Re: [Duplicity-talk] Asynchronous upload

[Peter Schuller]

> I noticed the --asynchronous-upload is labeled EXPERIMENTAL in the
> manual page for the latest version of duplicity. Is this just a generic
> cautionary warning (I.e., better safe than sorry) or are there any
> specific problems one should take into consideration before enabling
> this option?
>
> From my anecdotal testing it seems to work fine but I haven't tested
> this on a large scale. Perhaps someone with more experience could
> testify to the reliability of this feature in practice?

I am the source of the "EXPERIMENTAL". The reason was one of being
conservative, because the change introduced concurrency in a codebase
that in general was not written under the assumption that things
should be concurrency-safe.

I had plans to develop it further to support more arbitrary
concurrency levels and for both upload and download, but I haven't had
time to work on duplicity in a while.

At the time, the "EXPERIMENTAL" was only for conservatism and there
was no known specific issue. I don't think there are known left-over
issues since then (but I have only been following the mailing lists
partially) - if anyone knows of an issue feel free to chime in.

Maybe it's time to remove the 'EXPERIMENTAL' claim?

-- 
/ Peter Schuller