Re: improving async support in psycopg

On 24/03/10 22:32, Jan Urbański wrote:
> I'll publish my psycopg2 branch after I clean it up a bit

And here it is:

http://git.wulczer.org/?p=psycopg2.git;a=summary

I have implemented all (or so I think) the API I wrote about earlier, 
with one change:

cursor.isready() got changed to cursor.poll(). This method just can't 
return a boolean, it has to somehow tell you which event should you wait 
for: socket writability or readability. I'm not terribly attached to the 
name, but since the API is changing, it might as well change the method 
name. Otherwise code that uses async might not notice that it's broken 
for a long time.
If someone has a better proposal than poll(), or if the consensus is 
that isready() could return something that's not a boolean, I'm all ears.

Documentation patches are missing from that repository, I'll wait with 
learning Sphinx until you guys give me some feedback on this ;) But of 
course if the feature would get accepted, I'll update the docs as well.

I haven't done any Windows testing, but I'm almost 100% sure it's going 
to work. I also haven't tested against other PGs than 9.0devel, but then 
again the worst thing that can happen is that the unit tests will need 
some adjusting.

Question: I don't really know what's the custom on this list: should I 
attach the patchset to a mail? Or is just pushing the changes to a git 
repo OK for you?

My previous patches were done against the master branch of the git repo 
on initd.org, but I then noticed that the python2 branch is much more 
advanced... So I rebased against python2. Maybe there should be a notice 
somewhere that the active branch is python2, it took me some time to 
realize I'm working on old code :(

Cheers,
Jan
_______________________________________________
Psycopg mailing list
[email protected]
http://lists.initd.org/mailman/listinfo/psycopg