On Sat, Jun 13, 2015 at 5:38 AM, Sean P. DeNigris <
[hidden email]>
wrote:
> I have a few Pharo apps that simply offer querying of an unchanging
> database
> of objects - think something like a train schedule. I often want to offer
> these via Amber. Creating the Amber UI is easy enough. The thing that
> usually slows me down is importing the data. I know I can use Pharo as a
> backend and keep the data there, but since there is no editing involved, it
> would greatly simplify things to keep everything in Amber. The typical ways
> to store objects in Pharo are Fuel or STON. But as far as I know, neither
> of
> these are compatible with Amber, are they? What I've done in the past is
> write a custom serializer that creates class-side constructor methods to
> re-create the data, but feels like too much extra work.
>
STON is an extension of the JSON notation, if I understood this correctly
from a recent thread on the Pharo mailinglist.
Is that right?
What Amber supports out of the box is JSON (obviously).
So far there is no support for Fuel of STON but STON might work if it could
be transformed to/queried as plain JSON.
> Is anyone else running into this complication i.e. wanting to move data
> from
> Pharo to Amber? If so, what is the best practice here? Would it be worth it
> to port something to facilitate... maybe STON?
>
STON is going to be integrated into Pharo 5, right?
In that case it might be worth to have a library written in Amber which
supports reading/writing STON so an Amber application can easily exchange
data with a Pharo server.
I think that using the regular JSON parser will not work here (correct me
if I am wrong).
The question is if writing a new parser (Amber uses PEG.js for parsing)
would be required for a STON library or if the Amber internal parser could
be reused in this case.
Best,
Manfred
Locked
