On 7/25/10 5:09 PM, slush wrote: > > is there any way how to download, say, second half of large file > without downloading first half? Is that even teoritically possible in > current implementation (can I decode file from any position)?
Yes and no. We carefully designed the file-encoding scheme to make random access like this fast and easy. However we didn't implement full support in the client right away. The downloader in 1.7.1 will respond to partial-read requests (expressed by passing a "Range:" header in your HTTP GET request) by downloading the whole file to a temporary directory, and then passing you just the range you cared about. If you request the first half of the file, it will give you that data as soon as it becomes available, so you wouldn't even notice an extra delay. But if you request the last half of the file, it has to download the whole thing before it gives you any data at all. The new downloader that I'm almost finished with (see ticket #798 for details) fixes this: if you ask for the last half of the file, it will download only that half. The new downloader does not use cache files or temporary directories. If you'd like to try it out, apply the latest patch form #798 and give it a spin.. I'm looking for feedback! It should work faster than the old downloader too (fewer roundtrips, although obviously the overall download rate for large files will remain the same). cheers, -Brian _______________________________________________ tahoe-dev mailing list [email protected] http://tahoe-lafs.org/cgi-bin/mailman/listinfo/tahoe-dev
