Back when I was making the multi person collab with MQTT blocks i noticed that you cant really send lists or costumes or really anything without making it super complex. Well that is no longer the case!!! with the blocks bellow you can send also lists and costumes without having to convert them in any format!!! I honestly feel this should be completely added to the MQTT system.
Your blocks don’t seem to be sending lists or costumes - just “hello”
Do you have modified blocks to do this or is this a feature request?
Maybe publish some more complete examples?
regards
Simon - MQTT Library maintainer
No this is the actual script. this is just a script pic of the blocks so people can use them and see what I mean. but here are some examples of what I mean.
oh binary mode does lists? so i made it do lists just regularly
Ok- now it’s daylight where I am - I can now see what your up to 
Using “magic numbers” to encode/decode different types of payload doesn’t immediately grab me
I think it should be possible to automatically detect if published payload is a costume, convert it to base64 and then decode it in the subscribe block without causing any compatibility problems
Although I introduced the auto encoding lists to JSON as an experiment, I was worried that automatically deciding to decode a JSON payload back into a list might cause issues with non-SNAP generated payloads.
I’ll have a play around and discuss with others who use MQTT
1st go at avoiding using magic numbers
Not sure it’s worth going to the trouble of auto encoding/decoding different payloads
I think the programmer should just manually encode and decode different payload types.
Ok
This my current thinking.
It is easy to simply add sound and image encoding to the list encoding
And adding a single simple auto decoding reporter is the simplest way of decoding.
I’ve added sound and costume reporters to also simplify things
Maybe
data:image, data:audio, and the performance penalty for decoding JSON seems to be neglectible.
I always avoid try/except programming except as a last resort - a personal preference
And I can see that possibly more types of payload might be able to to be encoded/decoded so probably will keep the multiple ifs
[edit]Also going to look into doing the auto decoding within the mqttExtension.js itself
Wym. Like the js?
Yes.
I’ve managed to handle lists but costumes are proving difficult.
attempt to convert to base64. if it works and returns either an image or audio then we know if its an audio or image and we return that
My problem is not identifying image/sound base64 payloads but trying to return a costume/sound from the base64 data
return costume as an image file.
Like I said - I know what to do but not how to do
its been done. look on forums
Please provide links
