1 # This testcase is part of GDB, the GNU debugger.
3 # Copyright 2001-2014 Free Software Foundation, Inc.
5 # This program is free software; you can redistribute it and/or modify
6 # it under the terms of the GNU General Public License as published by
7 # the Free Software Foundation; either version 3 of the License, or
8 # (at your option) any later version.
10 # This program is distributed in the hope that it will be useful,
11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 # GNU General Public License for more details.
15 # You should have received a copy of the GNU General Public License
16 # along with this program. If not, see <http://www.gnu.org/licenses/>.
18 # Please email any bugs, comments, and/or additions to this file to:
21 # Test GDB's character set support.
24 standard_testfile .c charset-malloc.c
26 if { [prepare_for_testing ${testfile}.exp ${testfile} [list $srcfile $srcfile2]] } {
30 # Parse the output from a `show charset' command. Return the host
31 # and target charset as a two-element list.
32 proc parse_show_charset_output {testname} {
36 -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 $" {
37 set host_charset $expect_out(1,string)
38 set target_charset $expect_out(2,string)
39 set retlist [list $host_charset $target_charset]
42 -re "The host character set is \"(.*)\"\\.\[\r\n\]+$gdb_prompt $" {
43 set host_charset $expect_out(1,string)
44 set retlist [list $host_charset]
47 -re "The target character set is \"(.*)\"\\.\[\r\n\]+$gdb_prompt $" {
48 set target_charset $expect_out(1,string)
49 set retlist [list $target_charset]
52 -re ".*$gdb_prompt $" {
56 fail "$testname (timeout)"
64 # Try the various `show charset' commands.
66 send_gdb "show charset\n"
67 set show_charset [parse_show_charset_output "show charset"]
69 send_gdb "show target-charset\n"
70 set show_target_charset \
71 [lindex [parse_show_charset_output "show target-charset"] 0]
73 if {[lsearch -exact $show_charset $show_target_charset] >= 0} {
74 pass "check `show target-charset' against `show charset'"
76 fail "check `show target-charset' against `show charset'"
79 send_gdb "show host-charset\n"
80 set show_host_charset \
81 [lindex [parse_show_charset_output "show host-charset"] 0]
83 if {[lsearch -exact $show_charset $show_host_charset] >= 0} {
84 pass "check `show host-charset' against `show charset'"
86 fail "check `show host-charset' against `show charset'"
89 # Try a malformed `set charset'.
90 gdb_test "set charset" \
91 "Requires an argument. Valid arguments are.*" \
92 "try malformed `set charset'"
94 # Try using `set host-charset' on an invalid character set.
95 gdb_test "set host-charset my_grandma_bonnie" \
96 "Undefined item: \"my_grandma_bonnie\"." \
97 "try `set host-charset' with invalid charset"
99 # Try using `set target-charset' on an invalid character set.
100 gdb_test "set target-charset my_grandma_bonnie" \
101 "Undefined item: \"my_grandma_bonnie\"." \
102 "try `set target-charset' with invalid charset"
104 # A Tcl array mapping the names of all the character sets we've seen
105 # to "1" if the character set can be used as a host character set, or
106 # "0" otherwise. We can use `array names charsets' just to get a list
107 # of all character sets.
108 array set charsets {}
110 proc all_charset_names {} {
112 return [array names charsets]
115 proc valid_host_charset {charset} {
117 return [expr {[info exists charsets($charset)] && $charsets($charset)}]
120 proc valid_target_charset {charset} {
122 return [info exists charsets($charset)]
125 send_gdb "set host-charset\n"
127 -re "Requires an argument. Valid arguments are (.*)\\.\r\n$gdb_prompt $" {
128 set host_charset_list $expect_out(1,string)
129 regsub -all {, } $host_charset_list {,} host_charset_list
130 foreach host_charset [split $host_charset_list ","] {
131 set charsets($host_charset) 1
133 pass "capture valid host charsets"
136 -re ".*$gdb_prompt $" {
137 fail "capture valid host charsets"
141 fail "(timeout) capture valid host charsets"
145 # If gdb was built with a phony iconv, it will only have two character
146 # sets: "auto" and the default. In this situation, this set of tests
148 if {[llength [array names charsets]] < 3} {
153 send_gdb "set target-charset\n"
155 -re "Requires an argument. Valid arguments are (.*)\\.\r\n$gdb_prompt $" {
156 set target_charset_list $expect_out(1,string)
157 regsub -all {, } $target_charset_list {,} target_charset_list
158 foreach target_charset [split $target_charset_list ","] {
159 if {! [info exists charsets($target_charset)]} {
160 set charsets($target_charset) 0
163 pass "capture valid target charsets"
166 -re ".*$gdb_prompt $" {
167 fail "capture valid target charsets"
171 fail "(timeout) capture valid target charsets"
175 # We don't want to test all the charset names here, since that would
176 # be too many combinations. We we pick a subset.
177 set charset_subset {ASCII ISO-8859-1 EBCDIC-US IBM1047}
178 foreach host_charset $charset_subset {
179 if {[valid_host_charset $host_charset]} {
181 set testname "try `set host-charset $host_charset'"
182 send_gdb "set host-charset $host_charset\n"
184 -re "GDB doesn't know of any character set named.*\[\r\n]+${gdb_prompt} $" {
185 # How did it get into `charsets' then?
186 fail "$testname (didn't recognize name)"
188 -re "GDB can't use `.*' as its host character set\\.\[\r\n]+${gdb_prompt} $" {
189 # Well, then why does its `charsets' entry say it can?
192 -re "${gdb_prompt} $" {
196 fail "$testname (timeout)"
200 # Check that the command actually had its intended effect:
201 # $host_charset should now be the host character set.
202 send_gdb "show charset\n"
203 set result [parse_show_charset_output "parse `show charset' after `set host-charset $host_charset'"]
204 if {! [string compare [lindex $result 0] $host_charset]} {
205 pass "check effect of `set host-charset $host_charset'"
207 fail "check effect of `set host-charset $host_charset'"
210 # Now try setting every possible target character set,
211 # given that host charset.
212 foreach target_charset $charset_subset {
213 if {![valid_target_charset $target_charset]} {
216 set testname "try `set target-charset $target_charset'"
217 send_gdb "set target-charset $target_charset\n"
219 -re "GDB doesn't know of any character set named.*\[\r\n]+${gdb_prompt} $" {
220 fail "$testname (didn't recognize name)"
222 -re "GDB can't convert from the .* character set to .*\\.\[\r\n\]+${gdb_prompt} $" {
223 # This is a serious problem. GDB should be able to convert
224 # between any arbitrary pair of character sets.
225 fail "$testname (can't convert)"
227 -re "${gdb_prompt} $" {
231 fail "$testname (timeout)"
235 # Check that the command actually had its intended effect:
236 # $target_charset should now be the target charset.
237 send_gdb "show charset\n"
238 set result [parse_show_charset_output "parse `show charset' after `set target-charset $target_charset'"]
239 if {! [string compare $result [list $host_charset $target_charset]]} {
240 pass "check effect of `set target-charset $target_charset'"
242 fail "check effect of `set target-charset $target_charset'"
245 # Test handling of characters in the host charset which
246 # can't be translated into the target charset. \xA2 is
247 # `cent' in ISO-8859-1, which has no equivalent in ASCII.
249 # On some systems, the pseudo-tty through which we
250 # communicate with GDB insists on stripping the high bit
251 # from input characters, meaning that `cent' turns into
252 # `"'. Since ISO-8859-1 and ASCII are identical in the
253 # lower 128 characters, it's tough to see how we can test
254 # this behavior on such systems, so we just xfail it.
256 # Note: the \x16 (Control-V) is an escape to allow \xA2 to
258 if {! [string compare $host_charset iso-8859-1] && ! [string compare $target_charset ascii]} {
260 set testname "untranslatable character in character literal"
261 send_gdb "print '\x16\xA2'\n"
263 -re "There is no character corresponding to .* in the target character set .*\\.\[\r\n\]+$gdb_prompt $" {
266 -re " = 34 '\"'\[\r\n\]+$gdb_prompt $" {
267 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
269 -re "$gdb_prompt $" {
273 fail "$testname (timeout)"
277 set testname "untranslatable character in string literal"
278 # If the PTTY zeros bit seven, then this turns into
280 # which gets us a syntax error. We don't care.
281 send_gdb "print \"\x16\xA2\"\n"
283 -re "There is no character corresponding to .* in the target character set .*\\.\[\r\n\]+$gdb_prompt $" {
286 -re "Unterminated string in expression.\[\r\n\]+$gdb_prompt $" {
287 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
289 -re "$gdb_prompt $" {
293 fail "$testname (timeout)"
297 set testname "untranslatable characters in backslash escape"
298 send_gdb "print '\\\x16\xA2'\n"
300 -re "The escape sequence .* is equivalent to plain .*, which has no equivalent\[\r\n\]+in the .* character set\\.\[\r\n\]+$gdb_prompt $" {
303 -re " = 34 '\"'\[\r\n\]+$gdb_prompt $" {
304 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
306 -re "$gdb_prompt $" {
310 fail "$testname (timeout)"
319 # Set the host character set to plain ASCII, and try actually printing
320 # some strings in various target character sets. We need to run the
321 # test program to the point at which the strings have been
323 gdb_test "break ${srcfile}:[gdb_get_line_number "all strings initialized"]" \
324 ".*Breakpoint.* at .*" \
325 "set breakpoint after all strings have been initialized"
328 -re "Breakpoint.*all strings initialized.*$gdb_prompt $" {
329 pass "run until all strings have been initialized"
331 -re "$gdb_prompt $" {
332 fail "run until all strings have been initialized"
335 fail "run until all strings have been initialized (timeout)"
340 # We only try the wide character tests on machines where the wchar_t
341 # typedef in the test case has the right size.
342 set wchar_size [get_sizeof wchar_t 99]
344 if {$wchar_size == 2} {
345 lappend charset_subset UTF-16
347 } elseif {$wchar_size == 4} {
348 lappend charset_subset UTF-32
352 gdb_test_no_output "set host-charset ASCII"
353 foreach target_charset $charset_subset {
354 if {![valid_target_charset $target_charset]} {
358 if {$target_charset == "UTF-32" || $target_charset == "UTF-16"} {
359 set param target-wide-charset
362 set param target-charset
365 gdb_test_no_output "set $param $target_charset"
367 # Try printing the null character. There seems to be a bug in
368 # gdb_test that requires us to use gdb_expect here.
369 send_gdb "print $L'\\0'\n"
371 -re "\\\$${decimal} = 0 $L'\\\\000'\[\r\n\]+$gdb_prompt $" {
372 pass "print the null character in ${target_charset}"
374 -re "$gdb_prompt $" {
375 fail "print the null character in ${target_charset}"
378 fail "print the null character in ${target_charset} (timeout)"
382 # Compute the name of the variable in the test program that holds
383 # a string in $target_charset. The variable's name is the
384 # character set's name, in lower-case, with all non-identifier
385 # characters replaced with '_', with "_string" stuck on the end.
386 if {$target_charset == "UTF-16"} {
387 # We still use the utf_32_string variable -- but the size is
388 # correct for UTF-16.
389 set var_name utf_32_string
391 set var_name [string tolower "${target_charset}_string"]
392 regsub -all -- "\[^a-z0-9_\]" $var_name "_" var_name
395 # Compute a regexp matching the results we expect. This is static,
396 # but it's easier than writing it out.
397 regsub -all "." "abfnrtv" "(\\\\&|x)" escapes
398 set uppercase "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
399 set lowercase "abcdefghijklmnopqrstuvwxyz"
400 set digits "0123456789"
401 set octal_escape "\\\\\[0-9\]+"
403 send_gdb "print $var_name\n"
404 # ${escapes}${uppercase}${lowercase}${digits}${octal}${octal}
406 -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 $" {
407 pass "print string in $target_charset"
409 -re "$gdb_prompt $" {
410 fail "print string in $target_charset"
413 fail "print string in $target_charset (timeout)"
417 # Try entering a character literal, and see if it comes back unchanged.
418 gdb_test "print $L'A'" \
419 " = \[0-9-\]+ $L'A'" \
420 "parse character literal in ${target_charset}"
422 # Check that the character literal was encoded correctly.
423 gdb_test "print $L'A' == $var_name\[7\]" \
425 "check value of parsed character literal in ${target_charset}"
427 # Try entering a string literal, and see if it comes back unchanged.
428 gdb_test "print $L\"abcdefABCDEF012345\"" \
429 " = $L\"abcdefABCDEF012345\"" \
430 "parse string literal in ${target_charset}"
432 # Check that the string literal was encoded correctly.
433 gdb_test "print $L\"q\"\[0\] == $var_name\[49\]" \
435 "check value of parsed string literal in ${target_charset}"
437 # Test handling of characters in the target charset which
438 # can't be translated into the host charset.
439 if {! [string compare $target_charset iso-8859-1]} {
440 gdb_test "print iso_8859_1_string\[69\]" \
441 " = \[0-9-\]+ '\\\\242'" \
442 "print character with no equivalent in host character set"
443 gdb_test "print iso_8859_1_string + 70" \
444 " = ${hex} \"\\\\242.*\"" \
445 "print string with no equivalent in host character set"
448 # Make sure that we don't apply the ISO-8859-1 `print_literally'
450 if {! [string compare $target_charset ascii]} {
451 gdb_test "print iso_8859_1_string\[69\]" \
452 " = \[0-9-\]+ '\\\\242'" \
453 "print ASCII unprintable character"
454 gdb_test "print iso_8859_1_string + 70" \
455 " = ${hex} \"\\\\242.*\"" \
456 "print ASCII unprintable string"
459 # Try printing characters with backslash escape equivalents.
460 set escapees {a b f n r t v}
461 for {set i 0} {$i < [llength $escapees]} {incr i} {
462 set escape [lindex $escapees $i]
463 send_gdb "print $var_name\[$i\]\n"
466 -re "= \[0-9-\]+ $L'\\\\${escape}'\[\r\n\]+$gdb_prompt $" {
467 pass "try printing '\\${escape}' in ${target_charset}"
469 -re "= \[0-9-\]+ 'x'\[\r\n\]+$gdb_prompt $" {
470 xfail "try printing '\\${escape}' in ${target_charset} (no such escape)"
473 -re "$gdb_prompt $" {
474 fail "try printing '\\${escape}' in ${target_charset}"
477 fail "try printing '\\${escape}' in ${target_charset} (timeout)"
483 # Try parsing a backslash escape in a character literal.
484 gdb_test "print $L'\\${escape}' == $var_name\[$i\]" \
486 "check value of '\\${escape}' in ${target_charset}"
488 # Try parsing a backslash escape in a string literal.
489 gdb_test "print $L\"\\${escape}\"\[0\] == $var_name\[$i\]" \
491 "check value of \"\\${escape}\" in ${target_charset}"
495 # Try printing a character escape that doesn't exist. We should
496 # get the unescaped character, in the target character set.
497 gdb_test "print $L'\\q'" " = \[0-9-\]+ $L'q'" \
498 "print escape that doesn't exist in $target_charset"
499 gdb_test "print $L'\\q' == $var_name\[49\]" " = 1" \
500 "check value of escape that doesn't exist in $target_charset"
503 # Reset the target charset.
504 gdb_test_no_output "set target-charset UTF-8"
506 # \242 is not a valid UTF-8 character.
507 gdb_test "print \"\\242\"" " = \"\\\\242\"" \
508 "non-representable target character"
510 gdb_test "print '\\x'" "\\\\x escape without a following hex digit."
511 gdb_test "print '\\u'" "\\\\u escape without a following hex digit."
512 gdb_test "print '\\9'" " = \[0-9\]+ '9'"
514 # An octal escape can only be 3 digits.
515 gdb_test "print \"\\1011\"" " = \"A1\""
517 # Tests for wide- or unicode- strings. L is the prefix letter to use,
518 # either "L" (for wide strings), "u" (for UTF-16), or "U" (for UTF-32).
519 # NAME is used in the test names and should be related to the prefix
520 # letter in some easy-to-undestand way.
521 proc test_wide_or_unicode {L name} {
522 gdb_test "print $L\"ab\" $L\"c\"" " = $L\"abc\"" \
523 "basic $name string concatenation"
524 gdb_test "print $L\"ab\" \"c\"" " = $L\"abc\"" \
525 "narrow and $name string concatenation"
526 gdb_test "print \"ab\" $L\"c\"" " = $L\"abc\"" \
527 "$name and narrow string concatenation"
528 gdb_test "print $L\"\\xe\" $L\"c\"" " = $L\"\\\\016c\"" \
529 "$name string concatenation with escape"
530 gdb_test "print $L\"\" \"abcdef\" \"g\"" \
532 "concatenate three strings with empty $name string"
534 gdb_test "print $L'a'" "= \[0-9\]+ $L'a'" \
535 "basic $name character"
539 test_wide_or_unicode L wide
542 set ucs2_ok [expr {[get_sizeof char16_t 99] == 2}]
544 if ![valid_host_charset "UTF-16"] {
545 verbose -log "Disabling UTF-16 tests."
550 test_wide_or_unicode u UTF-16
553 set ucs4_ok [expr {[get_sizeof char32_t 99] == 4}]
555 test_wide_or_unicode U UTF-32
558 # Test an invalid string combination.
559 proc test_combination {L1 name1 L2 name2} {
560 gdb_test "print $L1\"abc\" $L2\"def\"" \
561 "Undefined string concatenation." \
562 "undefined concatenation of $name1 and $name2"
565 if {$wchar_ok && $ucs2_ok} {
566 test_combination L wide u UTF-16
568 if {$wchar_ok && $ucs4_ok} {
569 test_combination L wide U UTF-32
570 # Regression test for a typedef to a typedef.
571 gdb_test "print myvar" "= \[0-9\]+ L'A'" \
574 if {$ucs2_ok && $ucs4_ok} {
575 test_combination u UTF-16 U UTF-32
580 gdb_test_multiple "python print ('hello, world!')" \
581 "verify python support for charset tests" {
582 -re "not supported.*$gdb_prompt $" {
583 unsupported "python support is disabled"
586 -re "$gdb_prompt $" {}
590 gdb_test "print u\"abcdef\"" " = u\"abcdef\"" \
591 "set up for python printing of utf-16 string"
593 gdb_test "python print (gdb.history(0).string())" "abcdef" \
594 "extract utf-16 string using python"
598 # Regression test for a cleanup bug in the charset code.
599 gdb_test "print 'a' == 'a' || 'b' == 'b'" \
601 "EVAL_SKIP cleanup handling regression test"
604 proc string_display { var_name set_prefix x_size x_type} {
605 gdb_test_no_output "set ${var_name} = ${set_prefix}\"Test String\\0with zeroes\"" "Assign ${var_name} with prefix ${set_prefix}"
606 gdb_test "x /2${x_size}s ${var_name}" ".*\t${x_type}\"Test String\"\[\r\n\]+.*\t${x_type}\"with zeroes\"" "Display String ${var_name} with x/${x_size}s"
610 string_display String16 u h u
611 if {$wchar_size == 2} {
612 string_display String16 L h u
616 string_display String32 U w U
617 if {$wchar_size == 4} {
618 string_display String32 L w U
622 foreach name {short int long} {
623 # We're really just checking to make sure this doesn't give an
625 gdb_test "print ${name}_array = \"hi\"" \
627 "assign string to $name array"