In order to support resume, you have to keep the partial upload, and 
then you need a process that occasionally goes through and finds the 
partial uploads that never completed.

More housekeeping, and probably not worth the savings.

I don't know if BOINC handles this now or not, just that it's another 
place where you're trading one problem for another.

-- Lynn

Nicolás Alvarez wrote:
> El Vie 14 Ago 2009 18:30:36 Martin escribió:
>> Lynn W. Taylor wrote:
>> [...]
>>
>>> As for the fact that the full upload has to complete before the error
>>> comes back, that's as much an artifact of HTTP as anything else, and
>>> can't be fixed without using something other than HTTP.
>> Include some mechanism to automatically break up large transfers into
>> smaller more manageable chunks?...
>>
>> A number of smaller chunks have a much better chance of success than one
>> very large lump if there any networking problems (congestion or otherwise).
> 
> No, why would that help? It's more efficient to send the whole thing at once 
> than to do multiple requests for smaller chunks, and if it's aborted halfway, 
> you can resume where it left off.
> _______________________________________________
> boinc_dev mailing list
> [email protected]
> http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
> To unsubscribe, visit the above URL and
> (near bottom of page) enter your email address.
> 
_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.

Reply via email to