Simon Urbanek

Sure, I have added explicit MIT license (with non-binding copyleft) to the original file: https://github.com/s-u/multicore/blob/master/src/forknt.c But please keep me posted if you get it working - a lot of people...

This is, unfortunately, by design. In order to be efficient and avoid loading the entire contents into memory first, `save_object` instructs the HTTP engine to write the result into the...

If you look at the `s3sync` code , it's entirely manual - there is no actual "sync" HTTP command, even the CLI has to do all the work - figuring...

Short answer, sorry, no. Long answer: it would make a lot of sense, since we could use parallel connections to make it a lot faster. However, that would require re-write...

Well, the entire error handling requires a re-write, because it's a mess. I have started a branch `unify-response` that was aimed at addressing that, but a) I wasn't sure how...

I started re-writing the error handling to unify the responses such that an error object is always returned which contains all vital information: https://github.com/cloudyr/aws.s3/tree/unify-response The idea is that all functions...

The simple answer is this is not supported. `curl` is not compatible with `mclapply`, because you cannot fork CURL handles (nor SSL handles). It will only work if you fork...

@fcomte well, the main problem is that `AWS_S3_ENDPOINT` seems to be non-standard (not used by other AWS clients) and is pretty much undocumented. It appears to be only used as...

@fcomte Agree, that's exactly what I said above. I wouldn't write it that way and I don't know why it was introduced the way it is, but that's what the...

Ideally, you should use `R CMD` to setup the environment, so if let's say if you are running sh startme.sh then for Java to work you'd run R CMD sh...