BTW I'm happy to write a quick example to illustrate any intended use case
you have in mind.

On 5 May 2017 at 21:36, Mindy <[email protected]> wrote:

> On 05/05/2017 02:56 PM, Tom Ridge wrote:
>
>> Yes it is very easy. Please see the disk_ops type in the API.
>>
>> (** Disk operations: read, write, and sync *)
>> type 't disk_ops = {
>>   block_size: BLK.sz;
>>   read: BLK.r -> (BLK.t,'t) m;
>>   write: BLK.r -> BLK.t -> (unit,'t) m;
>>   disk_sync: unit -> (unit,'t) m;
>> }
>>
>> Also the disk_to_store function, and store_to_map. So if you have a
>> disk_ops (or store_ops) you can get a map_ops. But really this is a
>> B-tree library. More work needs to be done to get an efficient KV store,
>> or filesystem or whatever. But yes, it is already done - I looked at the
>> Mirage block interface a while ago and it seems reasonable and close to
>> what disk_ops assumes (which isn't much).
>>
>
> Nice, thanks!
>
> BTW what was confusing about the structure? Did you read the overview?
>>
>> https://tomjridge.github.io/tjr_btree/Tjr_btree_doc.html
>>
>
> Nope, I just read your included example and went to the repository and
> started digging around in it; I didn't notice your "documentation" link at
> the bottom the README, but indeed it's there.
>
> Thanks for your interest!
>>
>
> Thanks for your work and the announcement!
>
> -Mindy
>
_______________________________________________
MirageOS-devel mailing list
[email protected]
https://lists.xenproject.org/cgi-bin/mailman/listinfo/mirageos-devel

Reply via email to