[Python-Dev] Collection of typedefs for integral types

Tom MacDonald tmacd@eintelligence-inc.com
Sun, 9 Jul 2000 22:27:05 -0500 (CDT)

> [Tim, pushing the adoption of some C9X-like typedefs in Python's source]
> > ...
> > the-committee-wasn't-just-jerking-off-the-last-10-years<wink>-ly y'rs  -
> > tim
> [Ken Manheimer]
> > (The obvious question is, where they doing those other things at the same
> > time, or during breaks?)
> It varies by member.  For example, the Reverend Tom MacDonald jerked off
> non-stop the entire time (& I've copied him on this for confirmation).  But
> a surprising number of the members have never jerked off at all!

You know me Tim.  Ever since you showed me the pictures of
mad dog and butcher vachon I can't stop.  Interesting question
but the committee never took a role call vote on that question.

> Rev Tom, after 118 years serving on X3J11, do you think C9X is a good thing
> or a bad thing?  And do you still get off on it regardless?  And what does
> God think about C9X?

If I have to boil C99 down to either good or bad, I'll pick good.  It
cleans up a lot of fuzzy words from C89, addresses some missing features
that should be there (e.g., variable length arrays, long long, hex
floating constants, // comments, <inttypes.h>, intmax_t, mixed
declarations and code, designated initializers).   There are features,
though, that I wish were not there.  The biggest being the complex
type.  This seems a tad ironic since I pushed for complex early on in
the C99 process.  I didn't realize that IEEE arithmetic made a complex
type grow like the blob - even an imaginary type is there for those
that want everything IEEE arithmetic has to offer.

I18N was added (I think).  I never could figure out if we made
everybody that spoke their own language happy or not.  It sure made C99
a lot bigger.  Almost all of it in the library - though an
implementation can support identifiers containing international

I don't like the following features:  flexible array members, the
number of new functions and operators required to support IEEE
arithmetic, the LIA compatibility annex, and C99 style boolean

I like the rest of it.  However, no matter what I think, I don't see
any vendors rushing to implement C99.  Instead vendors are picking some
C99 features and implementing those.  So, perhaps the the answer to
the question of whether C99 is good or bad is only interesting if
someone actually implements it.  Actually, you need several vendors
to implement it.  One reason C89 was so successful is that all
vendors implemented it.

I still like C.  In my new job I'm using C++ all the time.  I'm
probably getting old and can't appreciate all this stuff C++
does for me under the covers, but I can never figure out which
overloaded function or operator is actually instantiated by
template such-n-such.  With C, I can always see what's going on.
So yes, I sill got off on it.

I asked God what he thought about C99 and he said he didn't
have an opinion on C99 cause he always uses python.

Rev. Tom