Re: Unicode 15

[समीर सिंह Sameer Singh]

See the code: those are failing line numbers and character numbers.

When the test fails, you have a procedure to update those so it
passes:

  ;; If there are lines marked as failing (see
  ;; `ucs-normalize-tests--failing-lines-part1' and
  ;; `ucs-normalize-tests--failing-lines-part2'), they may need to be
  ;; adjusted when NormalizationTest.txt is updated.  Run the function
  ;; `ucs-normalize-check-failing-lines' to see what changes are needed.

Note the last sentence.  admin/notes/unicode mentions that comment
where it describes how to run this test.

Thanks, the tests are passing now.

 As for COLRv1, I think it's HarfBuzz and/or Cairo that need to support
that, no?  But I'm just guessing here.  One thing to try is to run
hb-view and see what it produces for these sequences.

I tried with the latest harfbuzz master and the hb-view is producing a blank file,

also this is what harfbuzz says about it COLRv1 support [1]:

HarfBuzz implementation will follow later. No major client relies on HarfBuzz for color fonts currently, but we certainly want to implement later as there are clients who like to remove FreeType dependency completely.

[1] https://github.com/googlefonts/colr-gradients-spec#harfbuzz

On Thu, Sep 15, 2022 at 11:26 AM Eli Zaretskii <eliz@gnu.org> wrote:

> From: समीर सिंह Sameer Singh <lumarzeli30@gmail.com>
> Date: Thu, 15 Sep 2022 06:56:48 +0530
> Cc: emacs-devel@gnu.org
>
> The lisp/international/ucs-normalize-tests.el is failing and the file says to adjust
> ucs-normalize-tests--failing-lines-part1 and ucs-normalize-tests--failing-lines-part2,
> but what do the numbers mean there?

See the code: those are failing line numbers and character numbers.

When the test fails, you have a procedure to update those so it
passes:

  ;; If there are lines marked as failing (see
  ;; `ucs-normalize-tests--failing-lines-part1' and
  ;; `ucs-normalize-tests--failing-lines-part2'), they may need to be
  ;; adjusted when NormalizationTest.txt is updated.  Run the function
  ;; `ucs-normalize-check-failing-lines' to see what changes are needed.

Note the last sentence.  admin/notes/unicode mentions that comment
where it describes how to run this test.

> Also, the latest noto emoji[1] is in a new format named COLRv1 which I guess is not yet supported by
> emacs therefore all of the emojis are appearing as blank, so we should support that in the future.
> in the meantime do you know of any other font which supports the newest emojis so that emoji support could
> be
> tested?

I don't know of any other fonts, sorry.

As for COLRv1, I think it's HarfBuzz and/or Cairo that need to support
that, no?  But I'm just guessing here.  One thing to try is to run
hb-view and see what it produces for these sequences.