--skip_if_larger does nothing and returns an error if pngquant would
generate a larger file than the original. Formerly we would not copy
the file in this case so later operations expecting the file would fail.
No image triggered this, though, so the issue went unnoticed. We want
the smaller of the two files. It's unclear if later compression using
optipng would still do better with the larger quantized file vs the
original unquantized file, but we need to have a file.
Now optionally takes '-c' flag to set background colors for the last
emoji set. Colors are 6-digit rgb hex values (for css after '#').
When more than one color is specified, the last emoji set is repeated
across columns, one for each color. '-c' with no arguments defaults
to a set of 11 colors. Omitting the flag uses the standard background
color.
- includes aliases
- checks coverage of sequences (assumes full coverage of all unicode
emoji and sequences for now)
- reports sequence names
(Some of this code needs to be shuffled into other places, sequence name
lookup and emoji_vs stripping doesn't belong here since these operations
are more generally useful. That will come.)
Emoji image files we get from upstream sometimes use the 'canonical'
sequences from the unicode data, which can contain emoji variation
selectors in the sequence. For our image data we wish to ignore
variation selectors. This tool renames files in a directory to the
corresponding sequence without emoji variation selectors, so that
other tooling doesn't need to account for them.
This maps a sequence of codepoints to another sequence, where the
first sequence should be an alias to the second.
Initial data contains emoji sequences where we expect an image named
according to the second sequence, but want to support the first
sequence with the same image.
- supports checking files with other extension besides .png
- checks all files under a root directory and not just the
files directly in a directory
- checks for duplicate files in multiple directories under a root
- reports the directory containing a file when there are problems
The generated html references images in multiple directories that
might not be in any defined location relative to the html file.
For sharing the results it's convenient if we have the images
and html file under the same parent directory. This option computes
the necessary images and copies them to directories under the directory
into which the html file is written, and makes the html file reference
the files in these new locations.
In addition, this removes some clutter from the generated table by
using the nth-last-of-type pseudo-class selector instead of tagging
all the cells in a column with a class name.
Previously our copy of waveflag took just an input and output filename.
Upstream takes a prefix and one or more input filenames, and concatenates
the prefix to the input filename as the output.
The makefile is changed to pass a prefix and the input filename, instead
of the input filename and the output filename as it formerly did.
Unfortunately for us, our inputs have a directory prefix since they're
not in the current directory, and we don't want this prefix in the output
file path. So we tweak our copy of waveflag.c to call basename on the
input file path before we append it to the prefix.
We also make the tool a little less noisy by putting more printfs
under the debug flag.
This adds an argument to generate_emoji_html that specifies
a file listing codepoint sequences. Emoji matching one of these
codepoint sequences are highlighted in the output.
When waveflag.c was forked from behdad's repo one of the changes that
was made effectively removed the ability to generate different sizes
of flags, despite leaving the SIZE value in the source. Recently we
needed to generate waved flags at a different power-of-two size, and
found it no longer functioned as the original. These changes restore
that while leaving the other changes in this forked version (mostly
formatting changes) intact.
It used to be difficult to find a sequence since the codepoints weren't
provided, just the images. This provides the codepoint list as
the 'name' of the sequence.
This also makes some other changes:
- the python template system doesn't like keyword names that have have
hyphens, so rename font-face-style to fontFaceStyle to get around this.
Thought this had been fixed earlier, but apparently it didn't end up in
a pushed commit.
- no longer insert emoji variation selector after some characters.
This was done to see what difference it made in browser behavior, but
we think now that browsers should be able to handle these sequences
without the selectors present.
- use a flag to pass name of output html file, rather than taking it as
a direct arg. other flags take multiple args, and if the html file
name comes after one of those, it gets swallowed by the other flag,
so it has to come first. This lets you put the file name anywhere
in the parameter list, the flag acts as a delimiter.
Images are shown in LTR and RTL contexts. Chrome currently doesn't
correctly render some emoji sequences, in part this is because it is
using Unicode 8 property data. At any rate, these are known Chrome
issues.
To handle forming emoji 'ligatures' in RTL contexts we generate
reversed ligature sequences for the GSUB table. Formerly we only did
this when there was a ZWJ in the sequence, and full reversal worked
because we had no sequences with both fitzpatrick modifiers and ZWJ.
However, now we do. Harfbuzz treats fitzpatrick modifiers as though
they were combining marks and so we need to as well so that GSUB data
is in the order Harfbuzz expects. So we 'unreverse' these pairs.
By default the tool uses all sequences that appear in any of the image
sets. To make it easier to see just the changes between a smaller
set of images and a large one, this lets you limit the sequences
to just those in the first set being compared.
Doug Felt