From 50b320952e99ea20f9b77c6c501280fe37fd2598 Mon Sep 17 00:00:00 2001
From: NIIBE Yutaka <gniibe@fsij.org>
Date: Wed, 15 Apr 2020 14:10:08 +0900
Subject: [PATCH] regexp: Fix generation of _unicode_mapping.c.

* configure.ac (AWK_HEX_NUMBER_OPTION): Detect GNU Awk.
* regexp/Makefile.am: Use AWK_HEX_NUMBER_OPTION.
* regexp/parse-unidata.awk: Don't use strtonum.

GnuPG-bug-id: 4915
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
---
 configure.ac             | 10 ++++++++++
 regexp/Makefile.am       |  3 ++-
 regexp/parse-unidata.awk |  6 +++---
 3 files changed, 15 insertions(+), 4 deletions(-)

diff --git a/configure.ac b/configure.ac
index 917420e35..9eb25bd02 100644
--- a/configure.ac
+++ b/configure.ac
@@ -633,6 +633,16 @@ AC_ISC_POSIX
 AC_SYS_LARGEFILE
 
 
+# GNU AWK requires -n option to interpret "0xHH" as a number
+if $AWK 'BEGIN { if (PROCINFO@<:@"version"@:>@) exit 1 }'; then
+  AWK_HEX_NUMBER_OPTION=''
+  AC_MSG_NOTICE([awk with no option for hexadecimal])
+else
+  AWK_HEX_NUMBER_OPTION='-n'
+  AC_MSG_NOTICE([awk with an option -n for hexadecimal])
+fi
+AC_SUBST(AWK_HEX_NUMBER_OPTION)
+
 # We need to compile and run a program on the build machine.  A
 # comment in libgpg-error says that the AC_PROG_CC_FOR_BUILD macro in
 # the AC archive is broken for autoconf 2.57.  Given that there is no
diff --git a/regexp/Makefile.am b/regexp/Makefile.am
index a4e9fce0b..307aacf9d 100644
--- a/regexp/Makefile.am
+++ b/regexp/Makefile.am
@@ -34,5 +34,6 @@ BUILT_SOURCES = _unicode_mapping.c
 MAINTAINERCLEANFILES = _unicode_mapping.c
 
 _unicode_mapping.c: parse-unidata.awk UnicodeData.txt
-	$(AWK) -f $(srcdir)/parse-unidata.awk $(srcdir)/UnicodeData.txt >$@
+	$(AWK) $(AWK_HEX_NUMBER_OPTION) -f $(srcdir)/parse-unidata.awk \
+	       $(srcdir)/UnicodeData.txt >$@
 endif
diff --git a/regexp/parse-unidata.awk b/regexp/parse-unidata.awk
index 8bdc0bf8b..b9ed351ee 100644
--- a/regexp/parse-unidata.awk
+++ b/regexp/parse-unidata.awk
@@ -32,17 +32,17 @@ BEGIN {
 }
 
 {
-    code = strtonum(("0x" $1))
+    code = int("0x" $1)
     name = $2
     class = $3
     upper = $13
     lower = $14
     title = $15
 
-    if (code <= 0x7f) {
+    if (code <= 127) {
 	next
     }
-    if (code > 0xffff) {
+    if (code > 65535) {
 	next
     }
     if ($3 !~ /^L.*/) {