what assumptions they can make.  For instance, I'd
> say C89 compliance is required, but not yet C99.

I think it's a good idea to list the supported platforms, but I'd go one 
step further, and define what "supported" means (whatever it means).

I'd also suggest that anything that gets "supported" status, must have 
at least one person who is willing to test builds on that platform/arch 
combo over the development cycle and provide bug reports, patches, etc. 
back via bugzilla, possibly running a tinderbox also.  If someone isn't 
regularly building on a given platform, personally I don't consider it 
officially supported, as we'd never know if it actually compiled or not, 
and if so, if anything actually worked.  Also, bug reports would get 
little attention for platforms nobody is interested in maintaining and 
contributing for, etc.

I think for 6.8.0, we should consider "supported" platforms, ones that 
have went through Kevin's test matrix on the website and passed some 
specific (yet to be defined) level of testing.  Any platforms not 
building, installing, and running with good test results once 6.8.0 is 
stamped, could be listed either in "untested" category, 
"deprecated/obsolete" category, or "unsupported/unmaintained" category 
or somesuch depending on circumstances.  For those not in the supported 
list, a request for volunteers to maintain the platform could be put there.

Just some thoughts that might be useful..

TTYL