2015-07-26 23:08 GMT+02:00 Fabio Pietrosanti (naif) - lists <[email protected]>: > Kudos! > > On 7/26/15 11:05 PM, Giovanni Pellerano wrote: >> next steps i will follow are: >> 1) implement decryption in streaming >> 2) try to apply the same optimization with reused buffers also for the >> non-streamed implementation of encrypt/decrypt. > > Are there use cases for "non-streamed" encrypt/decrypt when there is > encrypt/decrypt in streaming that's more efficient? > > I mean, does it make sense to have "streaming-way" of operating as it's > default or there's application-way to keep two way of operations > (streamed/non-streamed)? > generally a library implement both the possibilities as the usage depends on what you need to do and what are your requirements. for small payloads generally operating as a whole (eventually in a worker) is mor efficient in term of cpu time, while performing streaming becomes a strong requirement when you need to encrypt large payloads (e.g. the non streaming version in my tests require 1500mb of ram for a file of only 42mb on chrome making impossible to encrypt larger files) _______________________________________________
http://openpgpjs.org Subscribe/unsubscribe: http://list.openpgpjs.org
