APNG support

[Poker-sang]

Prerequisites

• I have written a descriptive pull-request title
• I have verified that there are no overlapping pull-requests open
• I have verified that I am following the existing coding patterns and practice as demonstrated in the repository. These follow strict Stylecop rules 👮.
• I have provided test coverage for my change (where applicable)

Description

Implement APNG decoding and encoding based on PNG

[CLAassistant]

All committers have signed the CLA.

[JimBobSquarePants]

Thanks for this, I really appreciate your efforts!

I've tried to offer guidance as to what need to change. Please let me know if anything doesn't make sense.

[Poker-sang]

Thanks for your review! It is rigorous and easy for me to understand. I'm sorry for accidentally formatting some code that confused you (e.g. x++ to ++x). Now it's all reverted.

[JimBobSquarePants]

@Poker-sang Sorry this is taking so long to merge. There's a bit of a backlog we need to shift first to ensure there are no functional conflicts.

[Poker-sang]

It's all right. Take your time please.

[JimBobSquarePants]

@Poker-sang Thanks again for your patience. Following the merge of #2485 there's a few conflicts here. Would you be able to update your PR to include the updates from main? If not, I can have a try.

[Poker-sang]

I can have a try. But I'm sorry I can't handle it for I'm busy these days. If you're not in a hurry you can wait a few days for me. ☺️ Or you can handle it the way you want if it doesn't bother you.

[JimBobSquarePants]

I'm having a go at merging main back into this PR now.

[Poker-sang]

We've deliberately kept any metadata separate to highlight this and to allow unlimited metadata variations.

If we convert GIF to APNG, WEBP or other animated image formats, the delay information may be lost. Or we need to write appropriate conversion logic for each Metadata. 🤔

[JimBobSquarePants]

Yes, you would also lose any disposal or blend mode information

[Poker-sang]

Ok I see. Thanks for explaining.

[JimBobSquarePants]

@Poker-sang I've added support for alpha blending and disposal.

I think it's correct based on how we work with gif/webp but I could do with a second pair of eyes. We desperately need to add test images now to validate the implementation so if you could add that, it would be great!

[Poker-sang]

@JimBobSquarePants Thank you very much for your help! 😀 If you need some images to validate the implementation about alpha blending and disposal, I thought these images would help:

• Dispose ops: https://philip.html5.org/tests/apng/tests.html#apng-dispose-op-none-basic
• Blend ops: https://philip.html5.org/tests/apng/tests.html#apng-blend-op-source-on-solid-colour

[JimBobSquarePants]

@Poker-sang Can you give me collaboration access to your fork? There's an open issue with Git LFS which prevents me pushing images to your fork without those permissions.

git-lfs/git-lfs#5001

[Poker-sang]

Sure. When I'm cloning this repo lately, Filtering Content always takes me a long time. Is it caused by too many test images?

[JimBobSquarePants]

Sure. When I'm cloning this repo lately, Filtering Content always takes me a long time. Is it caused by too many test images?

I'm not sure what you mean by Filtering Content.

[Poker-sang]

I'm not sure what you mean by Filtering Content.

[JimBobSquarePants]

Ah yeah, there's a lot of LFS files.

You should only be cloning the repo 1 time really though and using feature branches for your PRs. Every time someone with a fork clones the repo it actually uses up my LFS bandwidth and ends up costing me money if I go over my pre-purchased allowance.

[Poker-sang]

Oh I see. I'll pay attention to it.

[JimBobSquarePants]

I’ve started adding the tests locally.

[Poker-sang]

APNG Specification says:

If the default image is the first frame:

Sequence number    Chunk
(none)             `acTL`
0                  `fcTL` first frame
(none)             `IDAT` first frame / default image
1                  `fcTL` second frame
2                  first `fdAT` for second frame
3                  second `fdAT` for second frame
....

If the default image is not part of the animation:

Sequence number    Chunk
(none)             `acTL`
(none)             `IDAT` default image
0                  `fcTL` first frame
1                  first `fdAT` for first frame
2                  second `fdAT` for first frame
....

Seems like a simple and not very useful feature, do we need to implement it? Says adding an option in the Encoder to exclude the first frame.

[JimBobSquarePants]

I think that just means the first frame control is optional. The code I have locally handles that.

[JimBobSquarePants]

I'm mostly happy with this now. All the test images I've thrown at it pass with flying colors. Will review again in the morning to be sure.

[Poker-sang]

I am very happy to hear that. I am so grateful that you have helped me so much!❤

[JimBobSquarePants]

I'm happy with this now. Thanks @Poker-sang for a great addition to the library!