Strip comments when altering the DOM

[whi7ed3vil]

I've got some fairly complicated CSS overlays that end up in various parts of the DOM. In low resource environments, the repetitive/recursive application of these styles can become a bit bloated on a page. This is compounded by the comments left in the DOM from the original CSS. I'd request that the application default behavior is to strip comments and lexicographically condense the CSS when injecting into the DOM for faster parsing. This might be a marginal gain in most settings, but I believe in making good use of comments as these living documents are modified, and that seems opposed to the implementation strategy which allows all those useless bytes to remain recursively embedded in the live page. Over time, it does seem to cause some noticeable performance delays. Perhaps a setting would allow the purists to prevent contraction before injection? But I don't see any reason to allow comments and extra whitespace for that matter to be embedded after the injector knows what to put where.

[tophf]

Stripping of comments and whitespace in JS sounds like a much more expensive operation than doing it in the optimized native machine code inside the browser. Naturally, there may be a bug in the browser, but that should be reported and fixed, and the first step would be to verify there's a detectable difference by comparing results in devtools -> performance for your CSS's original vs minified state.

[tophf]

One possible implementation I can agree with at this point would be a [ ] minify injected CSS for performance gains option that stores the minified CSS inside the style database when a style is edited/updated. It would naturally nearly double the disk space for the database, but that's the user's decision to make.

[whi7ed3vil]

Yes I hear you — and your second comment quite well captures what I had in mind. Doing it once upon saving is a small task vs doing it in real time multiplied by hundreds or thousands of recursions on a page that’s actively mutated would be favorable no matter how fast the browser can get the job done over JS, IMO. (Highly mutated pages are what drive the consideration, mainly.) Disk space expense is a reasonable consideration, but also a minimal expense given the typical scenario. User option is good. Thanks for your thoughts! :)