Daniel Speed

So far, no. The intent is more to provide a library for the algorithm for binary data than a final tool though, and I believe it would be relatively simple...

I'd personally initially avoid a binary format, and go for json for this - it would make it easier to extend and update while taking advantage of the marshalling support...

Off the top of my head, MD5 or equivalent on a whole file should be pretty fast for an initial comparison. File length, modification date would be other potential indicators....

You have to be careful there - I have some experience with using FTP as a way of populating CDN origins with data files, but I think it would probably...

It's worth having a look at the latest changes (particularly noticeable in patch.go). Most of the changes shouldn't be breaking (or should be quick fixes), but could make things significantly...

You're right - the existing implementation does do something that I explicitly warn against. Note that hypothetically speaking, including cache-busting headers in the responses can help. The main reason that...

I think we've managed to reproduce the failure on python.org using urllib, I'm not sure that it's enough just to know which site is causing the problem, the code needs...