RFR: 8353230: Emoji rendering regression after JDK-8208377

[Daniel Gredler]

On Tue, 29 Apr 2025 09:55:56 GMT, Nikita Gubarkov <ngubarkov at openjdk.org> wrote:

>> It looks like this regression actually fits into a longer series of fixes / regressions in this area:
>> 
>> - [JDK-4517298](https://bugs.openjdk.org/browse/JDK-4517298) fixed metrics for zero-width characters, but broke some ligatures / glyph substitutions
>> - [JDK-7017058](https://bugs.openjdk.org/browse/JDK-7017058) fixed the ligatures / glyph substitutions, but broke some zero-width metrics
>> - [JDK-8208377](https://bugs.openjdk.org/browse/JDK-8208377) fixed some metrics and rendering for zero-width characters, but broke some ligatures / glyph substitutions
>> - Now, with this PR, we aim to fix the ligatures without re-breaking zero-width metrics and display
>> 
>> We have two different types of use cases pulling `CharToGlyphMapper` in two different directions: the users who need raw, untransformed glyph info, and the users who need normalized / transformed glyph info.
>> 
>> It looks to me like, in the current code base, the only `CharToGlyphMapper` user which requires raw font data is HarfBuzz (explicitly confirmed with the HarfBuzz team here: https://github.com/harfbuzz/harfbuzz/discussions/5234).
>> 
>> The regression mechanism at play here is that the HarfBuzz font callbacks are currently providing HarfBuzz with transformed glyph info (e.g. ZWJ -> INVISIBLE_GLYPH_ID), which prevents HarfBuzz from recognizing and applying the correct font GSUB substitutions (which involve ZWJ).
>> 
>> In order to fix this without (yet again) breaking metrics and display behavior elsewhere, I've added two methods to `CharToGlyphMapper` which provide access to raw glyph info, to be used by the HarfBuzz font callbacks: `charToGlyphRaw(int)` and `charToVariationGlyphRaw(int)`.
>> 
>> Note two intricacies related to `CompositeGlyphMapper`:
>> 1. We need to be careful to only cache raw (untransformed) values, to avoid conflicts between requests for a raw version of a glyph and a transformed version of the same glyph. Another option would have been two separate caches, but I don't think that's necessary.
>> 2. Consumers who are using `CompositeGlyphMapper.SLOTMASK` to check glyph slots (e.g. `FontRunIterator` and `CTextPipe`) will "see" invisible glyphs as having come from slot 0. This isn't new, and I think it's OK, but something to be aware of.
>> 
>> The glyph cache handling in `CCharToGlyphMapper` (for macOS) also requires care to avoid mixing value types.
>> 
>> Please also note that I'm not sure if the tweak to `sunFont.c` is being tested, since FFM is being used by default for Harf...
>
> By the way, I see that in each implementation, both `charToGlyph` and `charToGlyphRaw` call a common method, like `getGlyph(int uniciode, boolean raw)`. At first there was just `charToGlyph`, then `charToVariationGlyph` was added and now you added a "raw" version for each of them, I see that in the future we will need other variants and how it's already starting an exponential explosion. Overriding all of those methods in each implementation brings quite a bit of boilerplate, and it becomes easier to miss something. Maybe take a step back and refactor this into a single `charToGlyph(int unicode, int variationSelector, boolean raw)` version?
> Also, this `raw` parameter only really controls `isDefaultIgnorable` check in the end of each method. Maybe we could factor this out without bringing it separately into each mapper implementation?

@YaaZ: Thanks for the additional feedback, please see my thoughts below:

> By the way, I see that in each implementation, both charToGlyph and charToGlyphRaw call a common method, like getGlyph(int uniciode, boolean raw). At first there was just charToGlyph, then charToVariationGlyph was added and now you added a "raw" version for each of them, I see that in the future we will need other variants and how it's already starting an exponential explosion.

I don't know if I would call two changes to `CharToGlyphMapper` in 20 years an exponential explosion, but I get your point :-)

> Overriding all of those methods in each implementation brings quite a bit of boilerplate, and it becomes easier to miss something.

True, but again keep in mind that there are only 5 implementations, only one of which (the macOS `CCharToGlyphMapper`) has been added in the last 20 years.

> Maybe take a step back and refactor this into a single charToGlyph(int unicode, int variationSelector, boolean raw) version?

We'd still need separate methods for `int` vs. `char`, but I think this might reduce 5 methods down to 3? The changeset would be a bit more intrusive (lots of callers would need to change to reflect the new method signature). I'd be interested to hear thoughts from some of the reviewers on this one.

> Also, this raw parameter only really controls isDefaultIgnorable check in the end of each method. Maybe we could factor this out without bringing it separately into each mapper implementation?

I prefer to think of it as controlling whether or not any transformations to `INVISIBLE_GLYPH_ID` happen (right now it's just for default-ignorable characters, but there may be other scenarios in the future, e.g. `\r`, `\n` and `\t` which currently are handled elsewhere).

Any ideas for what this refactoring might look like?

-------------

PR Comment: https://git.openjdk.org/jdk/pull/24412#issuecomment-2845646011