ICU update to 54 step 3

source/data/mappings/ucmlocal.mk

+# Note: A number of encodings are handled with purely algorithmic converters,
+# without any mapping tables:
+# US-ASCII, ISO 8859-1, UTF-7/8/16/32, SCSU
+
+# Listed here:
+
+# * ISO 8859-2..8,10,13,14,15,16
+#   - 8859-11 table is not included. It's rather treated as a synonym of
+#     Windows-874
+# * Windows-125[0-8]
+# * Simplified Chinese : GBK(Windows cp936), GB 18030
+#   - GB2312 table was removed and 4 aliases for GB2312 were added
+#     to GBK in convrtrs.txt to treat GB2312 as a synonym of GBK.
+#   - GB-HZ is supported now that it uses the GBK table.
+# * Traditional Chinese : Big5
+# * Japanese : SJIS (shift_jis-html), EUC-JP (euc-jp-html)
+# * Korean : Windows-949
+#  - EUC-KR table was removed. It's different from Windows-949, but
+#    in practice EUC-KR and Windows-949 are treated synonymously.
+#  - ISO-2022-KR is now supported by with a one-line change
+#    in common/ucnv2022.c and other changes in convrtrs.txt to make it
+#    use the windows-949 table.
+# * Thai : Windows-874
+#   - TIS-620 and ISO-8859-11 are treated as synonyms of Windows-874
+#     although they're not the same.
+# * Mac encodings : MacRoman, MacCyrillic
+# * Cyrillic : KOI8-R, KOI8-U, IBM-866
+#
+# * Missing
+#  - Armenian, Georgian  : extremly rare
+#  - Mac encodings (other than Roman and Cyrillic) : extremly rare
+
+UCM_SOURCE_FILES=
+
+
+UCM_SOURCE_CORE=iso-8859-2-html.ucm iso-8859-3-html.ucm iso-8859-4-html.ucm\
+iso-8859-5-html.ucm iso-8859-6-html.ucm iso-8859-7-html.ucm \
+iso-8859-8-html.ucm iso-8859-10-html.ucm iso-8859-13-html.ucm \
+iso-8859-14-html.ucm iso-8859-15-html.ucm iso-8859-16-html.ucm \
+windows-1250-html.ucm windows-1251-html.ucm windows-1252-html.ucm\
+windows-1253-html.ucm windows-1254-html.ucm windows-1255-html.ucm\
+windows-1256-html.ucm windows-1257-html.ucm windows-1258-html.ucm\
+windows-936-2000.ucm gb18030.ucm big5-html.ucm\
+shift_jis-html.ucm euc-jp-html.ucm\
+windows-949-2000.ucm\
+windows-874-html.ucm \
+macintosh-html.ucm x-mac-cyrillic-html.ucm\
+ibm866-html.ucm koi8-r-html.ucm koi8-u-html.ucm
+
+
+
+# Do not build EBCDIC converters.
+# ibm-37 and ibm-1047 are hardcoded in Makefile.in and
+# they're removed by modifying the file. It's also hard-coded in makedata.mak for
+# Winwodws, but we don't have to touch it because the data dll is generated out of
+# icu*.dat file generated on Linux.
+UCM_SOURCE_EBCDIC =