IDs are random. Makes it hard to compare saved graphs

[pachacamac]

export default function generateId() {
    return Date.now().toString() + (counter++).toString();
}

So currently the IDs for nodes, interfaces, connections are generated with the help of this function.

This is an issue because it makes it impossible to compare one saved version of a graph to another as they always differ.

Would it be possible to use some sort of hashing instead?

• for nodes over position, name, interfaces, options etc
• for connections over source- and destination interface
• for interfaces over parent node, type, options?

Something like that?

[newcat]

I don't think this is feasible, since the ids are also used in the editor internally and in Vue components as keys. So when calculating a hash based on the state of the node, it would change the hash whenever you move the node or change the value. Also it is certainly possible to have two different nodes at the same position and with the same state, which would result in the same id even though they are two different instances.

When loading a saved graph, the existing ids are loaded as well. So it should be possible to compare two versions of the same graph, because nodes and node interfaces will keep their id. Connections currently don't load their id, but this shouldn't really matter, since the from.id and to.id are loaded from the saved graph, so they can be compared as well.

If it would make life easier for you I can change it so that connections also load their id.

[pachacamac]

Thanks for clarifying! Now that you say it I think I only had an issue with the connections indeed. Then I checked how the IDs are generated and saw that they all use the same function. But yes I think this would make it more consistent if connections follow the same logic as nodes and interfaces.

[newcat]

Implemented in v1.9.1