1 # This testcase is part of GDB, the GNU debugger.
3 # Copyright 2001, 2004, 2007, 2008, 2009 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.
30 set testfile "charset"
31 set srcfile ${testfile}.c
32 set binfile ${objdir}/${subdir}/${testfile}
33 if { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" executable {debug}] != "" } {
34 untested "couldn't compile ${srcdir}/${subdir}/${srcfile}"
38 # Start with a fresh gdb.
41 gdb_reinitialize_dir $srcdir/$subdir
44 # Parse the output from a `show charset' command. Return the host
45 # and target charset as a two-element list.
46 proc parse_show_charset_output {testname} {
50 -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 $" {
51 set host_charset $expect_out(1,string)
52 set target_charset $expect_out(2,string)
53 set retlist [list $host_charset $target_charset]
56 -re "The host character set is \"(.*)\"\\.\[\r\n\]+$gdb_prompt $" {
57 set host_charset $expect_out(1,string)
58 set retlist [list $host_charset]
61 -re "The target character set is \"(.*)\"\\.\[\r\n\]+$gdb_prompt $" {
62 set target_charset $expect_out(1,string)
63 set retlist [list $target_charset]
66 -re ".*$gdb_prompt $" {
70 fail "$testname (timeout)"
78 # Try the various `show charset' commands.
80 send_gdb "show charset\n"
81 set show_charset [parse_show_charset_output "show charset"]
83 send_gdb "show target-charset\n"
84 set show_target_charset \
85 [lindex [parse_show_charset_output "show target-charset"] 0]
87 if {[lsearch -exact $show_charset $show_target_charset] >= 0} {
88 pass "check `show target-charset' against `show charset'"
90 fail "check `show target-charset' against `show charset'"
93 send_gdb "show host-charset\n"
94 set show_host_charset \
95 [lindex [parse_show_charset_output "show host-charset"] 0]
97 if {[lsearch -exact $show_charset $show_host_charset] >= 0} {
98 pass "check `show host-charset' against `show charset'"
100 fail "check `show host-charset' against `show charset'"
103 # Try a malformed `set charset'.
104 gdb_test "set charset" \
105 "Requires an argument. Valid arguments are.*" \
106 "try malformed `set charset'"
108 # Try using `set host-charset' on an invalid character set.
109 gdb_test "set host-charset my_grandma_bonnie" \
110 "Undefined item: \"my_grandma_bonnie\"." \
111 "try `set host-charset' with invalid charset"
113 # Try using `set target-charset' on an invalid character set.
114 gdb_test "set target-charset my_grandma_bonnie" \
115 "Undefined item: \"my_grandma_bonnie\"." \
116 "try `set target-charset' with invalid charset"
118 # A Tcl array mapping the names of all the character sets we've seen
119 # to "1" if the character set can be used as a host character set, or
120 # "0" otherwise. We can use `array names charsets' just to get a list
121 # of all character sets.
122 array set charsets {}
124 proc all_charset_names {} {
126 return [array names charsets]
129 proc valid_host_charset {charset} {
131 return [expr {[info exists charsets($charset)] && $charsets($charset)}]
134 proc valid_target_charset {charset} {
136 return [info exists charsets($charset)]
139 send_gdb "set host-charset\n"
141 -re "Requires an argument. Valid arguments are (\[^ \t\n\r,.\]*)" {
142 #set host_charset_list $expect_out(1,string)
143 set charsets($expect_out(1,string)) 1
145 #pass "capture valid host charsets"
148 -re ", (\[^ \t\n\r,.\]*)" {
149 #set host_charset_list $expect_out(1,string)
150 set charsets($expect_out(1,string)) 1
152 #pass "capture valid host charsets"
155 -re "\\.\r\n$gdb_prompt $" {
156 #set host_charset_list $expect_out(1,string)
157 pass "capture valid host charsets"
160 -re ".*$gdb_prompt $" {
161 fail "capture valid host charsets"
164 fail "(timeout) capture valid host charsets"
168 # If gdb was built with a phony iconv, it will only have two character
169 # sets: "auto" and the default. In this situation, this set of tests
171 if {[llength [array names charsets]] < 3} {
176 send_gdb "set target-charset\n"
178 -re "Requires an argument. Valid arguments are (\[^ \t\n\r,.\]*)" {
179 set target_charset $expect_out(1,string)
180 if {! [info exists charsets($target_charset)]} {
181 set charsets($target_charset) 0
186 -re ", (\[^ \t\n\r,.\]*)" {
187 set target_charset $expect_out(1,string)
188 if {! [info exists charsets($target_charset)]} {
189 set charsets($target_charset) 0
194 -re "\\.\r\n$gdb_prompt $" {
195 pass "capture valid target charsets"
199 -re ".*$gdb_prompt $" {
200 fail "capture valid target charsets"
204 fail "(timeout) capture valid target charsets"
208 # We don't want to test all the charset names here, since that would
209 # be too many combinations. We we pick a subset.
210 set charset_subset {ASCII ISO-8859-1 EBCDIC-US IBM1047}
211 foreach host_charset $charset_subset {
212 if {[valid_host_charset $host_charset]} {
214 set testname "try `set host-charset $host_charset'"
215 send_gdb "set host-charset $host_charset\n"
217 -re "GDB doesn't know of any character set named.*\[\r\n]+${gdb_prompt} $" {
218 # How did it get into `charsets' then?
219 fail "$testname (didn't recognize name)"
221 -re "GDB can't use `.*' as its host character set\\.\[\r\n]+${gdb_prompt} $" {
222 # Well, then why does its `charsets' entry say it can?
225 -re "${gdb_prompt} $" {
229 fail "$testname (timeout)"
233 # Check that the command actually had its intended effect:
234 # $host_charset should now be the host character set.
235 send_gdb "show charset\n"
236 set result [parse_show_charset_output "parse `show charset' after `set host-charset $host_charset'"]
237 if {! [string compare [lindex $result 0] $host_charset]} {
238 pass "check effect of `set host-charset $host_charset'"
240 fail "check effect of `set host-charset $host_charset'"
243 # Now try setting every possible target character set,
244 # given that host charset.
245 foreach target_charset $charset_subset {
246 if {![valid_target_charset $target_charset]} {
249 set testname "try `set target-charset $target_charset'"
250 send_gdb "set target-charset $target_charset\n"
252 -re "GDB doesn't know of any character set named.*\[\r\n]+${gdb_prompt} $" {
253 fail "$testname (didn't recognize name)"
255 -re "GDB can't convert from the .* character set to .*\\.\[\r\n\]+${gdb_prompt} $" {
256 # This is a serious problem. GDB should be able to convert
257 # between any arbitrary pair of character sets.
258 fail "$testname (can't convert)"
260 -re "${gdb_prompt} $" {
264 fail "$testname (timeout)"
268 # Check that the command actually had its intended effect:
269 # $target_charset should now be the target charset.
270 send_gdb "show charset\n"
271 set result [parse_show_charset_output "parse `show charset' after `set target-charset $target_charset'"]
272 if {! [string compare $result [list $host_charset $target_charset]]} {
273 pass "check effect of `set target-charset $target_charset'"
275 fail "check effect of `set target-charset $target_charset'"
278 # Test handling of characters in the host charset which
279 # can't be translated into the target charset. \xA2 is
280 # `cent' in ISO-8859-1, which has no equivalent in ASCII.
282 # On some systems, the pseudo-tty through which we
283 # communicate with GDB insists on stripping the high bit
284 # from input characters, meaning that `cent' turns into
285 # `"'. Since ISO-8859-1 and ASCII are identical in the
286 # lower 128 characters, it's tough to see how we can test
287 # this behavior on such systems, so we just xfail it.
289 # Note: the \x16 (Control-V) is an escape to allow \xA2 to
291 if {! [string compare $host_charset iso-8859-1] && ! [string compare $target_charset ascii]} {
293 set testname "untranslatable character in character literal"
294 send_gdb "print '\x16\xA2'\n"
296 -re "There is no character corresponding to .* in the target character set .*\\.\[\r\n\]+$gdb_prompt $" {
299 -re " = 34 '\"'\[\r\n\]+$gdb_prompt $" {
300 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
302 -re "$gdb_prompt $" {
306 fail "$testname (timeout)"
310 set testname "untranslatable character in string literal"
311 # If the PTTY zeros bit seven, then this turns into
313 # which gets us a syntax error. We don't care.
314 send_gdb "print \"\x16\xA2\"\n"
316 -re "There is no character corresponding to .* in the target character set .*\\.\[\r\n\]+$gdb_prompt $" {
319 -re "Unterminated string in expression.\[\r\n\]+$gdb_prompt $" {
320 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
322 -re "$gdb_prompt $" {
326 fail "$testname (timeout)"
330 set testname "untranslatable characters in backslash escape"
331 send_gdb "print '\\\x16\xA2'\n"
333 -re "The escape sequence .* is equivalent to plain .*, which has no equivalent\[\r\n\]+in the .* character set\\.\[\r\n\]+$gdb_prompt $" {
336 -re " = 34 '\"'\[\r\n\]+$gdb_prompt $" {
337 xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
339 -re "$gdb_prompt $" {
343 fail "$testname (timeout)"
352 # Set the host character set to plain ASCII, and try actually printing
353 # some strings in various target character sets. We need to run the
354 # test program to the point at which the strings have been
356 gdb_test "break ${srcfile}:[gdb_get_line_number "all strings initialized"]" \
357 ".*Breakpoint.* at .*" \
358 "set breakpoint after all strings have been initialized"
361 -re "Breakpoint.*all strings initialized.*$gdb_prompt $" {
362 pass "run until all strings have been initialized"
364 -re "$gdb_prompt $" {
365 fail "run until all strings have been initialized"
368 fail "run until all strings have been initialized (timeout)"
373 # We only try the wide character tests on machines where the wchar_t
374 # typedef in the test case has the right size.
375 set wchar_size [get_sizeof wchar_t 99]
377 if {$wchar_size == 2} {
378 lappend charset_subset UCS-2
380 } elseif {$wchar_size == 4} {
381 lappend charset_subset UCS-4
385 gdb_test "set host-charset ASCII" ""
386 foreach target_charset $charset_subset {
387 if {![valid_target_charset $target_charset]} {
391 if {$target_charset == "UCS-4" || $target_charset == "UCS-2"} {
392 set param target-wide-charset
395 set param target-charset
398 send_gdb "set $param $target_charset\n"
400 -re "$gdb_prompt $" {
401 pass "set $param $target_charset"
404 fail "set $param $target_charset (timeout)"
408 # Try printing the null character. There seems to be a bug in
409 # gdb_test that requires us to use gdb_expect here.
410 send_gdb "print $L'\\0'\n"
412 -re "\\\$${decimal} = 0 $L'\\\\0'\[\r\n\]+$gdb_prompt $" {
413 pass "print the null character in ${target_charset}"
415 -re "$gdb_prompt $" {
416 fail "print the null character in ${target_charset}"
419 fail "print the null character in ${target_charset} (timeout)"
423 # Compute the name of the variable in the test program that holds
424 # a string in $target_charset. The variable's name is the
425 # character set's name, in lower-case, with all non-identifier
426 # characters replaced with '_', with "_string" stuck on the end.
427 if {$target_charset == "UCS-2"} {
428 # We still use the ucs_4_string variable -- but the size is
430 set var_name ucs_4_string
432 set var_name [string tolower "${target_charset}_string"]
433 regsub -all -- "\[^a-z0-9_\]" $var_name "_" var_name
436 # Compute a regexp matching the results we expect. This is static,
437 # but it's easier than writing it out.
438 regsub -all "." "abfnrtv" "(\\\\&|x)" escapes
439 set uppercase "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
440 set lowercase "abcdefghijklmnopqrstuvwxyz"
441 set digits "0123456789"
442 set octal_escape "\\\\\[0-9\]+"
444 send_gdb "print $var_name\n"
445 # ${escapes}${uppercase}${lowercase}${digits}${octal}${octal}
447 -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 $" {
448 pass "print string in $target_charset"
450 -re "$gdb_prompt $" {
451 fail "print string in $target_charset"
454 fail "print string in $target_charset (timeout)"
458 # Try entering a character literal, and see if it comes back unchanged.
459 gdb_test "print $L'A'" \
460 " = \[0-9-\]+ $L'A'" \
461 "parse character literal in ${target_charset}"
463 # Check that the character literal was encoded correctly.
464 gdb_test "print $L'A' == $var_name\[7\]" \
466 "check value of parsed character literal in ${target_charset}"
468 # Try entering a string literal, and see if it comes back unchanged.
469 gdb_test "print $L\"abcdefABCDEF012345\"" \
470 " = $L\"abcdefABCDEF012345\"" \
471 "parse string literal in ${target_charset}"
473 # Check that the string literal was encoded correctly.
474 gdb_test "print $L\"q\"\[0\] == $var_name\[49\]" \
476 "check value of parsed string literal in ${target_charset}"
478 # Test handling of characters in the target charset which
479 # can't be translated into the host charset.
480 if {! [string compare $target_charset iso-8859-1]} {
481 gdb_test "print iso_8859_1_string\[69\]" \
482 " = \[0-9-\]+ '\\\\242'" \
483 "print character with no equivalent in host character set"
484 gdb_test "print iso_8859_1_string + 70" \
485 " = ${hex} \"\\\\242.*\"" \
486 "print string with no equivalent in host character set"
489 # Make sure that we don't apply the ISO-8859-1 `print_literally'
491 if {! [string compare $target_charset ascii]} {
492 gdb_test "print iso_8859_1_string\[69\]" \
493 " = \[0-9-\]+ '\\\\242'" \
494 "print ASCII unprintable character"
495 gdb_test "print iso_8859_1_string + 70" \
496 " = ${hex} \"\\\\242.*\"" \
497 "print ASCII unprintable string"
500 # Try printing characters with backslash escape equivalents.
501 set escapees {a b f n r t v}
502 for {set i 0} {$i < [llength $escapees]} {incr i} {
503 set escape [lindex $escapees $i]
504 send_gdb "print $var_name\[$i\]\n"
507 -re "= \[0-9-\]+ $L'\\\\${escape}'\[\r\n\]+$gdb_prompt $" {
508 pass "try printing '\\${escape}' in ${target_charset}"
510 -re "= \[0-9-\]+ 'x'\[\r\n\]+$gdb_prompt $" {
511 xfail "try printing '\\${escape}' in ${target_charset} (no such escape)"
514 -re "$gdb_prompt $" {
515 fail "try printing '\\${escape}' in ${target_charset}"
518 fail "try printing '\\${escape}' in ${target_charset} (timeout)"
524 # Try parsing a backslash escape in a character literal.
525 gdb_test "print $L'\\${escape}' == $var_name\[$i\]" \
527 "check value of '\\${escape}' in ${target_charset}"
529 # Try parsing a backslash escape in a string literal.
530 gdb_test "print $L\"\\${escape}\"\[0\] == $var_name\[$i\]" \
532 "check value of \"\\${escape}\" in ${target_charset}"
536 # Try printing a character escape that doesn't exist. We should
537 # get the unescaped character, in the target character set.
538 gdb_test "print $L'\\q'" " = \[0-9-\]+ $L'q'" \
539 "print escape that doesn't exist in $target_charset"
540 gdb_test "print $L'\\q' == $var_name\[49\]" " = 1" \
541 "check value of escape that doesn't exist in $target_charset"
544 # Reset the target charset.
545 gdb_test "set target-charset UTF-8" ""
547 # \242 is not a valid UTF-8 character.
548 gdb_test "print \"\\242\"" " = \"\\\\242\"" \
549 "non-representable target character"
551 gdb_test "print '\\x'" "\\\\x escape without a following hex digit."
552 gdb_test "print '\\u'" "\\\\u escape without a following hex digit."
553 gdb_test "print '\\9'" " = \[0-9\]+ '9'"
555 # Tests for wide- or unicode- strings. L is the prefix letter to use,
556 # either "L" (for wide strings), "u" (for UCS-2), or "U" (for UCS-4).
557 # NAME is used in the test names and should be related to the prefix
558 # letter in some easy-to-undestand way.
559 proc test_wide_or_unicode {L name} {
560 gdb_test "print $L\"ab\" $L\"c\"" " = $L\"abc\"" \
561 "basic $name string concatenation"
562 gdb_test "print $L\"ab\" \"c\"" " = $L\"abc\"" \
563 "narrow and $name string concatenation"
564 gdb_test "print \"ab\" $L\"c\"" " = $L\"abc\"" \
565 "$name and narrow string concatenation"
566 gdb_test "print $L\"\\xe\" $L\"c\"" " = $L\"\\\\16c\"" \
567 "$name string concatenation with escape"
568 gdb_test "print $L\"\" \"abcdef\" \"g\"" \
570 "concatenate three strings with empty $name string"
572 gdb_test "print $L'a'" "= \[0-9\]+ $L'a'" \
573 "basic $name character"
577 test_wide_or_unicode L wide
580 set ucs2_ok [expr {[get_sizeof char16_t 99] == 2}]
582 test_wide_or_unicode u UCS-2
585 set ucs4_ok [expr {[get_sizeof char32_t 99] == 4}]
587 test_wide_or_unicode U UCS-4
590 # Test an invalid string combination.
591 proc test_combination {L1 name1 L2 name2} {
592 gdb_test "print $L1\"abc\" $L2\"def\"" \
593 "Undefined string concatenation." \
594 "undefined concatenation of $name1 and $name2"
597 if {$wchar_ok && $ucs2_ok} {
598 test_combination L wide u UCS-2
600 if {$wchar_ok && $ucs4_ok} {
601 test_combination L wide U UCS-4
603 if {$ucs2_ok && $ucs4_ok} {
604 test_combination u UCS-2 U UCS-4