1 # This testcase is part of GDB, the GNU debugger.
3 # Copyright 2001, 2004, 2007, 2008, 2009, 2010 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.
28 set testfile "charset"
29 set srcfile ${testfile}.c
30 set srcmallocfile ${testfile}-malloc.c
31 if { [prepare_for_testing ${testfile}.exp ${testfile} [list $srcfile $srcmallocfile]] } {
35 # Parse the output from a `show charset' command. Return the host
36 # and target charset as a two-element list.
37 proc parse_show_charset_output {testname} {
41 -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 $" {
42 set host_charset $expect_out(1,string)
43 set target_charset $expect_out(2,string)
44 set retlist [list $host_charset $target_charset]
47 -re "The host character set is \"(.*)\"\\.\[\r\n\]+$gdb_prompt $" {
48 set host_charset $expect_out(1,string)
49 set retlist [list $host_charset]
52 -re "The target character set is \"(.*)\"\\.\[\r\n\]+$gdb_prompt $" {
53 set target_charset $expect_out(1,string)
54 set retlist [list $target_charset]
57 -re ".*$gdb_prompt $" {
61 fail "$testname (timeout)"
69 # Try the various `show charset' commands.
71 send_gdb "show charset\n"
72 set show_charset [parse_show_charset_output "show charset"]
74 send_gdb "show target-charset\n"
75 set show_target_charset \
76 [lindex [parse_show_charset_output "show target-charset"] 0]
78 if {[lsearch -exact $show_charset $show_target_charset] >= 0} {
79 pass "check `show target-charset' against `show charset'"
81 fail "check `show target-charset' against `show charset'"
84 send_gdb "show host-charset\n"
85 set show_host_charset \
86 [lindex [parse_show_charset_output "show host-charset"] 0]
88 if {[lsearch -exact $show_charset $show_host_charset] >= 0} {
89 pass "check `show host-charset' against `show charset'"
91 fail "check `show host-charset' against `show charset'"
94 # Try a malformed `set charset'.
95 gdb_test "set charset" \
96 "Requires an argument. Valid arguments are.*" \
97 "try malformed `set charset'"
99 # Try using `set host-charset' on an invalid character set.
100 gdb_test "set host-charset my_grandma_bonnie" \
101 "Undefined item: \"my_grandma_bonnie\"." \
102 "try `set host-charset' with invalid charset"
104 # Try using `set target-charset' on an invalid character set.
105 gdb_test "set target-charset my_grandma_bonnie" \
106 "Undefined item: \"my_grandma_bonnie\"." \
107 "try `set target-charset' with invalid charset"
109 # A Tcl array mapping the names of all the character sets we've seen
110 # to "1" if the character set can be used as a host character set, or
111 # "0" otherwise. We can use `array names charsets' just to get a list
112 # of all character sets.
113 array set charsets {}
115 proc all_charset_names {} {
117 return [array names charsets]
120 proc valid_host_charset {charset} {
122 return [expr {[info exists charsets($charset)] && $charsets($charset)}]
125 proc valid_target_charset {charset} {
127 return [info exists charsets($charset)]
130 send_gdb "set host-charset\n"
132 -re "Requires an argument. Valid arguments are (.*)\\.\r\n$gdb_prompt $" {
133 set host_charset_list $expect_out(1,string)
134 regsub -all {, } $host_charset_list {,} host_charset_list
135 foreach host_charset [split $host_charset_list ","] {
136 set charsets($host_charset) 1
138 pass "capture valid host charsets"
141 -re ".*$gdb_prompt $" {
142 fail "capture valid host charsets"
146 fail "(timeout) capture valid host charsets"
150 # If gdb was built with a phony iconv, it will only have two character
151 # sets: "auto" and the default. In this situation, this set of tests
153 if {[llength [array names charsets]] < 3} {
158 send_gdb "set target-charset\n"
160 -re "Requires an argument. Valid arguments are (.*)\\.\r\n$gdb_prompt $" {
161 set target_charset_list $expect_out(1,string)
162 regsub -all {, } $target_charset_list {,} target_charset_list
163 foreach target_charset [split $target_charset_list ","] {
164 if {! [info exists charsets($target_charset)]} {
165 set charsets($target_charset) 0
168 pass "capture valid target charsets"
171 -re ".*$gdb_prompt $" {
172 fail "capture valid target charsets"
176 fail "(timeout) capture valid target charsets"
180 # We don't want to test all the charset names here, since that would
181 # be too many combinations. We we pick a subset.
182 set charset_subset {ASCII ISO-8859-1 EBCDIC-US IBM1047}
183 foreach host_charset $charset_subset {
184 if {[valid_host_charset $host_charset]} {
186 set testname "try `set host-charset $host_charset'"
187 send_gdb "set host-charset $host_charset\n"
189 -re "GDB doesn't know of any character set named.*\[\r\n]+${gdb_prompt} $" {
190 # How did it get into `charsets' then?
191 fail "$testname (didn't recognize name)"
193 -re "GDB can't use `.*' as its host character set\\.\[\r\n]+${gdb_prompt} $" {
194 # Well, then why does its `charsets' entry say it can?
197 -re "${gdb_prompt} $" {
201 fail "$testname (timeout)"
205 # Check that the command actually had its intended effect:
206 # $host_charset should now be the host character set.
207 send_gdb "show charset\n"
208 set result [parse_show_charset_output "parse `show charset' after `set host-charset $host_charset'"]
209 if {! [string compare [lindex $result 0] $host_charset]} {
210 pass "check effect of `set host-charset $host_charset'"
212 fail "check effect of `set host-charset $host_charset'"
215 # Now try setting every possible target character set,
216 # given that host charset.
217 foreach target_charset $charset_subset {
218 if {![valid_target_charset $target_charset]} {
221 set testname "try `set target-charset $target_charset'"
222 send_gdb "set target-charset $target_charset\n"
224 -re "GDB doesn't know of any character set named.*\[\r\n]+${gdb_prompt} $" {
225 fail "$testname (didn't recognize name)"
227 -re "GDB can't convert from the .* character set to .*\\.\[\r\n\]+${gdb_prompt} $" {
228 # This is a serious problem. GDB should be able to convert
229 # between any arbitrary pair of character sets.
230 fail "$testname (can't convert)"
232 -re "${gdb_prompt} $" {
236 fail "$testname (timeout)"
240 # Check that the command actually had its intended effect:
241 # $target_charset should now be the target charset.
242 send_gdb "show charset\n"
243 set result [parse_show_charset_output "parse `show charset' after `set target-charset $target_charset'"]
244 if {! [string compare $result [list $host_charset $target_charset]]} {
245 pass "check effect of `set target-charset $target_charset'"
247 fail "check effect of `set target-charset $target_charset'"
250 # Test handling of characters in the host charset which
251 # can't be translated into the target charset. \xA2 is
252 # `cent' in ISO-8859-1, which has no equivalent in ASCII.
254 # On some systems, the pseudo-tty through which we
255 # communicate with GDB insists on stripping the high bit
256 # from input characters, meaning that `cent' turns into
257 # `"'. Since ISO-8859-1 and ASCII are identical in the
258 # lower 128 characters, it's tough to see how we can test
259 # this behavior on such systems, so we just xfail it.
261 # Note: the \x16 (Control-V) is an escape to allow \xA2 to
263 if {! [string compare $host_charset iso-8859-1] && ! [string compare $target_charset ascii]} {
265 set testname "untranslatable character in character literal"
266 send_gdb "print '\x16\xA2'\n"
268 -re "There is no character corresponding to .* in the target character set .*\\.\[\r\n\]+$gdb_prompt $" {
271 -re " = 34 '\"'\[\r\n\]+$gdb_prompt $" {
272 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
274 -re "$gdb_prompt $" {
278 fail "$testname (timeout)"
282 set testname "untranslatable character in string literal"
283 # If the PTTY zeros bit seven, then this turns into
285 # which gets us a syntax error. We don't care.
286 send_gdb "print \"\x16\xA2\"\n"
288 -re "There is no character corresponding to .* in the target character set .*\\.\[\r\n\]+$gdb_prompt $" {
291 -re "Unterminated string in expression.\[\r\n\]+$gdb_prompt $" {
292 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
294 -re "$gdb_prompt $" {
298 fail "$testname (timeout)"
302 set testname "untranslatable characters in backslash escape"
303 send_gdb "print '\\\x16\xA2'\n"
305 -re "The escape sequence .* is equivalent to plain .*, which has no equivalent\[\r\n\]+in the .* character set\\.\[\r\n\]+$gdb_prompt $" {
308 -re " = 34 '\"'\[\r\n\]+$gdb_prompt $" {
309 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
311 -re "$gdb_prompt $" {
315 fail "$testname (timeout)"
324 # Set the host character set to plain ASCII, and try actually printing
325 # some strings in various target character sets. We need to run the
326 # test program to the point at which the strings have been
328 gdb_test "break ${srcfile}:[gdb_get_line_number "all strings initialized"]" \
329 ".*Breakpoint.* at .*" \
330 "set breakpoint after all strings have been initialized"
333 -re "Breakpoint.*all strings initialized.*$gdb_prompt $" {
334 pass "run until all strings have been initialized"
336 -re "$gdb_prompt $" {
337 fail "run until all strings have been initialized"
340 fail "run until all strings have been initialized (timeout)"
345 # We only try the wide character tests on machines where the wchar_t
346 # typedef in the test case has the right size.
347 set wchar_size [get_sizeof wchar_t 99]
349 if {$wchar_size == 2} {
350 lappend charset_subset UTF-16
352 } elseif {$wchar_size == 4} {
353 lappend charset_subset UTF-32
357 gdb_test_no_output "set host-charset ASCII"
358 foreach target_charset $charset_subset {
359 if {![valid_target_charset $target_charset]} {
363 if {$target_charset == "UTF-32" || $target_charset == "UTF-16"} {
364 set param target-wide-charset
367 set param target-charset
370 gdb_test_no_output "set $param $target_charset"
372 # Try printing the null character. There seems to be a bug in
373 # gdb_test that requires us to use gdb_expect here.
374 send_gdb "print $L'\\0'\n"
376 -re "\\\$${decimal} = 0 $L'\\\\000'\[\r\n\]+$gdb_prompt $" {
377 pass "print the null character in ${target_charset}"
379 -re "$gdb_prompt $" {
380 fail "print the null character in ${target_charset}"
383 fail "print the null character in ${target_charset} (timeout)"
387 # Compute the name of the variable in the test program that holds
388 # a string in $target_charset. The variable's name is the
389 # character set's name, in lower-case, with all non-identifier
390 # characters replaced with '_', with "_string" stuck on the end.
391 if {$target_charset == "UTF-16"} {
392 # We still use the utf_32_string variable -- but the size is
393 # correct for UTF-16.
394 set var_name utf_32_string
396 set var_name [string tolower "${target_charset}_string"]
397 regsub -all -- "\[^a-z0-9_\]" $var_name "_" var_name
400 # Compute a regexp matching the results we expect. This is static,
401 # but it's easier than writing it out.
402 regsub -all "." "abfnrtv" "(\\\\&|x)" escapes
403 set uppercase "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
404 set lowercase "abcdefghijklmnopqrstuvwxyz"
405 set digits "0123456789"
406 set octal_escape "\\\\\[0-9\]+"
408 send_gdb "print $var_name\n"
409 # ${escapes}${uppercase}${lowercase}${digits}${octal}${octal}
411 -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 $" {
412 pass "print string in $target_charset"
414 -re "$gdb_prompt $" {
415 fail "print string in $target_charset"
418 fail "print string in $target_charset (timeout)"
422 # Try entering a character literal, and see if it comes back unchanged.
423 gdb_test "print $L'A'" \
424 " = \[0-9-\]+ $L'A'" \
425 "parse character literal in ${target_charset}"
427 # Check that the character literal was encoded correctly.
428 gdb_test "print $L'A' == $var_name\[7\]" \
430 "check value of parsed character literal in ${target_charset}"
432 # Try entering a string literal, and see if it comes back unchanged.
433 gdb_test "print $L\"abcdefABCDEF012345\"" \
434 " = $L\"abcdefABCDEF012345\"" \
435 "parse string literal in ${target_charset}"
437 # Check that the string literal was encoded correctly.
438 gdb_test "print $L\"q\"\[0\] == $var_name\[49\]" \
440 "check value of parsed string literal in ${target_charset}"
442 # Test handling of characters in the target charset which
443 # can't be translated into the host charset.
444 if {! [string compare $target_charset iso-8859-1]} {
445 gdb_test "print iso_8859_1_string\[69\]" \
446 " = \[0-9-\]+ '\\\\242'" \
447 "print character with no equivalent in host character set"
448 gdb_test "print iso_8859_1_string + 70" \
449 " = ${hex} \"\\\\242.*\"" \
450 "print string with no equivalent in host character set"
453 # Make sure that we don't apply the ISO-8859-1 `print_literally'
455 if {! [string compare $target_charset ascii]} {
456 gdb_test "print iso_8859_1_string\[69\]" \
457 " = \[0-9-\]+ '\\\\242'" \
458 "print ASCII unprintable character"
459 gdb_test "print iso_8859_1_string + 70" \
460 " = ${hex} \"\\\\242.*\"" \
461 "print ASCII unprintable string"
464 # Try printing characters with backslash escape equivalents.
465 set escapees {a b f n r t v}
466 for {set i 0} {$i < [llength $escapees]} {incr i} {
467 set escape [lindex $escapees $i]
468 send_gdb "print $var_name\[$i\]\n"
471 -re "= \[0-9-\]+ $L'\\\\${escape}'\[\r\n\]+$gdb_prompt $" {
472 pass "try printing '\\${escape}' in ${target_charset}"
474 -re "= \[0-9-\]+ 'x'\[\r\n\]+$gdb_prompt $" {
475 xfail "try printing '\\${escape}' in ${target_charset} (no such escape)"
478 -re "$gdb_prompt $" {
479 fail "try printing '\\${escape}' in ${target_charset}"
482 fail "try printing '\\${escape}' in ${target_charset} (timeout)"
488 # Try parsing a backslash escape in a character literal.
489 gdb_test "print $L'\\${escape}' == $var_name\[$i\]" \
491 "check value of '\\${escape}' in ${target_charset}"
493 # Try parsing a backslash escape in a string literal.
494 gdb_test "print $L\"\\${escape}\"\[0\] == $var_name\[$i\]" \
496 "check value of \"\\${escape}\" in ${target_charset}"
500 # Try printing a character escape that doesn't exist. We should
501 # get the unescaped character, in the target character set.
502 gdb_test "print $L'\\q'" " = \[0-9-\]+ $L'q'" \
503 "print escape that doesn't exist in $target_charset"
504 gdb_test "print $L'\\q' == $var_name\[49\]" " = 1" \
505 "check value of escape that doesn't exist in $target_charset"
508 # Reset the target charset.
509 gdb_test_no_output "set target-charset UTF-8"
511 # \242 is not a valid UTF-8 character.
512 gdb_test "print \"\\242\"" " = \"\\\\242\"" \
513 "non-representable target character"
515 gdb_test "print '\\x'" "\\\\x escape without a following hex digit."
516 gdb_test "print '\\u'" "\\\\u escape without a following hex digit."
517 gdb_test "print '\\9'" " = \[0-9\]+ '9'"
519 # An octal escape can only be 3 digits.
520 gdb_test "print \"\\1011\"" " = \"A1\""
522 # Tests for wide- or unicode- strings. L is the prefix letter to use,
523 # either "L" (for wide strings), "u" (for UTF-16), or "U" (for UTF-32).
524 # NAME is used in the test names and should be related to the prefix
525 # letter in some easy-to-undestand way.
526 proc test_wide_or_unicode {L name} {
527 gdb_test "print $L\"ab\" $L\"c\"" " = $L\"abc\"" \
528 "basic $name string concatenation"
529 gdb_test "print $L\"ab\" \"c\"" " = $L\"abc\"" \
530 "narrow and $name string concatenation"
531 gdb_test "print \"ab\" $L\"c\"" " = $L\"abc\"" \
532 "$name and narrow string concatenation"
533 gdb_test "print $L\"\\xe\" $L\"c\"" " = $L\"\\\\016c\"" \
534 "$name string concatenation with escape"
535 gdb_test "print $L\"\" \"abcdef\" \"g\"" \
537 "concatenate three strings with empty $name string"
539 gdb_test "print $L'a'" "= \[0-9\]+ $L'a'" \
540 "basic $name character"
544 test_wide_or_unicode L wide
547 set ucs2_ok [expr {[get_sizeof char16_t 99] == 2}]
549 test_wide_or_unicode u UTF-16
552 set ucs4_ok [expr {[get_sizeof char32_t 99] == 4}]
554 test_wide_or_unicode U UTF-32
557 # Test an invalid string combination.
558 proc test_combination {L1 name1 L2 name2} {
559 gdb_test "print $L1\"abc\" $L2\"def\"" \
560 "Undefined string concatenation." \
561 "undefined concatenation of $name1 and $name2"
564 if {$wchar_ok && $ucs2_ok} {
565 test_combination L wide u UTF-16
567 if {$wchar_ok && $ucs4_ok} {
568 test_combination L wide U UTF-32
569 # Regression test for a typedef to a typedef.
570 gdb_test "print myvar" "= \[0-9\]+ L'A'" \
573 if {$ucs2_ok && $ucs4_ok} {
574 test_combination u UTF-16 U UTF-32
579 gdb_test_multiple "python print 'hello, world!'" \
580 "verify python support for charset tests" {
581 -re "not supported.*$gdb_prompt $" {
582 unsupported "python support is disabled"
585 -re "$gdb_prompt $" {}
589 gdb_test "print u\"abcdef\"" " = u\"abcdef\"" \
590 "set up for python printing of utf-16 string"
592 gdb_test "python print gdb.history(0).string()" "abcdef" \
593 "extract utf-16 string using python"
597 # Regression test for a cleanup bug in the charset code.
598 gdb_test "print 'a' == 'a' || 'b' == 'b'" \
600 "EVAL_SKIP cleanup handling regression test"
603 proc string_display { var_name set_prefix x_size x_type} {
604 gdb_test_no_output "set ${var_name} = ${set_prefix}\"Test String\\0with zeroes\"" "Assign ${var_name} with prefix ${set_prefix}"
605 gdb_test "x /2${x_size}s ${var_name}" ".* ${x_type}\"Test String\"\[\r\n\]+.* ${x_type}\"with zeroes\"" "Display String ${var_name} with x/${x_size}s"
608 string_display String16 u h u
609 if {$wchar_size == 2} {
610 string_display String16 L h u
613 string_display String32 U w U
614 if {$wchar_size == 4} {
615 string_display String32 L w U