Greg Stein <gstein@lyra.org> writes:
> Karl: I said we should be using that for the marshalling format (marshalling
> mean store it to disk or on the wire, and we need to read it back at some
> later point). And that we should use apr_time_t inside the code, wherever
> possible, and that we would convert the values to strings only when
> displayed to the user.
>
> Right now, we store in a bogus format, and we pass around strings in many
> cases, when we should be passing apr_time_t instead.
Sure, I certainly agree, and always have.
The reason I'm confused is that you were responding to commit 1124,
which is about converting the value to a human-readable string for
display to the user, something we're going to have to continue doing
(as you point out above). The conversion method itself will of course
be tweaked if we change the way Subversion handles dates internally --
maybe it will convert from apr_time_t, or from ISO 8601 -- but the
conversion will still happen at the same place, and for the same
purpose. In essence 1124 is unrelated to problem of Subversion's
internal dates needing improvement, I think.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:37:02 2006