1 # This testcase is part of GDB, the GNU debugger.
3 # Copyright 2001, 2004, 2007, 2008, 2009, 2010, 2011
4 # Free Software Foundation, Inc.
6 # This program is free software; you can redistribute it and/or modify
7 # it under the terms of the GNU General Public License as published by
8 # the Free Software Foundation; either version 3 of the License, or
9 # (at your option) any later version.
11 # This program is distributed in the hope that it will be useful,
12 # but WITHOUT ANY WARRANTY; without even the implied warranty of
13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 # GNU General Public License for more details.
16 # You should have received a copy of the GNU General Public License
17 # along with this program. If not, see <http://www.gnu.org/licenses/>.
19 # Please email any bugs, comments, and/or additions to this file to:
22 # Test GDB's character set support.
29 set testfile "charset"
30 set srcfile ${testfile}.c
31 set srcmallocfile ${testfile}-malloc.c
32 if { [prepare_for_testing ${testfile}.exp ${testfile} [list $srcfile $srcmallocfile]] } {
36 # Parse the output from a `show charset' command. Return the host
37 # and target charset as a two-element list.
38 proc parse_show_charset_output {testname} {
42 -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 $" {
43 set host_charset $expect_out(1,string)
44 set target_charset $expect_out(2,string)
45 set retlist [list $host_charset $target_charset]
48 -re "The host character set is \"(.*)\"\\.\[\r\n\]+$gdb_prompt $" {
49 set host_charset $expect_out(1,string)
50 set retlist [list $host_charset]
53 -re "The target character set is \"(.*)\"\\.\[\r\n\]+$gdb_prompt $" {
54 set target_charset $expect_out(1,string)
55 set retlist [list $target_charset]
58 -re ".*$gdb_prompt $" {
62 fail "$testname (timeout)"
70 # Try the various `show charset' commands.
72 send_gdb "show charset\n"
73 set show_charset [parse_show_charset_output "show charset"]
75 send_gdb "show target-charset\n"
76 set show_target_charset \
77 [lindex [parse_show_charset_output "show target-charset"] 0]
79 if {[lsearch -exact $show_charset $show_target_charset] >= 0} {
80 pass "check `show target-charset' against `show charset'"
82 fail "check `show target-charset' against `show charset'"
85 send_gdb "show host-charset\n"
86 set show_host_charset \
87 [lindex [parse_show_charset_output "show host-charset"] 0]
89 if {[lsearch -exact $show_charset $show_host_charset] >= 0} {
90 pass "check `show host-charset' against `show charset'"
92 fail "check `show host-charset' against `show charset'"
95 # Try a malformed `set charset'.
96 gdb_test "set charset" \
97 "Requires an argument. Valid arguments are.*" \
98 "try malformed `set charset'"
100 # Try using `set host-charset' on an invalid character set.
101 gdb_test "set host-charset my_grandma_bonnie" \
102 "Undefined item: \"my_grandma_bonnie\"." \
103 "try `set host-charset' with invalid charset"
105 # Try using `set target-charset' on an invalid character set.
106 gdb_test "set target-charset my_grandma_bonnie" \
107 "Undefined item: \"my_grandma_bonnie\"." \
108 "try `set target-charset' with invalid charset"
110 # A Tcl array mapping the names of all the character sets we've seen
111 # to "1" if the character set can be used as a host character set, or
112 # "0" otherwise. We can use `array names charsets' just to get a list
113 # of all character sets.
114 array set charsets {}
116 proc all_charset_names {} {
118 return [array names charsets]
121 proc valid_host_charset {charset} {
123 return [expr {[info exists charsets($charset)] && $charsets($charset)}]
126 proc valid_target_charset {charset} {
128 return [info exists charsets($charset)]
131 send_gdb "set host-charset\n"
133 -re "Requires an argument. Valid arguments are (.*)\\.\r\n$gdb_prompt $" {
134 set host_charset_list $expect_out(1,string)
135 regsub -all {, } $host_charset_list {,} host_charset_list
136 foreach host_charset [split $host_charset_list ","] {
137 set charsets($host_charset) 1
139 pass "capture valid host charsets"
142 -re ".*$gdb_prompt $" {
143 fail "capture valid host charsets"
147 fail "(timeout) capture valid host charsets"
151 # If gdb was built with a phony iconv, it will only have two character
152 # sets: "auto" and the default. In this situation, this set of tests
154 if {[llength [array names charsets]] < 3} {
159 send_gdb "set target-charset\n"
161 -re "Requires an argument. Valid arguments are (.*)\\.\r\n$gdb_prompt $" {
162 set target_charset_list $expect_out(1,string)
163 regsub -all {, } $target_charset_list {,} target_charset_list
164 foreach target_charset [split $target_charset_list ","] {
165 if {! [info exists charsets($target_charset)]} {
166 set charsets($target_charset) 0
169 pass "capture valid target charsets"
172 -re ".*$gdb_prompt $" {
173 fail "capture valid target charsets"
177 fail "(timeout) capture valid target charsets"
181 # We don't want to test all the charset names here, since that would
182 # be too many combinations. We we pick a subset.
183 set charset_subset {ASCII ISO-8859-1 EBCDIC-US IBM1047}
184 foreach host_charset $charset_subset {
185 if {[valid_host_charset $host_charset]} {
187 set testname "try `set host-charset $host_charset'"
188 send_gdb "set host-charset $host_charset\n"
190 -re "GDB doesn't know of any character set named.*\[\r\n]+${gdb_prompt} $" {
191 # How did it get into `charsets' then?
192 fail "$testname (didn't recognize name)"
194 -re "GDB can't use `.*' as its host character set\\.\[\r\n]+${gdb_prompt} $" {
195 # Well, then why does its `charsets' entry say it can?
198 -re "${gdb_prompt} $" {
202 fail "$testname (timeout)"
206 # Check that the command actually had its intended effect:
207 # $host_charset should now be the host character set.
208 send_gdb "show charset\n"
209 set result [parse_show_charset_output "parse `show charset' after `set host-charset $host_charset'"]
210 if {! [string compare [lindex $result 0] $host_charset]} {
211 pass "check effect of `set host-charset $host_charset'"
213 fail "check effect of `set host-charset $host_charset'"
216 # Now try setting every possible target character set,
217 # given that host charset.
218 foreach target_charset $charset_subset {
219 if {![valid_target_charset $target_charset]} {
222 set testname "try `set target-charset $target_charset'"
223 send_gdb "set target-charset $target_charset\n"
225 -re "GDB doesn't know of any character set named.*\[\r\n]+${gdb_prompt} $" {
226 fail "$testname (didn't recognize name)"
228 -re "GDB can't convert from the .* character set to .*\\.\[\r\n\]+${gdb_prompt} $" {
229 # This is a serious problem. GDB should be able to convert
230 # between any arbitrary pair of character sets.
231 fail "$testname (can't convert)"
233 -re "${gdb_prompt} $" {
237 fail "$testname (timeout)"
241 # Check that the command actually had its intended effect:
242 # $target_charset should now be the target charset.
243 send_gdb "show charset\n"
244 set result [parse_show_charset_output "parse `show charset' after `set target-charset $target_charset'"]
245 if {! [string compare $result [list $host_charset $target_charset]]} {
246 pass "check effect of `set target-charset $target_charset'"
248 fail "check effect of `set target-charset $target_charset'"
251 # Test handling of characters in the host charset which
252 # can't be translated into the target charset. \xA2 is
253 # `cent' in ISO-8859-1, which has no equivalent in ASCII.
255 # On some systems, the pseudo-tty through which we
256 # communicate with GDB insists on stripping the high bit
257 # from input characters, meaning that `cent' turns into
258 # `"'. Since ISO-8859-1 and ASCII are identical in the
259 # lower 128 characters, it's tough to see how we can test
260 # this behavior on such systems, so we just xfail it.
262 # Note: the \x16 (Control-V) is an escape to allow \xA2 to
264 if {! [string compare $host_charset iso-8859-1] && ! [string compare $target_charset ascii]} {
266 set testname "untranslatable character in character literal"
267 send_gdb "print '\x16\xA2'\n"
269 -re "There is no character corresponding to .* in the target character set .*\\.\[\r\n\]+$gdb_prompt $" {
272 -re " = 34 '\"'\[\r\n\]+$gdb_prompt $" {
273 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
275 -re "$gdb_prompt $" {
279 fail "$testname (timeout)"
283 set testname "untranslatable character in string literal"
284 # If the PTTY zeros bit seven, then this turns into
286 # which gets us a syntax error. We don't care.
287 send_gdb "print \"\x16\xA2\"\n"
289 -re "There is no character corresponding to .* in the target character set .*\\.\[\r\n\]+$gdb_prompt $" {
292 -re "Unterminated string in expression.\[\r\n\]+$gdb_prompt $" {
293 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
295 -re "$gdb_prompt $" {
299 fail "$testname (timeout)"
303 set testname "untranslatable characters in backslash escape"
304 send_gdb "print '\\\x16\xA2'\n"
306 -re "The escape sequence .* is equivalent to plain .*, which has no equivalent\[\r\n\]+in the .* character set\\.\[\r\n\]+$gdb_prompt $" {
309 -re " = 34 '\"'\[\r\n\]+$gdb_prompt $" {
310 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
312 -re "$gdb_prompt $" {
316 fail "$testname (timeout)"
325 # Set the host character set to plain ASCII, and try actually printing
326 # some strings in various target character sets. We need to run the
327 # test program to the point at which the strings have been
329 gdb_test "break ${srcfile}:[gdb_get_line_number "all strings initialized"]" \
330 ".*Breakpoint.* at .*" \
331 "set breakpoint after all strings have been initialized"
334 -re "Breakpoint.*all strings initialized.*$gdb_prompt $" {
335 pass "run until all strings have been initialized"
337 -re "$gdb_prompt $" {
338 fail "run until all strings have been initialized"
341 fail "run until all strings have been initialized (timeout)"
346 # We only try the wide character tests on machines where the wchar_t
347 # typedef in the test case has the right size.
348 set wchar_size [get_sizeof wchar_t 99]
350 if {$wchar_size == 2} {
351 lappend charset_subset UTF-16
353 } elseif {$wchar_size == 4} {
354 lappend charset_subset UTF-32
358 gdb_test_no_output "set host-charset ASCII"
359 foreach target_charset $charset_subset {
360 if {![valid_target_charset $target_charset]} {
364 if {$target_charset == "UTF-32" || $target_charset == "UTF-16"} {
365 set param target-wide-charset
368 set param target-charset
371 gdb_test_no_output "set $param $target_charset"
373 # Try printing the null character. There seems to be a bug in
374 # gdb_test that requires us to use gdb_expect here.
375 send_gdb "print $L'\\0'\n"
377 -re "\\\$${decimal} = 0 $L'\\\\000'\[\r\n\]+$gdb_prompt $" {
378 pass "print the null character in ${target_charset}"
380 -re "$gdb_prompt $" {
381 fail "print the null character in ${target_charset}"
384 fail "print the null character in ${target_charset} (timeout)"
388 # Compute the name of the variable in the test program that holds
389 # a string in $target_charset. The variable's name is the
390 # character set's name, in lower-case, with all non-identifier
391 # characters replaced with '_', with "_string" stuck on the end.
392 if {$target_charset == "UTF-16"} {
393 # We still use the utf_32_string variable -- but the size is
394 # correct for UTF-16.
395 set var_name utf_32_string
397 set var_name [string tolower "${target_charset}_string"]
398 regsub -all -- "\[^a-z0-9_\]" $var_name "_" var_name
401 # Compute a regexp matching the results we expect. This is static,
402 # but it's easier than writing it out.
403 regsub -all "." "abfnrtv" "(\\\\&|x)" escapes
404 set uppercase "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
405 set lowercase "abcdefghijklmnopqrstuvwxyz"
406 set digits "0123456789"
407 set octal_escape "\\\\\[0-9\]+"
409 send_gdb "print $var_name\n"
410 # ${escapes}${uppercase}${lowercase}${digits}${octal}${octal}
412 -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 $" {
413 pass "print string in $target_charset"
415 -re "$gdb_prompt $" {
416 fail "print string in $target_charset"
419 fail "print string in $target_charset (timeout)"
423 # Try entering a character literal, and see if it comes back unchanged.
424 gdb_test "print $L'A'" \
425 " = \[0-9-\]+ $L'A'" \
426 "parse character literal in ${target_charset}"
428 # Check that the character literal was encoded correctly.
429 gdb_test "print $L'A' == $var_name\[7\]" \
431 "check value of parsed character literal in ${target_charset}"
433 # Try entering a string literal, and see if it comes back unchanged.
434 gdb_test "print $L\"abcdefABCDEF012345\"" \
435 " = $L\"abcdefABCDEF012345\"" \
436 "parse string literal in ${target_charset}"
438 # Check that the string literal was encoded correctly.
439 gdb_test "print $L\"q\"\[0\] == $var_name\[49\]" \
441 "check value of parsed string literal in ${target_charset}"
443 # Test handling of characters in the target charset which
444 # can't be translated into the host charset.
445 if {! [string compare $target_charset iso-8859-1]} {
446 gdb_test "print iso_8859_1_string\[69\]" \
447 " = \[0-9-\]+ '\\\\242'" \
448 "print character with no equivalent in host character set"
449 gdb_test "print iso_8859_1_string + 70" \
450 " = ${hex} \"\\\\242.*\"" \
451 "print string with no equivalent in host character set"
454 # Make sure that we don't apply the ISO-8859-1 `print_literally'
456 if {! [string compare $target_charset ascii]} {
457 gdb_test "print iso_8859_1_string\[69\]" \
458 " = \[0-9-\]+ '\\\\242'" \
459 "print ASCII unprintable character"
460 gdb_test "print iso_8859_1_string + 70" \
461 " = ${hex} \"\\\\242.*\"" \
462 "print ASCII unprintable string"
465 # Try printing characters with backslash escape equivalents.
466 set escapees {a b f n r t v}
467 for {set i 0} {$i < [llength $escapees]} {incr i} {
468 set escape [lindex $escapees $i]
469 send_gdb "print $var_name\[$i\]\n"
472 -re "= \[0-9-\]+ $L'\\\\${escape}'\[\r\n\]+$gdb_prompt $" {
473 pass "try printing '\\${escape}' in ${target_charset}"
475 -re "= \[0-9-\]+ 'x'\[\r\n\]+$gdb_prompt $" {
476 xfail "try printing '\\${escape}' in ${target_charset} (no such escape)"
479 -re "$gdb_prompt $" {
480 fail "try printing '\\${escape}' in ${target_charset}"
483 fail "try printing '\\${escape}' in ${target_charset} (timeout)"
489 # Try parsing a backslash escape in a character literal.
490 gdb_test "print $L'\\${escape}' == $var_name\[$i\]" \
492 "check value of '\\${escape}' in ${target_charset}"
494 # Try parsing a backslash escape in a string literal.
495 gdb_test "print $L\"\\${escape}\"\[0\] == $var_name\[$i\]" \
497 "check value of \"\\${escape}\" in ${target_charset}"
501 # Try printing a character escape that doesn't exist. We should
502 # get the unescaped character, in the target character set.
503 gdb_test "print $L'\\q'" " = \[0-9-\]+ $L'q'" \
504 "print escape that doesn't exist in $target_charset"
505 gdb_test "print $L'\\q' == $var_name\[49\]" " = 1" \
506 "check value of escape that doesn't exist in $target_charset"
509 # Reset the target charset.
510 gdb_test_no_output "set target-charset UTF-8"
512 # \242 is not a valid UTF-8 character.
513 gdb_test "print \"\\242\"" " = \"\\\\242\"" \
514 "non-representable target character"
516 gdb_test "print '\\x'" "\\\\x escape without a following hex digit."
517 gdb_test "print '\\u'" "\\\\u escape without a following hex digit."
518 gdb_test "print '\\9'" " = \[0-9\]+ '9'"
520 # An octal escape can only be 3 digits.
521 gdb_test "print \"\\1011\"" " = \"A1\""
523 # Tests for wide- or unicode- strings. L is the prefix letter to use,
524 # either "L" (for wide strings), "u" (for UTF-16), or "U" (for UTF-32).
525 # NAME is used in the test names and should be related to the prefix
526 # letter in some easy-to-undestand way.
527 proc test_wide_or_unicode {L name} {
528 gdb_test "print $L\"ab\" $L\"c\"" " = $L\"abc\"" \
529 "basic $name string concatenation"
530 gdb_test "print $L\"ab\" \"c\"" " = $L\"abc\"" \
531 "narrow and $name string concatenation"
532 gdb_test "print \"ab\" $L\"c\"" " = $L\"abc\"" \
533 "$name and narrow string concatenation"
534 gdb_test "print $L\"\\xe\" $L\"c\"" " = $L\"\\\\016c\"" \
535 "$name string concatenation with escape"
536 gdb_test "print $L\"\" \"abcdef\" \"g\"" \
538 "concatenate three strings with empty $name string"
540 gdb_test "print $L'a'" "= \[0-9\]+ $L'a'" \
541 "basic $name character"
545 test_wide_or_unicode L wide
548 set ucs2_ok [expr {[get_sizeof char16_t 99] == 2}]
550 if ![valid_host_charset "UTF-16"] {
551 verbose -log "Disabling UTF-16 tests."
556 test_wide_or_unicode u UTF-16
559 set ucs4_ok [expr {[get_sizeof char32_t 99] == 4}]
561 test_wide_or_unicode U UTF-32
564 # Test an invalid string combination.
565 proc test_combination {L1 name1 L2 name2} {
566 gdb_test "print $L1\"abc\" $L2\"def\"" \
567 "Undefined string concatenation." \
568 "undefined concatenation of $name1 and $name2"
571 if {$wchar_ok && $ucs2_ok} {
572 test_combination L wide u UTF-16
574 if {$wchar_ok && $ucs4_ok} {
575 test_combination L wide U UTF-32
576 # Regression test for a typedef to a typedef.
577 gdb_test "print myvar" "= \[0-9\]+ L'A'" \
580 if {$ucs2_ok && $ucs4_ok} {
581 test_combination u UTF-16 U UTF-32
586 gdb_test_multiple "python print 'hello, world!'" \
587 "verify python support for charset tests" {
588 -re "not supported.*$gdb_prompt $" {
589 unsupported "python support is disabled"
592 -re "$gdb_prompt $" {}
596 gdb_test "print u\"abcdef\"" " = u\"abcdef\"" \
597 "set up for python printing of utf-16 string"
599 gdb_test "python print gdb.history(0).string()" "abcdef" \
600 "extract utf-16 string using python"
604 # Regression test for a cleanup bug in the charset code.
605 gdb_test "print 'a' == 'a' || 'b' == 'b'" \
607 "EVAL_SKIP cleanup handling regression test"
610 proc string_display { var_name set_prefix x_size x_type} {
611 gdb_test_no_output "set ${var_name} = ${set_prefix}\"Test String\\0with zeroes\"" "Assign ${var_name} with prefix ${set_prefix}"
612 gdb_test "x /2${x_size}s ${var_name}" ".* ${x_type}\"Test String\"\[\r\n\]+.* ${x_type}\"with zeroes\"" "Display String ${var_name} with x/${x_size}s"
616 string_display String16 u h u
617 if {$wchar_size == 2} {
618 string_display String16 L h u
622 string_display String32 U w U
623 if {$wchar_size == 4} {
624 string_display String32 L w U
628 foreach name {short int long} {
629 # We're really just checking to make sure this doesn't give an
631 gdb_test "print ${name}_array = \"hi\"" \
633 "assign string to $name array"