Override some chars marked as control in unicode tables but are

expected to be found in normal text files.
gwsw · Jul 25, 2017 · 18e0088 · 18e0088
1 parent b4f5fe1
commit 18e0088
Show file tree

Hide file tree

Showing 4 changed files with 17 additions and 4 deletions.
diff --git a/compose.uni b/compose.uni
@@ -1,4 +1,4 @@
-/* Generated by "./mkutable -f2 Mn Me Cf -- unicode/UnicodeData.txt" on Mon Jun 26 10:40:58 PDT 2017 */
+/* Generated by "./mkutable -f2 Mn Me Cf -- unicode/UnicodeData.txt" on Tue Jul 25 09:04:35 PDT 2017 */
  { 0x00ad, 0x00ad }, /* Cf */
  { 0x0300, 0x036f }, /* Mn */
  { 0x0483, 0x0487 }, /* Mn */

diff --git a/mkutable b/mkutable
@@ -11,6 +11,16 @@ use vars qw( $opt_f $opt_n );
 use Getopt::Std;
 my $type_field = 2;
 
+# Override Unicode tables for certain control chars
+# that are expected to be found in normal text files.
+my %force_space = (
+ 0x08 => 1, # backspace
+ 0x09 => 1, # tab
+ 0x0a => 1, # newline
+ 0x0c => 1, # form feed
+ 0x0d => 1, # carriage return
+);
+
 exit (main() ? 0 : 1);
 
 sub main {
@@ -47,6 +57,7 @@ sub main {
  my $type = $fields[$type_field];
  $type =~ s/\s//g;
  for ($last_code = $lo_code; $last_code <= $hi_code; ++$last_code) {
+ $type = 'Zs' if $force_space{$last_code};
  output(\%out, $last_code, $type);
  }
  }

diff --git a/ubin.uni b/ubin.uni
@@ -1,5 +1,7 @@
-/* Generated by "./mkutable -f2 Cc Cs Co Zl Zp -- unicode/UnicodeData.txt" on Mon Jun 26 10:40:58 PDT 2017 */
- { 0x0000, 0x001f }, /* Cc */
+/* Generated by "./mkutable -f2 Cc Cs Co Zl Zp -- unicode/UnicodeData.txt" on Tue Jul 25 09:04:35 PDT 2017 */
+ { 0x0000, 0x0007 }, /* Cc */
+ { 0x000b, 0x000b }, /* Cc */
+ { 0x000e, 0x001f }, /* Cc */
  { 0x007f, 0x009f }, /* Cc */
  { 0x2028, 0x2028 }, /* Zl */
  { 0x2029, 0x2029 }, /* Zp */

diff --git a/wide.uni b/wide.uni
@@ -1,4 +1,4 @@
-/* Generated by "./mkutable -f1 W F -- unicode/EastAsianWidth.txt" on Mon Jun 26 10:40:58 PDT 2017 */
+/* Generated by "./mkutable -f1 W F -- unicode/EastAsianWidth.txt" on Tue Jul 25 09:04:35 PDT 2017 */
  { 0x1100, 0x115f }, /* W */
  { 0x231a, 0x231b }, /* W */
  { 0x2329, 0x232a }, /* W */