add GB18030-2022 charmap and test the entire GB18030 charmap [BZ #30243]

support GB18030-2022 after add and change some transcoding relationship
of GB18030-2022.Details are as follows:
add 25 transcoding relationship
  UE81E 0x82359037
  UE826 0x82359038
  UE82B 0x82359039
  UE82C 0x82359130
  UE832 0x82359131
  UE843 0x82359132
  UE854 0x82359133
  UE864 0x82359134
  UE78D 0x84318236
  UE78F 0x84318237
  UE78E 0x84318238
  UE790 0x84318239
  UE791 0x84318330
  UE792 0x84318331
  UE793 0x84318332
  UE794 0x84318333
  UE795 0x84318334
  UE796 0x84318335
  UE816 0xfe51
  UE817 0xfe52
  UE818 0xfe53
  UE831 0xfe6c
  UE83B 0xfe76
  UE855 0xfe91
change 6 transcoding relationship
  U20087 0x95329031
  U20089 0x95329033
  U200CC 0x95329730
  U215D7 0x9536b937
  U2298F 0x9630ba35
  U241FE 0x9635b630
Test the entire GB18030 charmap, not only the Unicode BMP part.

Co-authored-by: yangyanchao <yangyanchao6@huawei.com>
Co-authored-by: liqingqing <liqingqing3@huawei.com>
Co-authored-by: Bruno Haible <bruno@clisp.org>
Reviewed-by: Andreas Schwab <schwab@suse.de>
Reviewed-by: Mike FABIAN <mfabian@redhat.com>

1 files changed, 34 insertions, 16 deletions

diff --git a/iconvdata/tst-table.sh b/iconvdata/tst-table.sh
index bc6f542b24..7ba15bbf5c 100755
--- a/iconvdata/tst-table.sh
+++ b/iconvdata/tst-table.sh
@@ -37,7 +37,8 @@ set -e
   < ../localedata/charmaps/${charmap:-$charset} \
   > ${objpfx}tst-${charset}.charmap.table
 # When the charset is GB18030, truncate this table because for this encoding,
-# the tst-table-from and tst-table-to programs scan the Unicode BMP only.
+# the charmap contains ranges (<Unnnn>..<Ummmm> notation), which the
+# tst-table-charmap.sh script does not grok.
 if test ${charset} = GB18030; then
   grep '0x....$' < ${objpfx}tst-${charset}.charmap.table \
     > ${objpfx}tst-${charset}.truncated.table
@@ -73,25 +74,42 @@ diff ${objpfx}tst-${charset}.charmap.table ${objpfx}tst-${charset}.inverse.table
 
 # Check 1: charmap and iconv forward should be identical, except for
 # precomposed characters.
-if test -f ${precomposed}; then
-  cat ${objpfx}tst-${charset}.table ${precomposed} | sort | uniq -u \
-    > ${objpfx}tst-${charset}.tmp.table
-  cmp -s ${objpfx}tst-${charset}.charmap.table ${objpfx}tst-${charset}.tmp.table ||
+{ if test -f ${precomposed}; then
+    cat ${objpfx}tst-${charset}.table ${precomposed} | sort | uniq -u
+  else
+    cat ${objpfx}tst-${charset}.table
+  fi
+} | { if test ${charset} = GB18030; then grep '0x....$'; else cat; fi; } \
+  > ${objpfx}tst-${charset}.tmp1.table
+cmp -s ${objpfx}tst-${charset}.charmap.table ${objpfx}tst-${charset}.tmp1.table ||
   exit 1
-else
-  cmp -s ${objpfx}tst-${charset}.charmap.table ${objpfx}tst-${charset}.table ||
-  exit 1
-fi
 
 # Check 2: the difference between the charmap and iconv backward.
-if test -f ${irreversible}; then
-  cat ${objpfx}tst-${charset}.charmap.table ${irreversible} | sort | uniq -u \
-    > ${objpfx}tst-${charset}.tmp.table
-  cmp -s ${objpfx}tst-${charset}.tmp.table ${objpfx}tst-${charset}.inverse.table ||
-  exit 1
-else
-  cmp -s ${objpfx}tst-${charset}.charmap.table ${objpfx}tst-${charset}.inverse.table ||
+{ if test -f ${irreversible}; then
+    cat ${objpfx}tst-${charset}.charmap.table ${irreversible} | sort | uniq -u
+  else
+    cat ${objpfx}tst-${charset}.charmap.table
+  fi
+} | { if test ${charset} = GB18030; then grep '0x....$'; else cat; fi; } \
+  > ${objpfx}tst-${charset}.tmp2c.table
+cat ${objpfx}tst-${charset}.inverse.table \
+  | { if test ${charset} = GB18030; then grep '0x....$'; else cat; fi; } \
+  > ${objpfx}tst-${charset}.tmp2i.table
+cmp -s ${objpfx}tst-${charset}.tmp2c.table ${objpfx}tst-${charset}.tmp2i.table ||
   exit 1
+
+# Check 3: the difference between iconv forward and iconv backward. This is
+# necessary only for GB18030, because ${objpfx}tst-${charset}.charmap.table
+# is truncated for this encoding (see above).
+if test ${charset} = GB18030; then
+  { if test -f ${irreversible}; then
+      cat ${objpfx}tst-${charset}.table ${irreversible} | sort | uniq -u
+    else
+      cat ${objpfx}tst-${charset}.table
+    fi
+  } > ${objpfx}tst-${charset}.tmp3.table
+  cmp -s ${objpfx}tst-${charset}.tmp3.table ${objpfx}tst-${charset}.inverse.table ||
+    exit 1
 fi
 
 exit 0