On 20/07/2012 12:45, Sven Van Caekenberghe wrote:
On 19 Jul 2012, at 21:09, Sven Van Caekenberghe wrote:
Yes, this seems to be a problem: your analysis is correct, since the size is
unknown upfront - the whole idea of chunked transfer -, #streamFrom:to:size: in
its current version cannot work. I will look at this tomorrow, I am pretty sure
this can quite easily be fixed (minus the progress bar).
I made some changes and commits. You can load the lastest version of Zn with
the following load script
Gofer it
squeaksource: 'ZincHTTPComponents';
package: 'Zinc-HTTP';
package: 'Zinc-FileSystem';
package: 'Zinc-Tests';
load
If you are not on Pharo 2.0, use the following
Gofer it
squeaksource: 'ZincHTTPComponents';
package: 'Zinc-HTTP';
package: 'Zinc-FileSystem-Legacy';
package: 'Zinc-Tests';
load
Now, your example should work fine
'/tmp/foo.txt' asFileReference ensureDeleted.
ZnClient new url: 'http://www.google.com'; downloadTo: '/tmp/foo.txt'.
Thanks for that. I'm actually still having a bit of trouble when the
file is bigger than the chunk size.
ZnChunkedReadStream>>#readInto:startingAt:count: uses a limit variable
which is bigger than the collection and requestedCount.
Regards,
Chris
