If a file fails to download, it is deleted. That's
better than leaving a partial file, but the whole file
must be retrieved again, even if the download failed at
99%. Obviously, this is suboptimal.
A suggested solution is to use the filename hash and/or
a checksum to match the file and resume the download
without corrupting the file. All files should be put
into a temporary directory with some other extension
until they are completely downloaded, at which point
they are moved to the download directly.
This is a great concept with great potential. Clearly
it's earned the "Alpha" status, but I think it could do
great things to bring down obsolete business models and
reestablish truly Free (speech) interchange.
Logged In: YES
user_id=781921
A much needed feature i agree, the temp folder is already
implemented in the latest CVS btw.
Logged In: NO
if you rezume and have a temp folder, then partial filesharing
should be implemented as well.
Logged In: NO
yup, resume support would be nice,..
Logged In: NO
All it needs to do on failure is automatically search for the
hash_1234... code of the failed file and resume the download
with the next chunk. This may re-find the same file if the
original source got disconnected and reconnected with an
new ID, or it will find a new route if the old route died, or it
may find a mirror of the file on some other system that's still
up; doesn't matter, it's all good. The download should never
give up until the user cancels it; it should just keep re-
searching and resuming.
Logged In: YES
user_id=690572
we need this resume.
WHen you have the hash, the hash search is simple to find
same file on another source.
Then you can req the other chunks from the other source.
I like an option to activate this redume automatically.
Fighting since the job is done =)) !!!
cheers
Logged In: YES
user_id=690572
IMPORTANT, maximum consistency in data...to be sure that the
chunks saved are downloaded correctly.
I prey for this