pcb-rnd devlog

subcircuit roadmap: element names

by Igor2

> So that brings up another question: how is the displayed element name
> going to change behavior in future versions of pcb-rnd?

The current situation in full details

there are exactly 3 pcb text objects attached to each pcb element object. A text object has size, coordinate, font ID, rotation and string (plus the usual background info like rtree entry). Although you see only one of these at any given time, the code maintains all three.

And although the coordinate and rotation and size and font, or in other words, everything but the string, need to be the same, we still store all these data 3 times... Every code that needs to touch an element text also needs to repeat all these ugly hacks; like the code that can move an element text really needs to move 3 to keep all coords in sync.

On one hand I like to be able to globally change whether we show refdes or value, but on the other hand I often want to see both. And even more: it's a different use case to see data on the screen or see the artwork on silk.

So my plan for subcircuits

• 1. We will be able to have as many text objects as the user wants, on whichever layer (including mask or even paste)
• 2. we'll have some metadata for each object (text objects included) that tells the role of the object.
• 3. The role won't be one of 3 sacred roles some holy priests invents in the initial implementation, but user defined. The default convention obviously will be our usual 3 roles: refdes, value, footprint name.
• 4. Because of 2. and 3., we will be able to identify what objects make up a refdes; these objects can be multiple texts, rectangles, can be on silk or any other layer. For example if I don't have silk but have mask, I could move my refdes' to mask. Or even to copper.
• 5. The UI... we'd have two different things there:
  • We'd have some GUI aid to display data on some overlay UI layer, independenlty of whether the data land on silk or not. So you'd be able to look at the value of all components on screen, without this affecting the silk layer in any way. This feature would be close to what we have now in the sense that you could globally tell it which role to show; but more like a set of checkboxes than a radio button: I often want to see both refdes and value.
  • Then we'd have some per-subcircuit control over which elements are hidden, so you won't need to get all them on silk. E.g. you'd show refdes and element-outline related objects but hide the others. This would be subcircuit-local, we wouldn't have a global setting for this. We could still have scripts and actions to make mass-changes possible, but the actual bit for show/hide is stored per subcircuit instance. This would let us print the value of some important elements, or even all resistors while none of the ICs.

Most of this wouldn't happen in the first implementation of subcircuits but would be added gradually.