On Tue, 8 Jan 2008, Philippe HAMEAU wrote:

Only on the users list now, and please don't resort to top-posting.

> To me, trying to continue a file where it last stopped should produce a
> valid file in the end, no matter how many retries/continues it took.

Of course, I'm not trying to imply otherwise. I'm just saying that getting
actual facts and info is better than guessing.

> The rollback/garbage bytes things is something I have read long ago in the
> documentation of some download manager (maybe GetRight) : it was some
> security feature in case the server had a bad behaviour when giving up.

Yes, it makes sense to do rollbacks at times, and especially when proxies are
used, but for the plain timeout abort that you showed I can't see that it
would make any difference. And the truncate approach currently used could be
seen as a pretty major rollback...

After all, I didn't say rollback would be a totally bad idea - curl just
doesn't offer any such feature atm.

-- 
  Commercial curl and libcurl Technical Support: http://haxx.se/curl.html