On Wed, Oct 15, 2014 at 12:03 AM, Jason Yan <[email protected]> wrote:
> What do you think about supporting multiple concurrency backends?  We
> could provide support for both gevent and multiprocessing.  Another
> option, and one that might be particularly suitable if you want to get
> rid of gevent, would be to look at using the concurrent.futures
> backport from Python 3.2 and instead support threads and
> multiprocessing.  I'd be happy to help take a stab at this.

Among the goals I'd have for removing gevent is python3 support.  It
would seem all my major dependencies support it...so it's about time
to move on:

https://github.com/boto/boto
https://pypi.python.org/pypi/azurepython3/1.7.7
https://wiki.openstack.org/wiki/Python3#OpenStack_clients

Another reason to write such a shim is because I have some doubts as
to how VM-light approaches using "fork" might be, unless we can keep
the parent process small somehow.

I am not that excited about targeting multiple backends as a use
feature unless an express need arrives, but it may be worth carrying
as a jig to figure out what's going on in some cases.
concurrent.futures looks interesting.

I wouldn't let this be a drag on your momentum, it seems like you have
enough effort ahead of you already.  I can figure that out after the
fact and if and only if it winds up fixing a severe issue with vm
footprint.

As a bonus, IIRC the memory consumption of "tarfile" on Python3 has a
shot at being drastically lower, because it uses __slots__.  I had to
go to somewhat cruel and unusual lengths to get backups on some
pathological systems and backporting tarfile looked like too big a
slog.

-- 
You received this message because you are subscribed to the Google Groups 
"wal-e" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to