On 30/10/13 00:11, xor wrote:
> On Saturday, October 26, 2013 09:11:03 PM Matthew Toseland wrote:
>> The initial state could be huge. 
... Skip discussion of filters, which are of course absolutely vital ...
>> Please don't send it as part of the FCP
>> message's SimpleFieldSet. 
> Do we currently have a limit on the size of a SFS?
They are generally parsed and kept in RAM, at least until we send them.
>> Either split it up into multiple messages
>> (e.g. one per identity; but then make sure there is some way of telling
>> that you got all of them), 
> Would be tricky because the synchronization should be an atomic operation 
> which guarantees that the client is in a synchronized state after it :| It 
> might be easy using multiple synchronous FCP calls, but I'm reluctant to 
> change it since it will overcomplicate a lot.
Right, hence the second option ...
>> or (probably better) put it in a Bucket i.e.
>> a data field.
> Which storage format should I use then?
Another SFS of course!
> We use SFS because it allows easy implementation of clients since it is human 
> readble. Do we have something similar which can be shoved into buckets?
>
> Could we do 1 SFS per Identity and shove multiple SFS into the Bucket?
>
> Would suck a bucket be obtainable by FCP clients which are not directly 
> compiled against fred but connected via network only and written in a 
> completely different language?
You can shove a single SFS or a series of SFS's into a Bucket. Just make
sure you standardise the charset. I recommend a series of SFS's rather
than a single SFS since the current Freenet SFS code will read the whole
thing into RAM.

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
Devl mailing list
[email protected]
https://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to