Re: RFC: date parser rewrite

On Fri, 2003-12-19 at 19:34, Julian Foad wrote:
> The only thing as far as I am concerned that would be intolerable is
> guaranteeing to support all of the currently accepted date formats in
> future. As I mentioned, we can solve that by documenting that the
> nasty formats are not supported even though the code happens to accept
> them currently.

I think it would be naive of us to assume that people generally read the
documentation before using date formats. We would be taking only a
futile swing at the problem, not solving it.

> > 1.) protect the logic with a big mutex

> That sounds like a moderately important thing to do, and a small
> enough change that porting it into 1.0 is a reasonable option.

It's not a small change when you consider that we don't use mutexes
anywhere in the code base right now. Or that APR provides no way to
statically initialize mutexes. (Nor do we have a library-wide
initialization function or global context in which to store a
runtime-initialized mutex.)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Dec 20 01:58:53 2003