TRE: Check for 32-bit wchar_t at configure time

This check used to be in src/elinks.h.  Move it to configure.in so
that (1) the result can be logged and (2) ELinks won't even link with
TRE if wchar_t prevents its use.

Also, rename HAVE_TRE_REGEX_H to CONFIG_TRE, to reflect that it is not
always defined if the header exists.

configure.in

@@ -925,11 +925,20 @@ fi
 # ===================================================================
 # Check for TRE library
 # ===================================================================
+#
+# This section only checks that --without-tre is not given and the
+# library seems to work, and sets TRE_CFLAGS, TRE_LIBS, and
+# tre_log.  It does not define CONFIG_TRE, and always resets
+# LIBS and CFLAGS back to their original values.
+#
+# After any --enable-utf-8 and --disable-utf-8 options have been
+# handled, a separate section decides whether to actually use TRE.
 
 AC_ARG_WITH([[tre]], [[  --without-tre           disable TRE regex search support]])
 if test "$with_tre" = no; then
 	AC_MSG_CHECKING([[for TRE]])
 	AC_MSG_RESULT([[disabled]])
+	tre_log="no (explicitly disabled)"
 else
 	AC_MSG_CHECKING([[for TRE in pkg-config]])
 	if pkg-config tre; then
@@ -955,22 +964,12 @@ else
 		     regwcomp(&re, L"zap", REG_ICASE);
 		     regwexec(&re, L"ELIZAPROGRAM", 1, match, 0);],
 		    [AC_MSG_RESULT([[yes]])
-		     AC_DEFINE([HAVE_TRE_REGEX_H], [1],
+		     tre_log="available"],
-			       [Define to 1 if you have the <tre/regex.h> header file.])
-		     # TRE_CFLAGS will be used only where needed.
-		     # TRE_LIBS will be kept in LIBS and used everywhere.
-		     EL_RESTORE_FLAGS
-		     LIBS="$TRE_LIBS $LIBS"],
 		    [AC_MSG_RESULT([[no]])
-		     TRE_CFLAGS=
+		     tre_log="no (TRE not found)"])
-		     TRE_LIBS=
+	EL_RESTORE_FLAGS
-		     EL_RESTORE_FLAGS])
-	AC_SUBST(TRE_CFLAGS)
-	AC_SUBST(TRE_LIBS)
 fi
 
-AC_CHECK_SIZEOF([wchar_t], [4], [[#include <wchar.h>]])
-
 # ===================================================================
 # Check for Ruby, optional even if installed.
 # ===================================================================
@@ -1385,6 +1384,66 @@ AC_ARG_ENABLE(weehoofooboomookerchoo,
 	      [AC_MSG_ERROR(Are you strange, or what?)])
 
 
+# ===================================================================
+# Decide whether to use TRE
+# ===================================================================
+#
+# This must be done after the CONFIG_UTF8 check above.
+# The first part of the TRE check is separate, to get
+# the configure --help output in a sensible order.
+
+if test "$tre_log" = "available"; then
+	if test "$CONFIG_UTF8" = "yes"; then
+		# When CONFIG_UTF8 and CONFIG_TRE are both defined,
+		# src/viewer/text/search.c makes a string of
+		# unicode_val_T and gives it to regwexec(), which
+		# expects a string of wchar_t.  If the unicode_val_T
+		# and wchar_t types are too different, this won't
+		# work, so try to detect that and disable regexp
+		# operations entirely in that case.
+		#
+		# Currently, this code only compares the sizes of the
+		# types.  src/intl/charsets.h defines unicode_val_T as
+		# uint32_t, so we check whether wchar_t has exactly 32
+		# bits.  But don't use AC_CHECK_SIZEOF for this, because
+		# there doesn't seem to be a documented way to get the
+		# result of that for use in the configure script.
+		#
+		# C99 says the implementation can define
+		# __STDC_ISO_10646__ if wchar_t values match ISO 10646
+		# (or Unicode) numbers in all locales.  Do not check
+		# that macro here, because it is too restrictive: it
+		# should be enough for ELinks if the values match in
+		# the locales where ELinks is actually run.
+		AC_MSG_CHECKING([[whether wchar_t is exactly 32-bit]])
+		AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[
+#include <limits.h>
+#include <stddef.h>
+int dummy[(sizeof(wchar_t) * CHAR_BIT == 32) ? 1 : -1];]], [])],
+			[AC_MSG_RESULT([[yes]])
+			 tre_log="TRE"],
+			[AC_MSG_RESULT([[no]])
+			 tre_log="no (unsuitable wchar_t)"])
+	else
+		# If UTF-8 is not used, then wchar_t doesn't matter.
+		tre_log="TRE"
+	fi
+fi
+
+if test "$tre_log" = "TRE"; then
+	AC_DEFINE([CONFIG_TRE], [1],
+		[Define as 1 to use the TRE library for regular expression searching.  This requires the <tre/regex.h> header file.  If you define CONFIG_UTF8 too, then wchar_t must be exactly 32-bit so that it matches unicode_val_T.])
+	# TRE_CFLAGS will be used only where needed.
+	LIBS="$LIBS $TRE_LIBS"
+else
+	TRE_LIBS=
+	TRE_CFLAGS=
+fi
+AC_SUBST(TRE_CFLAGS)
+AC_SUBST(TRE_LIBS)
+EL_LOG_CONFIG([[CONFIG_TRE]], [[Regexp searching]], [[$tre_log]])
+
+
 # ===================================================================
 # Further LDFLAGS tweaks
 # ===================================================================

src/config/options.inc

@@ -407,7 +407,7 @@ static struct option_info config_options_info[] = {
 		N_("Whether the search should match the document text while "
 		"maintaining case sensitivity.")),
 
-#ifdef HAVE_TRE_REGEX_H
+#ifdef CONFIG_TRE
 	INIT_OPT_INT("document.browse.search", N_("Regular expressions"),
 		"regex", 0, 0, 2, 0,
 		N_("Enable searching with regular expressions:\n"

src/elinks.h

@@ -29,30 +29,6 @@
 #define DEBUG_MEMLEAK
 #endif
 
-
-/* When CONFIG_UTF8 is defined, src/viewer/text/search.c makes a string
- * of unicode_val_T and gives it to regwexec(), which expects a string
- * of wchar_t.  If the unicode_val_T and wchar_t types are too different,
- * this won't work, so try to detect that and disable regexp operations
- * entirely in that case.
- *
- * Currently, this code only compares the sizes of the types.  Because
- * unicode_val_T is defined as uint32_t and POSIX says bytes are 8-bit,
- * sizeof(unicode_val_T) is 4 and the following compares SIZEOF_WCHAR_T
- * to that.
- *
- * C99 says the implementation can define __STDC_ISO_10646__ if wchar_t
- * values match ISO 10646 (or Unicode) numbers in all locales.  Do not
- * check that macro here, because it is too restrictive: it should be
- * enough for ELinks if the values match in the locales where ELinks is
- * actually run.  */
-
-#ifdef CONFIG_UTF8
-#if SIZEOF_WCHAR_T != 4
-#undef HAVE_TRE_REGEX_H
-#endif
-#endif
-
 /* This maybe overrides some of the standard high-level functions, to ensure
  * the expected behaviour. These overrides are not system specific. */
 #include "osdep/stub.h"

src/intl/charsets.h

@@ -1,6 +1,7 @@
 #ifndef EL__INTL_CHARSETS_H
 #define EL__INTL_CHARSETS_H
 
+/* The TRE check in configure.in assumes unicode_val_T is uint32_t.  */
 typedef uint32_t unicode_val_T;
 
 /* U+0020 SPACE.  Normally the same as ' ' or L' ' but perhaps ELinks

src/viewer/text/search.c

@@ -18,7 +18,7 @@
 #include <sys/types.h> /* FreeBSD needs this before regex.h */
 #include <stdlib.h>
 #include <string.h>
-#ifdef HAVE_TRE_REGEX_H
+#ifdef CONFIG_TRE
 #include <tre/regex.h>
 #endif
 
@@ -270,7 +270,7 @@ get_range(struct document *document, int y, int height, int l,
 	return 0;
 }
 
-#ifdef HAVE_TRE_REGEX_H
+#ifdef CONFIG_TRE
 /** Returns a string @c doc that is a copy of the text in the search
  * nodes from @a s1 to (@a s1 + @a doclen - 1) with the space at the
  * end of each line converted to a new-line character (LF). */
@@ -465,7 +465,7 @@ is_in_range_regex(struct document *document, int y, int height,
 
 	return common_ctx.found;
 }
-#endif /* HAVE_TRE_REGEX_H */
+#endif /* CONFIG_TRE */
 
 static UCHAR *
 memacpy_u(unsigned char *text, int textlen, int utf8)
@@ -602,7 +602,7 @@ is_in_range(struct document *document, int y, int height,
 	if (get_range(document, y, height, textlen, &s1, &s2))
 		return 0;
 
-#ifdef HAVE_TRE_REGEX_H
+#ifdef CONFIG_TRE
 	if (get_opt_int("document.browse.search.regex"))
 		return is_in_range_regex(document, y, height, text, textlen,
 					 min, max, s1, s2, utf8);
@@ -681,7 +681,7 @@ srch_failed:
 	*pl = len;
 }
 
-#ifdef HAVE_TRE_REGEX_H
+#ifdef CONFIG_TRE
 struct get_searched_regex_context {
 	int xoffset;
 	int yoffset;
@@ -749,7 +749,7 @@ get_searched_regex(struct document_view *doc_view, struct point **pt, int *pl,
 	*pt = ctx.points;
 	*pl = ctx.len;
 }
-#endif /* HAVE_TRE_REGEX_H */
+#endif /* CONFIG_TRE */
 
 static void
 get_searched(struct document_view *doc_view, struct point **pt, int *pl, int utf8)
@@ -773,7 +773,7 @@ get_searched(struct document_view *doc_view, struct point **pt, int *pl, int utf
 		return;
 	}
 
-#ifdef HAVE_TRE_REGEX_H
+#ifdef CONFIG_TRE
 	if (get_opt_int("document.browse.search.regex"))
 		get_searched_regex(doc_view, pt, pl, l, s1, s2, utf8);
 	else
@@ -1594,7 +1594,7 @@ search_typeahead(struct session *ses, struct document_view *doc_view,
  * a nice cleanup target ;-). --pasky */
 
 enum search_option {
-#ifdef HAVE_TRE_REGEX_H
+#ifdef CONFIG_TRE
 	SEARCH_OPT_REGEX,
 #endif
 	SEARCH_OPT_CASE,
@@ -1602,7 +1602,7 @@ enum search_option {
 };
 
 static struct option_resolver resolvers[] = {
-#ifdef HAVE_TRE_REGEX_H
+#ifdef CONFIG_TRE
 	{ SEARCH_OPT_REGEX,	"regex" },
 #endif
 	{ SEARCH_OPT_CASE,	"case" },
@@ -1669,7 +1669,7 @@ search_dlg_do(struct terminal *term, struct memory_list *ml,
 			       hop->values, SEARCH_OPTIONS);
 	hop->data = data;
 
-#ifdef HAVE_TRE_REGEX_H
+#ifdef CONFIG_TRE
 #define SEARCH_WIDGETS_COUNT 8
 #else
 #define SEARCH_WIDGETS_COUNT 5
@@ -1693,7 +1693,7 @@ search_dlg_do(struct terminal *term, struct memory_list *ml,
 	field = get_dialog_offset(dlg, SEARCH_WIDGETS_COUNT);
 	add_dlg_field(dlg, text, 0, 0, NULL, MAX_STR_LEN, field, history);
 
-#ifdef HAVE_TRE_REGEX_H
+#ifdef CONFIG_TRE
 	add_dlg_radio(dlg, _("Normal search", term), 1, 0, &hop->values[SEARCH_OPT_REGEX].number);
 	add_dlg_radio(dlg, _("Regexp search", term), 1, 1, &hop->values[SEARCH_OPT_REGEX].number);
 	add_dlg_radio(dlg, _("Extended regexp search", term), 1, 2, &hop->values[SEARCH_OPT_REGEX].number);