| 1 |
commit: 78802b1270bf48bea2e39039877096c4867afbb3 |
| 2 |
Author: Mike Frysinger <vapier <AT> gentoo <DOT> org> |
| 3 |
AuthorDate: Mon Sep 27 06:31:11 2021 +0000 |
| 4 |
Commit: Mike Frysinger <vapier <AT> gentoo <DOT> org> |
| 5 |
CommitDate: Mon Sep 27 06:31:11 2021 +0000 |
| 6 |
URL: https://gitweb.gentoo.org/proj/locale-gen.git/commit/?id=78802b12 |
| 7 |
|
| 8 |
skip duplicate locales when normalized |
| 9 |
|
| 10 |
If you try to generate "en_US.UTF-8 UTF-8" and "en_US.UTF8 UTF-8", we |
| 11 |
end up generating the locale twice even though it normalizes down to |
| 12 |
the same value (or, in the case of --update, we skip the 1st and then |
| 13 |
always generate the 2nd). Keep track of all locales we've processed |
| 14 |
so we can dedupe as we go. |
| 15 |
|
| 16 |
Signed-off-by: Mike Frysinger <vapier <AT> gentoo.org> |
| 17 |
|
| 18 |
locale-gen | 33 +++++++++++++++++++++------------ |
| 19 |
1 file changed, 21 insertions(+), 12 deletions(-) |
| 20 |
|
| 21 |
diff --git a/locale-gen b/locale-gen |
| 22 |
index d313c23..d4417d8 100755 |
| 23 |
--- a/locale-gen |
| 24 |
+++ b/locale-gen |
| 25 |
@@ -192,9 +192,9 @@ if [[ -z ${locales_to_generate} ]] && [[ -e ${CONFIG} ]] ; then |
| 26 |
fi |
| 27 |
fi |
| 28 |
|
| 29 |
-# Transform the name in locales.gen to the name used when storing |
| 30 |
-# the locale data in /usr/lib/locale/ ... this normalize algo is |
| 31 |
-# taken out of the glibc localedef source code ... |
| 32 |
+# Transform the name in locales.gen to the name used when storing the locale data in |
| 33 |
+# /usr/lib/locale/. This normalize algo is taken out of the glibc localedef source: |
| 34 |
+# https://sourceware.org/git/?p=glibc.git;a=blob;f=locale/programs/localedef.c;hb=glibc-2.34#l562 |
| 35 |
normalize() { |
| 36 |
if [[ $1 == *.* ]] ; then |
| 37 |
local ret=$(echo ${1##*.} | tr '[[:upper:]]' '[[:lower:]]') |
| 38 |
@@ -316,6 +316,10 @@ JOB_IDX_S=0 |
| 39 |
JOB_IDX_E=0 |
| 40 |
cnt=0 |
| 41 |
lidx=0 |
| 42 |
+# Keep track of (normalized) locales generated in case the request has different inputs that |
| 43 |
+# normalize down to the same value. We trim $existing_locales as we go for later use which |
| 44 |
+# prevents its direct use. |
| 45 |
+generated_locales=" " |
| 46 |
while [[ -n ${locales_to_generate[${lidx}]} ]] ; do |
| 47 |
: $(( ++cnt )) |
| 48 |
locale=${locales_to_generate[$((lidx++))]} |
| 49 |
@@ -334,16 +338,21 @@ while [[ -n ${locales_to_generate[${lidx}]} ]] ; do |
| 50 |
|
| 51 |
disp=${locales_disp[$(( cnt - 1 ))]} |
| 52 |
|
| 53 |
- if [[ -n ${UPDATE} ]] ; then |
| 54 |
- normalized_locale=$(normalize ${locale}) |
| 55 |
- if [[ ${existing_locales} == *" ${normalized_locale} "* ]] ; then |
| 56 |
- existing_locales=${existing_locales/ ${normalized_locale} / } |
| 57 |
- if [[ ${QUIET} -eq 0 ]] ; then |
| 58 |
- cnt_fmt=$(printf "%${#total}i" ${cnt}) |
| 59 |
- einfo " (${cnt_fmt}/${total}) Skipping ${disp}" |
| 60 |
- fi |
| 61 |
- continue |
| 62 |
+ normalized_locale=$(normalize ${locale}) |
| 63 |
+ if [[ ${generated_locales} == *" ${normalized_locale} "* ]] ; then |
| 64 |
+ already_generated="true" |
| 65 |
+ else |
| 66 |
+ already_generated="false" |
| 67 |
+ fi |
| 68 |
+ generated_locales+="${normalized_locale} " |
| 69 |
+ if ${already_generated} || \ |
| 70 |
+ [[ -n ${UPDATE} && ${existing_locales} == *" ${normalized_locale} "* ]] ; then |
| 71 |
+ existing_locales=${existing_locales/ ${normalized_locale} / } |
| 72 |
+ if [[ ${QUIET} -eq 0 ]] ; then |
| 73 |
+ cnt_fmt=$(printf "%${#total}i" ${cnt}) |
| 74 |
+ einfo " (${cnt_fmt}/${total}) Skipping ${disp}" |
| 75 |
fi |
| 76 |
+ continue |
| 77 |
fi |
| 78 |
|
| 79 |
# If the locale is like 'en_US.UTF8', then we really want 'en_US' |
Archives