Issue #234 Paint in Place Removed Image Changes

[jrconway3]

Removed the spritesheet restructure from #313 so we can easier see the code changes.

There are still some image changes, but only the ones necessary to make skintones work for all assets (namely this still adds missing base assets for Wolf/Cat Ear Skintones and Wrinkles).

[jrconway3]

Thanks for the additional comments. I'll review them shortly.

[jrconway3]

#330 (comment)

For some stupid reason I cannot leave a response comment or mark this as resolved anymore when it was literally just working an hour ago.

Anyway, I looked into this and I watched the video. I still have no idea what the issue is, at no point did I find it being slow at all. I asked Copilot to find a resolution, I tested it, it works but its not any faster or slower than before, so frankly, I don't really see what the problem is.

I guess the argument is that if there's enough entries it COULD cause massive slowdown, but frankly the odds of that happening in the future are very slim because not many assets will have extra colors anyway. But feel free to look at this solution and let me know how it looks.

[Copilot]

Pull request overview

Copilot reviewed 139 out of 413 changed files in this pull request and generated 8 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Gaurav0]

#330 (comment)

For some stupid reason I cannot leave a response comment or mark this as resolved anymore when it was literally just working an hour ago.

Anyway, I looked into this and I watched the video. I still have no idea what the issue is, at no point did I find it being slow at all. I asked Copilot to find a resolution, I tested it, it works but its not any faster or slower than before, so frankly, I don't really see what the problem is.

I guess the argument is that if there's enough entries it COULD cause massive slowdown, but frankly the odds of that happening in the future are very slim because not many assets will have extra colors anyway. But feel free to look at this solution and let me know how it looks.

Well this certainly should be much faster just by looking at the new code, but without instrumenting it would be hard to tell if it actually is. Interesting that its using \u0000 as a delimiter in the key.

There is already code in the repo demonstrating how to instrument performance. See

Universal-LPC-Spritesheet-Character-Generator/sources/canvas/renderer.js

Lines 383 to 389 in df495db

	// Mark end and measure
	if (profiler) {
	profiler.mark("renderCharacter:end");
	profiler.measure(
	"renderCharacter",
	"renderCharacter:start",
	"renderCharacter:end"

[jrconway3]

#330 (comment)
For some stupid reason I cannot leave a response comment or mark this as resolved anymore when it was literally just working an hour ago.
Anyway, I looked into this and I watched the video. I still have no idea what the issue is, at no point did I find it being slow at all. I asked Copilot to find a resolution, I tested it, it works but its not any faster or slower than before, so frankly, I don't really see what the problem is.
I guess the argument is that if there's enough entries it COULD cause massive slowdown, but frankly the odds of that happening in the future are very slim because not many assets will have extra colors anyway. But feel free to look at this solution and let me know how it looks.

Well this certainly should be much faster just by looking at the new code, but without instrumenting it would be hard to tell if it actually is. Interesting that its using \u0000 as a delimiter in the key.

There is already code in the repo demonstrating how to instrument performance. See

Universal-LPC-Spritesheet-Character-Generator/sources/canvas/renderer.js

Lines 383 to 389 in df495db

	// Mark end and measure
	if (profiler) {
	profiler.mark("renderCharacter:end");
	profiler.measure(
	"renderCharacter",
	"renderCharacter:start",
	"renderCharacter:end"

I added a profiler, let me know how that looks. Its not reporting anything, but I checked and confirmed that it is hitting the profiler by throwing console logs into the if (profiler) { blocks.

[Gaurav0]

I added a profiler, let me know how that looks. Its not reporting anything, but I checked and confirmed that it is hitting the profiler by throwing console logs into the if (profiler) { blocks.

You just had to change verbose flag to true to see the report. We might want to put another querystring flag for this or default it to true for DEBUG.

While the new code is now more than 100x faster, ~150ms vs ~1.3ms, you don't notice because loading images and rerendering the character takes much longer. May as well keep the faster code though.