Re: [evolution-patches] report_writable e-book change

[Hans Petter Jansson <hpj ximian com>]

On Thu, 2004-07-15 at 00:25 -0700, Chris Toshok wrote:
> there were a few things wrong with the _report_writable handling in
> libebook.
> 
> 1) it used an idle handler, but didn't remove the handler in
> e_book_dispose (this isn't so much of an issue, since a ref is held).
> 2) it created a closure when it didn't need to
> 3) it wasn't MT-safe.

There's one problem with this patch:

In the WritableStatusEvent case of e_book_handle_response(), you ref the
book unconditionally, but attach the handler conditioned on there not
already being one. I guess that ref is supposed to be released in
e_book_idle_writable(), so we end up with a ref leak.

It'd be nice if you used cvs diff -up, by the way, so the patch says
what function it's in.

--
Hans Petter