Re: svn commit: r20677 - trunk/subversion/bindings/swig

David James wrote:
> On 7/14/06, maxb@tigris.org <maxb@tigris.org> wrote:
>> SWIG Ruby: Retain compatibility with 1.3.x bindings, by NOT changing the
>> return type of svn_repos_get_committed_info().
>>
>> * subversion/bindings/swig/svn_repos.i:
>> Delete a well-intentioned but compatibility-breaking typemap.
>
> Why does this typemap break compatibility?

Actually, it doesn't - not of the public APIs, anyway. Sorry about that.
Reverted.

It *does* raise an interesting question though:

So far, all the bindings have kept to a uniform structure, of a
SWIG-generated layer which *closely* wraps the C API, and a native
scripting language wrapper, which adapts the auto-generated code to be
more pleasant to use.

What the Ruby bindings *have* done at some point between 1.3.x and 1.4.x
is to move a small part of this adaptation from the Ruby layer into the
C layer.

This is something I think we should not do without careful
consideration, because it makes it much harder to make the typemaps as
language-neutral as possible, if the various languages are deliberately
typemapping the same thing in different ways.

So, Kouhei, could you explain what advantages r19545 "Convert date C
string to Ruby Time in C level not Ruby level." brings?

Thanks,
Max.