Last week, I started thinking about what features I’d want in the TextMate Dojo completion bundle in order to feel good about calling it releasable. I made a list, and started knocking them down. Here are a few that I managed to tackle over the past week:
- It scans and executes
dojo.require()s in the current file and adds their symbols to the completion context.
- It recognizes when you are typing inside a
dojo.require()and completes dojo package names instead of functions and properties.
- It recognizes when you are in a
dojo.querychain and completes NodeList functions while you’re in the chain.
- Because it understands both
dojo.query, it will complete functions from NodeList mixins, such as
dojo.NodeList-fx, that you’ve
dojo.required in the current file.
How does it do that?
A note about performance
Any completion framework has to be fast. I’m ruthless when it comes to development tools, and I’ll bet you are too. If something slows me down, I stop using it. If the pain outweighs the benefits, it’s gone. So, one of my goals is to make sure this bundle is as fast as it can be.
The current performance is very good on my laptop. The completion popup is nearly instantaneous, even with
dojo.require scanning, and hundreds of potential completions. That said, I’m using the most recent MacBook Pro rev with 8g RAM and an SSD. I realize that may not be the most common setup, so I will be testing it with other setups to make sure it doesn’t suck.
I can haz it?
A few folks have asked when they can get the bundle. Since it’s a side project, I don’t want to give a date. I’ll release an initial version when I feel it’s at the point that I’d actually use it, and I can tell you that it’s getting close :)
For now, here are a few more teasers that show the new stuff in action.