gdb:

2009-03-19  Tom Tromey  <tromey@redhat.com>
	    Julian Brown  <julian@codesourcery.com>

	PR i18n/7220, PR i18n/7821, PR exp/8815, PR exp/9103,
	PR i18n/9401, PR exp/9613:
	* NEWS: Update
	* value.h (value_typed_string): Declare.
	(val_print_string): Update.
	* valprint.h (print_char_chars): Update.
	* valprint.c (print_char_chars): Add type argument.  Update.
	(val_print_string): Likewise.
	* valops.c (value_typed_string): New function.
	* utils.c (host_char_to_target): New function.
	(parse_escape): Use host_char_to_target, host_hex_value.  Update.
	Remove '^' case.
	(no_control_char_error): Remove.
	* typeprint.c (print_type_scalar): Update.
	* scm-valprint.c (scm_scmval_print): Update.
	* scm-lang.h (scm_printchar, scm_printstr): Update.
	* scm-lang.c (scm_printchar): Add type argument.
	(scm_printstr): Likewise.
	* printcmd.c (print_formatted): Update.
	(print_scalar_formatted): Update.
	(printf_command) <wide_string_arg, wide_char_arg>: New constants.
	Handle '%lc' and '%ls'.
	* parser-defs.h (struct typed_stoken): New type.
	(struct stoken_vector): Likewise.
	(write_exp_string_vector): Declare.
	* parse.c (write_exp_string_vector): New function.
	* p-valprint.c (pascal_val_print): Update.
	* p-lang.h (is_pascal_string_type, pascal_printchar,
	pascal_printstr): Update.
	* p-lang.c (is_pascal_string_type): Remove 'char_size' argument.
	Add 'char_type' argument.
	(pascal_emit_char): Add type argument.
	(pascal_printchar): Likewise.
	(pascal_printstr): Likewise.
	* objc-lang.c (objc_emit_char): Add type argument.
	(objc_printchar): Likewise.
	(objc_printstr): Likewise.
	* macroexp.c (get_character_constant): Handle unicode characters.
	Use c_parse_escape.
	(get_string_literal): Handle unicode strings.  Use
	c_parse_escape.
	* m2-valprint.c (print_unpacked_pointer): Update.
	(m2_print_array_contents): Update.
	(m2_val_print): Update.
	* m2-lang.c (m2_emit_char): Add type argument.
	(m2_printchar): Likewise.
	(m2_printstr): Likewise.
	* language.h (struct language_defn) <la_printchar>: Add type
	argument.
	<la_printstr, la_emitchar>: Likewise.
	(LA_PRINT_CHAR): Likewise.
	(LA_PRINT_STRING): Likewise.
	(LA_EMIT_CHAR): Likewise.
	* language.c (unk_lang_emit_char): Add type argument.
	(unk_lang_printchar): Likewise.
	(unk_lang_printstr): Likewise.
	* jv-valprint.c (java_val_print): Update.
	* jv-lang.c (java_emit_char): Add type argument.
	* f-valprint.c (f_val_print): Update.
	* f-lang.c (f_emit_char): Add type argument.
	(f_printchar): Likewise.
	(f_printstr): Likewise.
	* expprint.c (print_subexp_standard): Update.
	* charset.h (target_wide_charset): Declare.
	(c_target_char_has_backslash_escape, c_parse_backslash,
	host_char_print_literally, host_char_to_target,
	target_char_to_host, target_char_to_control_char): Remove.
	(enum transliterations): New type.
	(convert_between_encodings): Declare.
	(HOST_ESCAPE_CHAR): New define.
	(host_letter_to_control_character, host_hex_value): Declare.
	(enum wchar_iterate_result): New enum.
	(struct wchar_iterator): Declare.
	(make_wchar_iterator, make_cleanup_wchar_iterator, wchar_iterator,
	wchar_push_back): Declare.
	* charset-list.h: New file.
	* c-valprint.c (textual_name): New function.
	(textual_element_type): Handle wide character types.
	(c_val_print): Pass original type to textual_element_type.  Handle
	wide character types.
	(c_value_print): Use textual_element_type.  Pass original type of
	value to val_print.
	* c-lang.h (enum c_string_type): New type.
	(c_printchar, c_printstr): Update.
	* c-lang.c (classify_type): New function.
	(print_wchar): Likewise.
	(c_emit_char): Add type argument.  Handle wide characters.
	(c_printchar): Likewise.
	(c_printstr): Add type argument.  Handle wide and multibyte
	character sets.
	(convert_ucn): New function.
	(emit_numeric_character): Likewise.
	(convert_octal): Likewise.
	(convert_hex): Likewise.
	(ADVANCE): New macro.
	(convert_escape): New function.
	(parse_one_string): Likewise.
	(evaluate_subexp_c): Likewise.
	(exp_descriptor_c): New global.
	(c_language_defn): Use exp_descriptor_c.
	(cplus_language_defn): Likewise.
	(asm_language_defn): Likewise.
	(minimal_language_defn): Likewise.
	(charset_for_string_type): New function.
	* c-exp.y (%union): Add 'svec' and 'tsval'.
	(CHAR): New token.
	(exp): Add CHAR production.
	(string_exp): Rewrite.
	(exp) <string_exp>: Rewrite.
	(tempbuf): Now global.
	(tempbuf_init): New global.
	(parse_string_or_char): New function.
	(yylex) <tempbuf>: Now global.
	<tokptr, tempbufindex, tempbufsize, token_string, class_prefix>:
	Remove.
	Handle 'u', 'U', and 'L' prefixes.  Call parse_string_or_char.
	(c_parse_escape): New function.
	* auxv.c (fprint_target_auxv): Update.
	* ada-valprint.c (ada_emit_char): Add type argument.
	(ada_printchar): Likewise.
	(ada_print_scalar): Update.
	(printstr): Add type argument.  Update calls to ada_emit_char.
	(ada_printstr): Add type argument.
	(ada_val_print_array): Update.
	(ada_val_print_1): Likewise.
	* ada-lang.c (emit_char): Add type argument.
	* ada-lang.h (ada_emit_char, ada_printchar, ada_printstr): Add
	type arguments.
	* gdb_locale.h: Include langinfo.h.
	* charset.c (_initialize_charset): Set default host charset from
	the locale.  Don't register charsets.  Add target-wide-charset
	commands.  Call find_charset_names.
	(struct charset, struct translation): Remove.
	(GDB_DEFAULT_HOST_CHARSET): Remove.
	(GDB_DEFAULT_TARGET_WIDE_CHARSET): New define.
	(target_wide_charset_name): New global.
	(show_host_charset_name): Handle "auto".
	(show_target_wide_charset_name): New function.
	(host_charset_enum, target_charset_enum): Remove.
	(charset_enum): New global.
	(all_charsets, register_charset, lookup_charset, all_translations,
	register_translation, lookup_translation): Remove.
	(simple_charset, ascii_print_literally, ascii_to_control): Remove.
	(iso_8859_print_literally, iso_8859_to_control,
	iso_8859_family_charset): Remove.
	(ebcdic_print_literally, ebcdic_to_control,
	ebcdic_family_charset): Remove.
	(struct cached_iconv, check_iconv_cache, cached_iconv_convert,
	register_iconv_charsets): Remove.
	(target_wide_charset_be_name, target_wide_charset_le_name): New
	globals.
	(identity_either_char_to_other): Remove.
	(set_be_le_names, validate): New functions.
	(backslashable, backslashed, represented): Remove.
	(default_c_target_char_has_backslash_escape): Remove.
	(default_c_parse_backslash, iconv_convert): Remove.
	(ascii_to_iso_8859_1_table, ascii_to_ebcdic_us_table,
	ascii_to_ibm1047_table, iso_8859_1_to_ascii_table,
	iso_8859_1_to_ebcdic_us_table, iso_8859_1_to_ibm1047_table,
	ebcdic_us_to_ascii_table, ebcdic_us_to_iso_8859_1_table,
	ebcdic_us_to_ibm1047_table, ibm1047_to_ascii_table,
	ibm1047_to_iso_8859_1_table, ibm1047_to_ebcdic_us_table): Remove.
	(table_convert_char, table_translation, simple_table_translation):
	Remove.
	(current_host_charset, current_target_charset,
	c_target_char_has_backslash_escape_func,
	c_target_char_has_backslash_escape_baton): Remove.
	(c_parse_backslash_func, c_parse_backslash_baton): Remove.
	(host_char_to_target_func, host_char_to_target_baton): Remove.
	(target_char_to_host_func, target_char_to_host_baton): Remove.
	(cached_iconv_host_to_target, cached_iconv_target_to_host):
	Remove.
	(lookup_charset_or_error, check_valid_host_charset): Remove.
	(set_host_and_target_charsets): Remove.
	(set_host_charset, set_target_charset): Remove.
	(set_host_charset_sfunc, set_target_charset_sfunc): Rewrite.
	(set_target_wide_charset_sfunc): New function.
	(show_charset): Print target wide character set.
	(host_charset, target_charset): Rewrite.
	(target_wide_charset): New function.
	(c_target_char_has_backslash_escape): Remove.
	(c_parse_backslash): Remove.
	(host_letter_to_control_character): New function.
	(host_char_print_literally): Remove.
	(host_hex_value): New function.
	(target_char_to_control_char): Remove.
	(cleanup_iconv): New function.
	(convert_between_encodings): New function.
	(target_char_to_host): Remove.
	(struct wchar_iterator): Define.
	(make_wchar_iterator, make_cleanup_wchar_iterator, wchar_iterator,
	wchar_push_back): New functions.
	(do_cleanup_iterator): New function.
	(char_ptr): New typedef.
	(charsets): New global.
	(add_one, find_charset_names): New functions.
	(default_charset_names): New global.
	(auto_host_charset_name): Likewise.
	* aclocal.m4, config.in, configure: Rebuild.
	* configure.ac: Call AM_LANGINFO_CODESET.
	(GDB_DEFAULT_HOST_CHARSET): Default to UTF-8.
	(AM_ICONV): Invoke earlier.
	* acinclude.m4: Include codeset.m4.  Subst LIBICONV_INCLUDE and
	LIBICONV_LIBDIR.  Check for libiconv in build tree.
	* Makefile.in (LIBICONV_LIBDIR, LIBICONV_INCLUDE): New macros.
	(INTERNAL_CFLAGS_BASE): Add LIBICONV_INCLUDE.
	(INTERNAL_LDFLAGS): Add LIBICONV_LIBDIR.
	* gdb_obstack.h (obstack_grow_wstr): New define.
        * gdb_wchar.h: New file.
        * defs.h: Include it.
gdb/testsuite:
	* gdb.base/store.exp: Update for change to escape output.
	* gdb.base/callfuncs.exp (fetch_all_registers): Update for change
	to escape output.
	* gdb.base/pointers.exp: Update for change to escape output.
	* gdb.base/long_long.exp (gdb_test_long_long): Update for change
	to escape output.
	* gdb.base/constvars.exp (do_constvar_tests): Update for change to
	escape output.
	* gdb.base/call-rt-st.exp (print_struct_call): Update for change
	to escape output.
	* gdb.cp/ref-types.exp (gdb_start_again): Update for change to
	escape output.
	* gdb.base/setvar.exp: Update for change to escape output.
	* lib/gdb.exp (default_gdb_start): Set LC_CTYPE to C.
	* gdb.base/printcmds.exp (test_print_all_chars): Update for change
	to escape output.
	(test_print_string_constants): Likewise.
	* gdb.base/charset.exp (valid_host_charset): Check size of
	wchar_t.  Handle UCS-2 and UCS-4.  Add tests for wide and unicode
	cases.  Handle "auto"-related output.
	* gdb.base/charset.c (char16_t, char32_t): New typedefs.
	(uvar, Uvar): New globals.
gdb/doc:
	* gdb.texinfo (Character Sets): Remove obsolete text.  Document
	set target-wide-charset.
	(Requirements): Mention iconv.

1 files changed, 130 insertions, 84 deletions

diff --git a/gdb/testsuite/gdb.base/charset.exp b/gdb/testsuite/gdb.base/charset.exp
index fa26521..a6368c9 100644
--- a/gdb/testsuite/gdb.base/charset.exp
+++ b/gdb/testsuite/gdb.base/charset.exp
@@ -47,13 +47,7 @@ proc parse_show_charset_output {testname} {
     global gdb_prompt
 
     gdb_expect {
-        -re "The current host and target character set is `(.*)'\\.\[\r\n\]+$gdb_prompt $" {
-            set host_charset $expect_out(1,string)
-            set target_charset $expect_out(1,string)
-	    set retlist [list $host_charset $target_charset]
-            pass $testname
-        }
-        -re "The current host character set is `(.*)'\\.\[\r\n\]+The current target character set is `(.*)'\\.\[\r\n\]+$gdb_prompt $" {
+        -re "The host character set is \"(.*)\"\\.\[\r\n\]+The target character set is \"(.*)\"\\.\[\r\n\]+The target wide character set is \"(.*)\"\\.\[\r\n\]+$gdb_prompt $" {
             set host_charset $expect_out(1,string)
             set target_charset $expect_out(2,string)
 	    set retlist [list $host_charset $target_charset]
@@ -81,76 +75,31 @@ proc parse_show_charset_output {testname} {
 }
 
 
-# Try the various `show charset' commands.  These are all aliases of each
-# other; `show target-charset' and `show host-charset' actually print
-# both the host and target charsets.
+# Try the various `show charset' commands.
 
 send_gdb "show charset\n"
 set show_charset [parse_show_charset_output "show charset"]
 
 send_gdb "show target-charset\n"
-set show_target_charset [parse_show_charset_output "show target-charset"]
+set show_target_charset \
+  [lindex [parse_show_charset_output "show target-charset"] 0]
 
-if {[lsearch $show_charset $show_target_charset] >= 0} {
+if {[lsearch -exact $show_charset $show_target_charset] >= 0} {
     pass "check `show target-charset' against `show charset'"
 } else {
     fail "check `show target-charset' against `show charset'"
 }
 
 send_gdb "show host-charset\n"
-set show_host_charset [parse_show_charset_output "show host-charset"]
+set show_host_charset \
+  [lindex [parse_show_charset_output "show host-charset"] 0]
 
-if {[lsearch $show_charset $show_host_charset] >= 0} {
+if {[lsearch -exact $show_charset $show_host_charset] >= 0} {
     pass "check `show host-charset' against `show charset'"
 } else {
     fail "check `show host-charset' against `show charset'"
 }
 
-
-# Get the list of supported (host) charsets as possible completions.
-send_gdb "set charset \t\t"
-
-# Check that we can at least use ASCII as a host character set.
-sleep 1
-gdb_expect {
-    -re "^set charset .*\r\nASCII.*\r\n$gdb_prompt set charset " {
-	# We got the output that we wanted, including ASCII as possible
-	# charset. Send a newline to get us back to the prompt. This will
-	# also generate an error message. Let's not check here that the error
-	# message makes sense, we do that below, as a separate testcase.
-        send_gdb "\n"
-        gdb_expect {
-	    -re ".*Requires an argument.*$gdb_prompt $" {
-		pass "get valid character sets"
-	    }
-	    -re ".*$gdb_prompt $" {
-		send_gdb "\n"
-		gdb_expect {
-		    -re ".*$gdb_prompt $" {
-			fail "get valid character sets"
-		    }
-		}
-	    }
-	    timeout {
-		fail "(timeout) get valid character sets"
-	    }
-	}
-    }
-    -re ".*$gdb_prompt $" {
-	# We got some output that ended with a regular prompt
-        fail "get valid character sets"
-    }
-    -re ".*$gdb_prompt set charset.*$" {
-	# We got some other output, send a cntrl-c to gdb to get us back
-        # to the prompt.
-	send_gdb "\003"
-        fail "get valid character sets"
-    }
-    timeout {
-        fail "get valid character sets (timeout)"
-    }
-}
-
 # Try a malformed `set charset'.
 gdb_test "set charset" \
          "Requires an argument. Valid arguments are.*" \
@@ -179,7 +128,7 @@ proc all_charset_names {} {
 
 proc valid_host_charset {charset} {
     global charsets
-    return $charsets($charset)
+    return [expr {[info exists charsets($charset)] && $charsets($charset)}]
 }
 
 send_gdb "set host-charset\n"
@@ -211,6 +160,13 @@ gdb_expect {
     }
 }
 
+# If gdb was built with a phony iconv, it will only have two character
+# sets: "auto" and the default.  In this situation, this set of tests
+# is pointless.
+if {[llength [array names charsets]] < 3} {
+    untested charset.exp
+    return -1
+}
 
 send_gdb "set target-charset\n"
 gdb_expect {
@@ -244,8 +200,10 @@ gdb_expect {
     }
 }
 
-# Make sure that GDB supports every host/target charset combination.
-foreach host_charset [all_charset_names] {
+# We don't want to test all the charset names here, since that would
+# be too many combinations.  We we pick a subset.
+set charset_subset {ASCII ISO-8859-1 EBCDIC-US IBM1047}
+foreach host_charset $charset_subset {
     if {[valid_host_charset $host_charset]} {
 
         set testname "try `set host-charset $host_charset'"
@@ -279,7 +237,7 @@ foreach host_charset [all_charset_names] {
 
         # Now try setting every possible target character set,
         # given that host charset.
-        foreach target_charset [all_charset_names] {
+        foreach target_charset $charset_subset {
             set testname "try `set target-charset $target_charset'"
             send_gdb "set target-charset $target_charset\n"
             gdb_expect {
@@ -404,23 +362,42 @@ gdb_expect {
 }
 
 
+# We only try the wide character tests on machines where the wchar_t
+# typedef in the test case has the right size.
+set wchar_size [get_sizeof wchar_t 99]
+set wchar_ok 0
+if {$wchar_size == 2} {
+    lappend charset_subset UCS-2
+    set wchar_ok 1
+} elseif {$wchar_size == 4} {
+    lappend charset_subset UCS-4
+    set wchar_ok 1
+}
+
 gdb_test "set host-charset ASCII" ""
-foreach target_charset [all_charset_names] {
-    send_gdb "set target-charset $target_charset\n" 
+foreach target_charset $charset_subset {
+    if {$target_charset == "UCS-4" || $target_charset == "UCS-2"} {
+	set param target-wide-charset
+	set L L
+    } else {
+	set param target-charset
+	set L ""
+    }
+    send_gdb "set $param $target_charset\n" 
     gdb_expect {
         -re "$gdb_prompt $" {
-            pass "set target-charset $target_charset"
+            pass "set $param $target_charset"
         }
         timeout {
-            fail "set target-charset $target_charset (timeout)"
+            fail "set $param $target_charset (timeout)"
         }
     }
 
     # Try printing the null character.  There seems to be a bug in
     # gdb_test that requires us to use gdb_expect here.
-    send_gdb "print '\\0'\n"
+    send_gdb "print $L'\\0'\n"
     gdb_expect {
-        -re "\\\$${decimal} = 0 '\\\\0'\[\r\n\]+$gdb_prompt $" {
+        -re "\\\$${decimal} = 0 $L'\\\\0'\[\r\n\]+$gdb_prompt $" {
             pass "print the null character in ${target_charset}"
         }
         -re "$gdb_prompt $" {
@@ -435,8 +412,14 @@ foreach target_charset [all_charset_names] {
     # a string in $target_charset.  The variable's name is the
     # character set's name, in lower-case, with all non-identifier
     # characters replaced with '_', with "_string" stuck on the end.
-    set var_name [string tolower "${target_charset}_string"]
-    regsub -all -- "\[^a-z0-9_\]" $var_name "_" var_name
+    if {$target_charset == "UCS-2"} {
+	# We still use the ucs_4_string variable -- but the size is
+	# correct for UCS-2.
+	set var_name ucs_4_string
+    } else {
+	set var_name [string tolower "${target_charset}_string"]
+	regsub -all -- "\[^a-z0-9_\]" $var_name "_" var_name
+    }
     
     # Compute a regexp matching the results we expect.  This is static,
     # but it's easier than writing it out.
@@ -444,12 +427,12 @@ foreach target_charset [all_charset_names] {
     set uppercase "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
     set lowercase "abcdefghijklmnopqrstuvwxyz"
     set digits "0123456789"
-    set octal_escape "\\\\\[0-9\]\[0-9\]\[0-9\]"
+    set octal_escape "\\\\\[0-9\]+"
 
     send_gdb "print $var_name\n"
     # ${escapes}${uppercase}${lowercase}${digits}${octal}${octal}
     gdb_expect {
-        -re ".* = \"(\\\\a|x)(\\\\b|x)(\\\\f|x)(\\\\n|x)(\\\\r|x)(\\\\t|x)(\\\\v|x)${uppercase}${lowercase}${digits}(\\\\\[0-9\]\[0-9\]\[0-9\]|x)(\\\\\[0-9\]\[0-9\]\[0-9\]|x).*\"\[\r\n\]+$gdb_prompt $" {
+        -re ".* = $L\"(\\\\a|x)(\\\\b|x)(\\\\f|x)(\\\\n|x)(\\\\r|x)(\\\\t|x)(\\\\v|x)${uppercase}${lowercase}${digits}(${octal_escape}|x)+\"\[\r\n\]+$gdb_prompt $" {
             pass "print string in $target_charset"
         }
         -re "$gdb_prompt $" {
@@ -461,22 +444,22 @@ foreach target_charset [all_charset_names] {
     }
 
     # Try entering a character literal, and see if it comes back unchanged.
-    gdb_test "print 'A'" \
-             " = \[0-9-\]+ 'A'" \
+    gdb_test "print $L'A'" \
+             " = \[0-9-\]+ $L'A'" \
              "parse character literal in ${target_charset}"
 
     # Check that the character literal was encoded correctly.
-    gdb_test "print 'A' == $var_name\[7\]" \
+    gdb_test "print $L'A' == $var_name\[7\]" \
              " = 1" \
              "check value of parsed character literal in ${target_charset}"
 
     # Try entering a string literal, and see if it comes back unchanged.
-    gdb_test "print \"abcdefABCDEF012345\"" \
-             " = \"abcdefABCDEF012345\"" \
+    gdb_test "print $L\"abcdefABCDEF012345\"" \
+             " = $L\"abcdefABCDEF012345\"" \
              "parse string literal in ${target_charset}"
 
     # Check that the string literal was encoded correctly.
-    gdb_test "print \"q\"\[0\] == $var_name\[49\]" \
+    gdb_test "print $L\"q\"\[0\] == $var_name\[49\]" \
              " = 1" \
              "check value of parsed string literal in ${target_charset}"
 
@@ -509,7 +492,7 @@ foreach target_charset [all_charset_names] {
         send_gdb "print $var_name\[$i\]\n"
         set have_escape 1
         gdb_expect {
-            -re "= \[0-9-\]+ '\\\\${escape}'\[\r\n\]+$gdb_prompt $" {
+            -re "= \[0-9-\]+ $L'\\\\${escape}'\[\r\n\]+$gdb_prompt $" {
                 pass "try printing '\\${escape}' in ${target_charset}"
             }
             -re "= \[0-9-\]+ 'x'\[\r\n\]+$gdb_prompt $" {
@@ -527,12 +510,12 @@ foreach target_charset [all_charset_names] {
         if {$have_escape} {
 
             # Try parsing a backslash escape in a character literal.
-            gdb_test "print '\\${escape}' == $var_name\[$i\]" \
+            gdb_test "print $L'\\${escape}' == $var_name\[$i\]" \
                      " = 1" \
                      "check value of '\\${escape}' in ${target_charset}"
 
             # Try parsing a backslash escape in a string literal.
-            gdb_test "print \"\\${escape}\"\[0\] == $var_name\[$i\]" \
+            gdb_test "print $L\"\\${escape}\"\[0\] == $var_name\[$i\]" \
                      " = 1" \
                      "check value of \"\\${escape}\" in ${target_charset}"
         }
@@ -540,10 +523,73 @@ foreach target_charset [all_charset_names] {
 
     # Try printing a character escape that doesn't exist.  We should 
     # get the unescaped character, in the target character set.
-    gdb_test "print '\\q'" " = \[0-9-\]+ 'q'" \
+    gdb_test "print $L'\\q'" " = \[0-9-\]+ $L'q'" \
              "print escape that doesn't exist in $target_charset"
-    gdb_test "print '\\q' == $var_name\[49\]" " = 1" \
+    gdb_test "print $L'\\q' == $var_name\[49\]" " = 1" \
              "check value of escape that doesn't exist in $target_charset"
 }
 
+# Reset the target charset.
+gdb_test "set target-charset UTF-8" ""
+
+# \242 is not a valid UTF-8 character.
+gdb_test "print \"\\242\"" " = \"\\\\242\"" \
+  "non-representable target character"
+
+gdb_test "print '\\x'" "\\\\x escape without a following hex digit."
+gdb_test "print '\\u'" "\\\\u escape without a following hex digit."
+gdb_test "print '\\9'" " = \[0-9\]+ '9'"
+
+# Tests for wide- or unicode- strings.  L is the prefix letter to use,
+# either "L" (for wide strings), "u" (for UCS-2), or "U" (for UCS-4).
+# NAME is used in the test names and should be related to the prefix
+# letter in some easy-to-undestand way.
+proc test_wide_or_unicode {L name} {
+    gdb_test "print $L\"ab\" $L\"c\"" " = $L\"abc\"" \
+      "basic $name string concatenation"
+    gdb_test "print $L\"ab\" \"c\"" " = $L\"abc\"" \
+      "narrow and $name string concatenation"
+    gdb_test "print \"ab\" $L\"c\"" " = $L\"abc\"" \
+      "$name and narrow string concatenation"
+    gdb_test "print $L\"\\xe\" $L\"c\"" " = $L\"\\\\16c\"" \
+      "$name string concatenation with escape"
+    gdb_test "print $L\"\" \"abcdef\" \"g\"" \
+      "$L\"abcdefg\"" \
+      "concatenate three strings with empty $name string"
+
+    gdb_test "print $L'a'" "= \[0-9\]+ $L'a'" \
+      "basic $name character"
+}
+
+if {$wchar_ok} {
+    test_wide_or_unicode L wide
+}
+
+set ucs2_ok [expr {[get_sizeof char16_t 99] == 2}]
+if {$ucs2_ok} {
+    test_wide_or_unicode u UCS-2
+}
+
+set ucs4_ok [expr {[get_sizeof char32_t 99] == 4}]
+if {$ucs4_ok} {
+    test_wide_or_unicode U UCS-4
+}
+
+# Test an invalid string combination.
+proc test_combination {L1 name1 L2 name2} {
+    gdb_test "print $L1\"abc\" $L2\"def\"" \
+      "Undefined string concatenation." \
+      "undefined concatenation of $name1 and $name2"
+}
+
+if {$wchar_ok && $ucs2_ok} {
+    test_combination L wide u UCS-2
+}
+if {$wchar_ok && $ucs4_ok} {
+    test_combination L wide U UCS-4
+}
+if {$ucs2_ok && $ucs4_ok} {
+    test_combination u UCS-2 U UCS-4
+}
+
 gdb_exit