Antoine Pitrou wrote:
Chris Withers <chris <at> simplistix.co.uk> writes:
However, I'm kinda stumped on where to start with debugging the difference. I've tried upping the buffer size as suggested in the issue, but it's had no effect...
Then perhaps it's not the same bug. Please take a look at CPU utilization during the download. If Python takes close to 100% CPU, it might be due to the lack of buffering or any other suboptimal situation in the implementation.
Well, it's locked at 25% on a quad core box, so yeah, I'd say something is wrong ;-)
I guess I could try profile it and finding out where most of the time is being spent?