Re: [G2R] ORBit2 / linc non-blocking API: change request
- From: Michael Meeks <michael ximian com>
- To: Jody Goldberg <jody gnome org>
- Cc: Release Team <gnome2-release-team gnome org>, Havoc Pennington <hp redhat com>, <gnome-2-0-list gnome org>
- Subject: Re: [G2R] ORBit2 / linc non-blocking API: change request
- Date: Fri, 28 Dec 2001 15:06:30 -0500 (EST)
Hi Jody,
On Fri, 28 Dec 2001, Jody Goldberg wrote:
> Hmm, the change itself would be ok to go in, but your description of
> it makes me wonder if it will be usable in all situations. The
> gnumeric -> guppi connection sometimes pumps big chunks of data
> across. I'd like to be sure that the connection breaker is either
> optional, or takes the number of pending msgs into account along
> with the outstanding byte count.
It's totaly optional yes - naturaly we still want to be able to
pass lots of data around. As a nice side effect it means we can do
_really_ asynchronous data transfer with the new async API - so we'll
can avoid blocking altogether even on normal interfaces ( like
Bonobo::Stream ).
> Given that linc is basicly an ORBit2 internal library and this change
> is strictly a signal addition it safe to go in even this far into the
> freeze. If there is the potential for more change in the future you
> may want to add some placeholders for future expansion while you're at
> it.
Sure - the 'blocking' signal is really for future expansion - it
seems like a useful thing to be able to be notified of; and the write
options being opaque mean that we can easily add other options such as
timeouts later without altering the message signatures in a purely
incremental way.
Thanks for looking it over,
Regards,
Michael.
--
mmeeks gnu org <><, Pseudo Engineer, itinerant idiot
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]