nerd-fonts/bin/scripts/gotta-patch-em-all-font-patcher!.sh

387 lines
14 KiB
Bash
Raw Normal View History

2018-07-09 16:20:40 +02:00
#!/usr/bin/env bash
2024-04-12 16:31:51 +02:00
# Nerd Fonts Version: 3.2.1
# Script Version: 1.4.5
#
# You can supply options to the font-patcher via environment variable NERDFONTS
# That option will override the defaults (also defaults of THIS script).
# used for debugging
# set -x
2017-05-14 00:02:53 +02:00
LINE_PREFIX="# [Nerd Fonts] "
test "${BASH_VERSION%%[^0-9]*}" -ge 4 || {
echo >&2 "$LINE_PREFIX A non-ancient version of Bash is needed (>= 4)"
echo >&2 "# Bash version 4 has been released in 2009, so it's about time to update"
echo >&2 "# (Most likely you are on MacOS; try Homebrew with \`brew install bash\`) ;-)"
exit 1
}
# Check for Fontforge
type fontforge >/dev/null 2>&1 || {
2017-05-14 00:02:53 +02:00
echo >&2 "$LINE_PREFIX FontForge must be installed before running this script."
echo >&2 "# Please see installation instructions at"
echo >&2 "# http://designwithfontforge.com/en-US/Installing_Fontforge.html"
exit 1
}
# Get script directory to set source and target dirs relative to it
sd="$( cd -- "$(dirname "$0")" >/dev/null 2>&1 || exit ; pwd -P )"
repo_root_dir=$(dirname "$(dirname "${sd}")") # two levels up (i.e. ../../)
# Set source and target directories
like_pattern='.*\.\(otf\|ttf\|sfd\)'
last_font_root=""
unpatched_parent_dir="src/unpatched-fonts"
patched_parent_dir="patched-fonts"
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
timestamp_parent_dir=${patched_parent_dir}
source_fonts_dir="${repo_root_dir}/${unpatched_parent_dir}"
max_parallel_process=8
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
function activate_keeptime {
type ttfdump >/dev/null 2>&1 || {
echo >&2 "$LINE_PREFIX ttfdump must be installed for option --keeptime"
exit 1
}
keeptime=TRUE
}
function activate_checkfont {
patched_parent_dir="check-fonts"
}
function activate_info {
info_only=TRUE
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
echo "${LINE_PREFIX} 'Info Only' option given, only generating font info (not patching)"
}
function show_help {
echo "Usage: $0 [OPTION] [FILTER]"
echo
echo " OPTION:"
echo " -c, --checkfont Create the font(s) in check-fonts/ instead"
echo " -t, --keeptime Try to preserve timestamp of previously patched"
echo " font in patched-fonts/ directory"
echo " -v, --verbose Show more information when running"
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
echo " -i, --info Rebuild JUST the readmes"
echo " -j, --jobs Run up to 8 patch processes in parallel"
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
echo " -h, --help Show this help"
echo
echo " FILTER:"
echo " The filter argument to this script is a filter for the fonts to patch."
echo " The filter is a regex (glob "*" is expressed as "[^/]*", see \`man 7 glob\`)"
echo " All font files that start with that filter (and are ttf, otf, or sfd files) will"
echo " be processed only."
echo " Example ./gotta-patch-em-all-font-patcher\!.sh \"iosevka\""
echo " Process all font files that start with \"iosevka\""
echo " If the argument starts with a '/' all font files in a directory that matches"
echo " the filter are processed only."
echo " Example ./gotta-patch-em-all-font-patcher\!.sh \"/iosevka\""
echo " Process all font files that are in directory \"iosevka\""
}
function find_font_root {
# e.g. /a/b/c/nerd-fonts/src/unpatched-fonts/Meslo
sed -E "s|(${unpatched_parent_dir}/[^/]*).*|\1|" <<< "$1"
}
while getopts ":chijtv-:" option; do
case "${option}" in
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
c)
activate_checkfont
;;
h)
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
show_help
exit 0;;
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
i)
activate_info
;;
j)
parallel=TRUE
;;
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
t)
activate_keeptime
;;
v)
verbose=TRUE
;;
-)
case "${OPTARG}" in
checkfont)
activate_checkfont
;;
help)
show_help
exit 0;;
info)
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
activate_info
;;
jobs)
parallel=TRUE
;;
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
keeptime)
activate_keeptime
;;
verbose)
verbose=TRUE
;;
*)
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
echo >&2 "Option '--${OPTARG}' unknown"
exit 1;;
esac;;
*)
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
echo >&2 "Option '-${OPTARG}' unknown"
exit 1;;
esac
done
shift $((OPTIND-1))
if [ $# -gt 1 ]
then
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
echo >&2 "Unknown parameter(s): $2 ..."
exit 1
fi
if [ $# -eq 1 ]
then
if [[ "${1:0:1}" == "/" ]]
then
like_pattern=".*$1/.*\.\(otf\|ttf\|sfd\)"
echo "$LINE_PREFIX Filter given, limiting search and patch to pathname pattern '$1'"
else
like_pattern=".*/$1[^/]*\.\(otf\|ttf\|sfd\)"
echo "$LINE_PREFIX Filter given, limiting search and patch to filename pattern '$1'"
fi
fi
# correct way to output find results into an array (when files have space chars, etc)
2017-07-28 22:29:44 +02:00
# source: https://stackoverflow.com/questions/8213328/bash-script-find-output-to-array
source_fonts=()
while IFS= read -d $'\0' -r file ; do
source_fonts=("${source_fonts[@]}" "$file")
done < <(find "$source_fonts_dir" -iregex "${like_pattern}" -type f -print0)
# print total number of source fonts found
2017-05-14 00:02:53 +02:00
echo "$LINE_PREFIX Total source fonts found: ${#source_fonts[*]}"
# Use one date-time for ALL fonts and for creation and modification date in the font file
if [ -z "${SOURCE_DATE_EPOCH}" ]
then
export SOURCE_DATE_EPOCH=$(date +%s)
fi
release_timestamp=$(date -R "--date=@${SOURCE_DATE_EPOCH}" 2>/dev/null) || {
echo >&2 "$LINE_PREFIX Invalid release timestamp SOURCE_DATE_EPOCH: ${SOURCE_DATE_EPOCH}"
exit 2
}
echo "$LINE_PREFIX Release timestamp is ${release_timestamp}"
function patch_font {
local f=$1; shift
local i=$1; shift
local purge=$1; shift
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
# Try to copy the release date from the 'original' patch
if [ -n "${keeptime}" ]
then
# take everything before the last slash (/) to start building the full path
local ts_font_dir="${f%/*}/"
local ts_font_dir="${ts_font_dir/$unpatched_parent_dir/$timestamp_parent_dir}"
local one_font=$(find "${ts_font_dir}" -name '*.[ot]tf' | head -n 1)
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
if [ -n "${one_font}" ]
then
orig_font_date=$(ttfdump -t head "${one_font}" | \
grep -E '[^a-z]modified:.*0x' | sed 's/.*x//' | tr 'a-f' 'A-F')
SOURCE_DATE_EPOCH=$(dc -e "16i ${orig_font_date} Ai 86400 24107 * - p")
echo "$LINE_PREFIX Release timestamp adjusted to $(date -R "--date=@${SOURCE_DATE_EPOCH}")"
gotta-patch-em-all: Add options to simplify patcher testing [why] When working on the font-patcher the developer needs to test the changes on a number of fonts. This is usually a manual call of `font-patcher` and afterwards a 'diff' of the newly created font with the 'old' font in the patched-fonts/ directory with fontforge (which has a font-compare option). If you run gotta-patch-em-all normally the newly generated font will replace the existing font and git will ALWAYS show it as different. The reason is that at least the timestamp in the generated font has changed. Far more easy would be if the new gotta-patch-em-all run could keep the previous timestamps, in that way one can immediately see that the old and new fonts are bitwise equal (via git). Furthermore if you expect a change and want to show the differences of old and new font in fontforge you need both fonts in the filesystem. But a normal gotta-patch-em-all run replaces the font. A different destination folder would help here. [how] Introduce two new (independent) options to a) keep the timestamp equal to previous patch run b) generate the fonts in a different directory While b) is straight forward, a) is a bit more complicated, esp because filenames can change and so on. So the script examines just one (1) random font in the specific font directory and uses its timestamp. In most cases this is correct enough if the developer uses gotta-patch-em-all consequently. Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>
2023-01-05 09:23:36 +01:00
fi
fi
# take everything before the last slash (/) to start building the full path
local patched_font_dir="${f%/*}/"
# find replace unpatched parent dir with patched parent dir:
local patched_font_dir="${patched_font_dir/$unpatched_parent_dir/$patched_parent_dir}"
[[ -d "$patched_font_dir" ]] || mkdir -p "$patched_font_dir"
if [ -n "${purge}" ]
then
if [ -n "${verbose}" ]
then
echo "Purging patched font dir ${patched_font_dir}"
fi
rm -- "${patched_font_dir}"/*
fi
config_parent_dir=$( cd "$( dirname "$f" )" && cd ".." && pwd)
config_dir=$( cd "$( dirname "$f" )" && pwd)
# find the font config file:
if [ -f "$config_dir/config.cfg" ]
then
font_config="--configfile=$config_dir/config.cfg"
elif [ -f "$config_parent_dir/config.cfg" ]
then
font_config="--configfile=$config_parent_dir/config.cfg"
elif [ -f "$(find_font_root "$config_parent_dir")/config.cfg" ]
then
font_config="--configfile=$(find_font_root "$config_parent_dir")/config.cfg"
else
# We need to give some argument because empty arguments will break the patcher call
font_config="-q"
fi
cd "$repo_root_dir" || {
echo >&2 "# Could not find project parent directory"
exit 3
}
# Add logfile always (but can be overridden by config.cfg and env var NERDFONTS)
# Use absolute path to allow fontforge being an AppImage (used in CI)
PWD=$(pwd)
# Create "Nerd Font"
if [ -n "${verbose}" ]
then
echo "fontforge -quiet -script \"${PWD}/font-patcher\" --debug 1 \"$f\" -q \"${font_config}\" -c --no-progressbars --outputdir \"${patched_font_dir}\" ${NERDFONTS}"
fi
# shellcheck disable=SC2086 # We want splitting for the unquoted variables to get multiple options out of them
{ OUT=$(fontforge -quiet -script "${PWD}/font-patcher" --debug 1 "$f" -q "${font_config}" -c --no-progressbars \
--outputdir "${patched_font_dir}" ${NERDFONTS} 2>&1 1>&3 3>&- ); } 3>&1
# shellcheck disable=SC2181 # Checking the code directly is very unreadable here, as we execute a whole block
if [ $? -ne 0 ]; then printf "%s\nPatcher run aborted!\n\n" "$OUT"; fi
# Create "Nerd Font Mono"
if [ -n "${verbose}" ]
then
echo "fontforge -quiet -script \"${PWD}/font-patcher\" --debug 1 \"$f\" -q -s \"${font_config}\" -c --no-progressbars --outputdir \"${patched_font_dir}\" ${NERDFONTS}"
fi
# shellcheck disable=SC2086 # We want splitting for the unquoted variables to get multiple options out of them
{ OUT=$(fontforge -quiet -script "${PWD}/font-patcher" --debug 1 "$f" -q -s "${font_config}" -c --no-progressbars \
--outputdir "${patched_font_dir}" ${NERDFONTS} 2>&1 1>&3 3>&- ); } 3>&1
# shellcheck disable=SC2181 # Checking the code directly is very unreadable here, as we execute a whole block
if [ $? -ne 0 ]; then printf "%s\nPatcher run aborted!\n\n" "$OUT"; fi
# Create "Nerd Font Propo"
if [ -n "${verbose}" ]
then
echo "fontforge -quiet -script \"${PWD}/font-patcher\" --debug 1 \"$f\" -q --variable \"${font_config}\" -c --no-progressbars --outputdir \"${patched_font_dir}\" ${NERDFONTS}"
fi
# shellcheck disable=SC2086 # We want splitting for the unquoted variables to get multiple options out of them
{ OUT=$(fontforge -quiet -script "${PWD}/font-patcher" --debug 1 "$f" -q --variable "${font_config}" -c --no-progressbars \
--outputdir "${patched_font_dir}" ${NERDFONTS} 2>&1 1>&3 3>&- ); } 3>&1
# shellcheck disable=SC2181 # Checking the code directly is very unreadable here, as we execute a whole block
if [ $? -ne 0 ]; then printf "%s\nPatcher run aborted!\n\n" "$OUT"; fi
# wait for this group of background processes to finish to avoid forking too many processes
# that can add up quickly with the number of combinations
#wait
}
# Generates font information: readmes, combinations, licenses, and variation counts
# $1 = fontdir path
# $2 = font file name (used for metadata)
function generate_info {
local f=$1; shift
local font_file=$1; shift
# take everything before the last slash (/) to start building the full path
local patched_font_dir="${f%/*}/"
# find replace unpatched parent dir with patched parent dir:
local patched_font_dir="${patched_font_dir/$unpatched_parent_dir/$patched_parent_dir}"
echo "$LINE_PREFIX Generating info for '$font_file':"
[[ -d "$patched_font_dir" ]] || mkdir -p "$patched_font_dir"
local font_root=$(echo "$patched_font_dir" | sed "s|.*$patched_parent_dir/||;s|/.*||")
# if first time with this font then re-build parent dir readme, else skip:
if [ "$last_font_root" != "$font_root" ]
then
echo "$LINE_PREFIX --- Calling standardize-and-complete-readmes for $font_root"
"${sd}/standardize-and-complete-readmes.sh" "$font_root" "$patched_parent_dir"
echo "$LINE_PREFIX ---"
last_font_root=$font_root
fi
# Copy 'all' license files found in the complete font`s source tree
# into the destination. This will overwrite all same-names files
# so make sure all licenses of one fontface are identical
echo "$LINE_PREFIX * Copying license files"
current_dir=$(dirname "$f")
copy_license "$(find_font_root "$current_dir")" "$patched_font_dir"
}
# Copy any license file to the patched font directory
# $1 = fontdir source path
# $2 = fontdir destination path
function copy_license {
local src=$1
local dest=$2
local license_file=""
while IFS= read -d $'\0' -r license_file ; do
[[ -d "$dest" ]] || mkdir -p "$dest"
cp "$license_file" -t "$dest"
done < <(find "$src" -iregex ".*\(licen[cs]e\|ofl\).*" -type f -print0)
# To check which files will or will not be copied and make sure all relevant
# licences do match
# find src/unpatched-fonts -not -iname "*.[ot]tf" -type f -not -name 'README.md' -not -name 'config.cfg' -iregex ".*\(licen[cs]e\|ofl\).*"
# find src/unpatched-fonts -not -iname "*.[ot]tf" -type f -not -name 'README.md' -not -name 'config.cfg' -not -iregex ".*\(licen[cs]e\|ofl\).*"
}
2018-01-13 04:20:52 +01:00
if [ ! "$info_only" ]
then
# Iterate through source fonts
for i in "${!source_fonts[@]}"
do
purge_destination=""
current_source_dir=$(dirname "${source_fonts[$i]}")
if [ "${current_source_dir}" != "${last_source_dir}" ]
then
# If we are going to patch ALL font files from a certain source directory
# the destination directory is purged (all font files therein deleted)
# to follow font naming changed. We can not do this if we patch only
# some of the source font files in that directory.
last_source_dir=${current_source_dir}
num_to_patch=$(find "${current_source_dir}" -iregex "${like_pattern}" -type f | wc -l)
num_existing=$(find "${current_source_dir}" -iname "*.[ot]tf" -o -iname "*.sfd" -type f | wc -l)
if [ "${num_to_patch}" -eq "${num_existing}" ]
then
purge_destination="TRUE"
fi
fi
echo "$LINE_PREFIX Processing font $((i+1))/${#source_fonts[@]}"
if [ -n "${parallel}" ]
then
patch_font "${source_fonts[$i]}" "$i" "$purge_destination" 2>/dev/null &
else
patch_font "${source_fonts[$i]}" "$i" "$purge_destination" 2>/dev/null
fi
# un-comment to test this script (patch 1 font)
#break
# wait for this set of bg commands to finish: dont do too many at once!
# if we spawn a background process for each set of fonts it will
# end up using too many system resources
# however we want to run a certain number in parallel to decrease
# the amount of time patching all the fonts will take
# for now set a 'wait' for each X set of processes:
if [[ $(((i + 1) % max_parallel_process)) == 0 ]];
then
wait
fi
done
# wait for all bg commands to finish
wait
fi
# update information in separate iteration (to avoid issues with bg processes and the counts):
# Iterate through source fonts
for i in "${!source_fonts[@]}"
do
2017-07-29 19:50:37 +02:00
# only output after last slash (/):
path=${source_fonts[$i]}
font_file=${path##*/}
generate_info "$path" "$font_file" 2>/dev/null
done