fundamental
commented
6 years ago Maybe we could define a protocol to describe what should be screenshoted, maybe even how it should be anotated and have all that be done automatically.
I started working on a first pass of the system, but wasn't happy with how it was turning out. Essentially for this to work we need a domain specific language to convey the sequence of actions needed to obtain the screenshots.
This includes:
- Taking a screenshot of the active region
- Identifying the active region based upon the bounding boxes of several widgets
- Changing the view of the GUI
- Changing some OSC state within the GUI
- Waiting either some real time or some frame-time before running a set of events
Setting up the waiting for events can be handled by a DAG. OSC and view changes are simple enough. Some work has to be put into the screenshot-ing, but that's fairly step by step coding.
Identifying the region of interest in an easy to use approach is trickier. I'm thinking you say "I want to see something that contains '.Pvolume' and '.Pstretch'" or something like that to build up a region of interest.
Beyond that the exact semantics of how you specify this series of commands is undecided.
unfa
In the Goals.adoc I've created two ideas for a system that could let us create and update screenshots automatically without having to re-shoot them all when the look of Zyn-Fusion changes.
I've asked @fundamental and he said it should be possible to have Zyn-Fusion provide screenshots itself from requested UI elements.
Maybe we could define a protocol to describe what should be screenshoted, maybe even ow it should be anotated and have all that be done automatically.
Then hopefully running one command will update all screenshots in the manual for a new Zyn-Fusion UI look or layout.