Re: CVS update: subversion/subversion/libsvn_delta vcdiff_parse.c

> Come on. It works *everywhere*.

I certainly know people who have made money writing code for some
of the arcane crufty platforms mentioned in the C FAQ.

At any rate, relax. I said, "If you're aware of this and don't care,
that's okay." I just didn't want anyone to be ignorant of the issue
when using that construction.

And it's not that I'm terribly concerned that we will some day target
an architecture where it doesn't work--although a lot of code
certainly breaks on the Alpha because the author simply didn't care
about platforms where you couldn't stuff a pointer into an int. I
just believe in staying within the constraints of formal
specifications. But I won't be religious about that.

> If it wasn't, then apr_pcalloc() wouldn't even be necessary.

Not a very strong argument. Not every structure contains pointers,
and not every application can spare the extra cycles to, for instance,
manually initialize a large numeric array.
Received on Sat Oct 21 14:36:10 2006