Add support for UTF-8 in EBCDIC environments.

git-svn-id: svn://vcs.exim.org/pcre/code/trunk@391 2f5784b3-3f2a-0410-8824-cb99058d5e15

1 files changed, 18 insertions, 4 deletions

diff --git a/configure.ac b/configure.ac
index f5dabf0..7b7c46f 100644
--- a/configure.ac
+++ b/configure.ac
@@ -102,7 +102,7 @@ AC_ARG_ENABLE(rebuild-chartables,
 # Handle --enable-utf8 (disabled by default)
 AC_ARG_ENABLE(utf8,
               AS_HELP_STRING([--enable-utf8],
-                             [enable UTF-8 support]),
+                             [enable UTF-8 support (incompatible with --enable-ebcdic)]),
               , enable_utf8=unset)
 
 # Handle --enable-unicode-properties
@@ -150,7 +150,7 @@ AC_ARG_ENABLE(bsr-anycrlf,
 # Handle --enable-ebcdic
 AC_ARG_ENABLE(ebcdic,
               AS_HELP_STRING([--enable-ebcdic],
-                             [assume EBCDIC coding rather than ASCII; use this only in (uncommon) EBCDIC environments; it implies --enable-rebuild-chartables]),
+                             [assume EBCDIC coding rather than ASCII; incompatible with --enable-utf8; use only in (uncommon) EBCDIC environments; it implies --enable-rebuild-chartables]),
               , enable_ebcdic=no)
 
 # Handle --disable-stack-for-recursion
@@ -226,10 +226,17 @@ then
 fi
 
 # Make sure that if enable_ebcdic is set, rebuild_chartables is also enabled.
+# Also check that UTF-8 support is not requested, because PCRE cannot handle
+# EBCDIC and UTF-8 in the same build. To do so it would need to use different
+# character constants depending on the mode.
 #
 if test "x$enable_ebcdic" = "xyes"
 then
   enable_rebuild_chartables=yes
+  if test "x$enable_utf8" = "xyes"
+  then
+    AC_MSG_ERROR([support for EBCDIC and UTF-8 cannot be enabled at the same time])
+  fi
 fi
 
 # Convert the newline identifier into the appropriate integer value.
@@ -381,7 +388,10 @@ fi
 
 if test "$enable_utf8" = "yes"; then
   AC_DEFINE([SUPPORT_UTF8], [], [
-    Define to enable support for the UTF-8 Unicode encoding.])
+    Define to enable support for the UTF-8 Unicode encoding. This will
+    work even in an EBCDIC environment, but it is incompatible with
+    the EBCDIC macro. That is, PCRE can support *either* EBCDIC code
+    *or* ASCII/UTF-8, but not both at once.])
 fi
 
 if test "$enable_unicode_properties" = "yes"; then
@@ -502,7 +512,11 @@ if test "$enable_ebcdic" = "yes"; then
   AC_DEFINE_UNQUOTED([EBCDIC], [], [
     If you are compiling for a system that uses EBCDIC instead of ASCII
     character codes, define this macro as 1. On systems that can use
-    "configure", this can be done via --enable-ebcdic.])
+    "configure", this can be done via --enable-ebcdic. PCRE will then
+    assume that all input strings are in EBCDIC. If you do not define 
+    this macro, PCRE will assume input strings are ASCII or UTF-8 Unicode. 
+    It is not possible to build a version of PCRE that supports both
+    EBCDIC and UTF-8.])
 fi
 
 # Platform specific issues