Re: What to return in async transfer callbacks?

[Matthias Kaeppler <noreply finitestate org>]

Alexander Larsson wrote:
> If you're copying a directory with a million small files in it, then
> progress_sync_callback will be called a million times, with the source
> and target filename for each file. progress_update_callback however,
> will at most be called every 100 msecs for progress updates. Both are
> called when you need user interaction, but its better to use the async
> one.
Well, if not more than 1 file is copied every 100ms (which seems to be 
very unlikely to me), the update-callback is in fact the one which is 
called WAY more often, and should be the faster one, right? ^^

For example, when copying a file which is sized e.g. 100MB, the sync 
callback would only be called before and after the transfer (if I got 
that part right), but the update callback would be called probably 
thousands of times.

Still, I am not sure about whether I was right with the confused 
parameters. Can you really confirm that? After reading the nautilus 
source code (or the relevant parts), it looks as if they are passing the 
arguments in the order specified by the documentation I claimed to be 
wrong (because it didn't work for me). In other words, I'm not so sure 
afterall if I was right with that observation, since it seems to work 
for the nautilus guys.

I'm really confused now.

> In short, unless you want to know the exact operation taking place in
> the operation just ignore the sync callback.

Okay.

Regards,
Matthias