gdb/testsuite/gdb.base/charset.exp

   1 # This testcase is part of GDB, the GNU debugger.
   2
   3 # Copyright 2001, 2004, 2007 Free Software Foundation, Inc.
   4
   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 2 of the License, or
   8 # (at your option) any later version.
   9 #
  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.
  14 #
  15 # You should have received a copy of the GNU General Public License
  16 # along with this program; if not, write to the Free Software
  17 # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  18
  19 # Please email any bugs, comments, and/or additions to this file to:
  20 # bug-gdb@gnu.org
  21
  22 # Test GDB's character set support.
  23
  24 if $tracelevel then {
  25         strace $tracelevel
  26 }
  27
  28 set prms_id 0
  29 set bug_id 0
  30
  31 set testfile "charset"
  32 set srcfile ${testfile}.c
  33 set binfile ${objdir}/${subdir}/${testfile}
  34 if  { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" executable {debug}] != "" } {
  35     untested "couldn't compile ${srcdir}/${subdir}/${srcfile}"
  36     return -1
  37 }
  38
  39 # Start with a fresh gdb.
  40 gdb_exit
  41 gdb_start
  42 gdb_reinitialize_dir $srcdir/$subdir
  43 gdb_load ${binfile}
  44
  45 # Parse the output from a `show charset' command.  Return the host
  46 # and target charset as a two-element list.
  47 proc parse_show_charset_output {testname} {
  48     global gdb_prompt
  49
  50     gdb_expect {
  51         -re "The current host and target character set is `(.*)'\\.\[\r\n\]+$gdb_prompt $" {
  52             set host_charset $expect_out(1,string)
  53             set target_charset $expect_out(1,string)
  54             set retlist [list $host_charset $target_charset]
  55             pass $testname
  56         }
  57         -re "The current host character set is `(.*)'\\.\[\r\n\]+The current target character set is `(.*)'\\.\[\r\n\]+$gdb_prompt $" {
  58             set host_charset $expect_out(1,string)
  59             set target_charset $expect_out(2,string)
  60             set retlist [list $host_charset $target_charset]
  61             pass $testname
  62         }
  63         -re "The host character set is \"(.*)\"\\.\[\r\n\]+$gdb_prompt $" {
  64             set host_charset $expect_out(1,string)
  65             set retlist [list $host_charset]
  66             pass $testname
  67         }
  68         -re "The target character set is \"(.*)\"\\.\[\r\n\]+$gdb_prompt $" {
  69             set target_charset $expect_out(1,string)
  70             set retlist [list $target_charset]
  71             pass $testname
  72         }
  73         -re ".*$gdb_prompt $" {
  74             fail $testname
  75         }
  76         timeout {
  77             fail "$testname (timeout)"
  78         }
  79     }
  80
  81     return $retlist
  82 }
  83
  84
  85 # Try the various `show charset' commands.  These are all aliases of each
  86 # other; `show target-charset' and `show host-charset' actually print
  87 # both the host and target charsets.
  88
  89 send_gdb "show charset\n"
  90 set show_charset [parse_show_charset_output "show charset"]
  91
  92 send_gdb "show target-charset\n"
  93 set show_target_charset [parse_show_charset_output "show target-charset"]
  94
  95 if {[lsearch $show_charset $show_target_charset] >= 0} {
  96     pass "check `show target-charset' against `show charset'"
  97 } else {
  98     fail "check `show target-charset' against `show charset'"
  99 }
 100
 101 send_gdb "show host-charset\n"
 102 set show_host_charset [parse_show_charset_output "show host-charset"]
 103
 104 if {[lsearch $show_charset $show_host_charset] >= 0} {
 105     pass "check `show host-charset' against `show charset'"
 106 } else {
 107     fail "check `show host-charset' against `show charset'"
 108 }
 109
 110
 111 # Get the list of supported (host) charsets as possible completions.
 112 send_gdb "set charset \t\t"
 113
 114 # Check that we can at least use ASCII as a host character set.
 115 sleep 1
 116 gdb_expect {
 117     -re "^set charset .*\r\nASCII.*\r\n$gdb_prompt set charset " {
 118         # We got the output that we wanted, including ASCII as possible
 119         # charset. Send a newline to get us back to the prompt. This will
 120         # also generate an error message. Let's not check here that the error
 121         # message makes sense, we do that below, as a separate testcase.
 122         send_gdb "\n"
 123         gdb_expect {
 124             -re ".*Requires an argument.*$gdb_prompt $" {
 125                 pass "get valid character sets"
 126             }
 127             -re ".*$gdb_prompt $" {
 128                 send_gdb "\n"
 129                 gdb_expect {
 130                     -re ".*$gdb_prompt $" {
 131                         fail "get valid character sets"
 132                     }
 133                 }
 134             }
 135             timeout {
 136                 fail "(timeout) get valid character sets"
 137             }
 138         }
 139     }
 140     -re ".*$gdb_prompt $" {
 141         # We got some output that ended with a regular prompt
 142         fail "get valid character sets"
 143     }
 144     -re ".*$gdb_prompt set charset.*$" {
 145         # We got some other output, send a cntrl-c to gdb to get us back
 146         # to the prompt.
 147         send_gdb "\003"
 148         fail "get valid character sets"
 149     }
 150     timeout {
 151         fail "get valid character sets (timeout)"
 152     }
 153 }
 154
 155 # Try a malformed `set charset'.
 156 gdb_test "set charset" \
 157          "Requires an argument. Valid arguments are.*" \
 158          "try malformed `set charset'"
 159
 160 # Try using `set host-charset' on an invalid character set.
 161 gdb_test "set host-charset my_grandma_bonnie" \
 162          "Undefined item: \"my_grandma_bonnie\"." \
 163          "try `set host-charset' with invalid charset"
 164
 165 # Try using `set target-charset' on an invalid character set.
 166 gdb_test "set target-charset my_grandma_bonnie" \
 167          "Undefined item: \"my_grandma_bonnie\"." \
 168          "try `set target-charset' with invalid charset"
 169
 170 # A Tcl array mapping the names of all the character sets we've seen
 171 # to "1" if the character set can be used as a host character set, or
 172 # "0" otherwise.  We can use `array names charsets' just to get a list
 173 # of all character sets.
 174 array set charsets {}
 175
 176 proc all_charset_names {} {
 177     global charsets
 178     return [array names charsets]
 179 }
 180
 181 proc valid_host_charset {charset} {
 182     global charsets
 183     return $charsets($charset)
 184 }
 185
 186 send_gdb "set host-charset\n"
 187 gdb_expect {
 188     -re "Requires an argument. Valid arguments are (\[^ \t\n\r,.\]*)" {
 189         #set host_charset_list $expect_out(1,string)
 190         set charsets($expect_out(1,string)) 1
 191         exp_continue
 192         #pass "capture valid host charsets"
 193     }
 194
 195     -re ", (\[^ \t\n\r,.\]*)" {
 196         #set host_charset_list $expect_out(1,string)
 197         set charsets($expect_out(1,string)) 1
 198         exp_continue
 199         #pass "capture valid host charsets"
 200     }
 201
 202     -re "\\.\r\n$gdb_prompt $" {
 203         #set host_charset_list $expect_out(1,string)
 204         pass "capture valid host charsets"
 205     }
 206
 207     -re ".*$gdb_prompt $" {
 208         fail "capture valid host charsets"
 209     }
 210     timeout {
 211         fail "(timeout) capture valid host charsets"
 212     }
 213 }
 214
 215
 216 send_gdb "set target-charset\n"
 217 gdb_expect {
 218     -re "Requires an argument. Valid arguments are (\[^ \t\n\r,.\]*)" {
 219         set target_charset $expect_out(1,string)
 220         if {! [info exists charsets($target_charset)]} {
 221             set charsets($target_charset) 0
 222         }
 223         exp_continue
 224     }
 225
 226     -re ", (\[^ \t\n\r,.\]*)" {
 227         set target_charset $expect_out(1,string)
 228         if {! [info exists charsets($target_charset)]} {
 229             set charsets($target_charset) 0
 230         }
 231         exp_continue
 232     }
 233
 234     -re "\\.\r\n$gdb_prompt $" {
 235         pass "capture valid target charsets"
 236
 237     }
 238
 239     -re ".*$gdb_prompt $" {
 240         fail "capture valid target charsets"
 241     }
 242
 243     timeout {
 244         fail "(timeout) capture valid target charsets"
 245     }
 246 }
 247
 248 # Make sure that GDB supports every host/target charset combination.
 249 foreach host_charset [all_charset_names] {
 250     if {[valid_host_charset $host_charset]} {
 251
 252         set testname "try `set host-charset $host_charset'"
 253         send_gdb "set host-charset $host_charset\n"
 254         gdb_expect {
 255             -re "GDB doesn't know of any character set named.*\[\r\n]+${gdb_prompt} $" {
 256                 # How did it get into `charsets' then?
 257                 fail "$testname (didn't recognize name)"
 258             }
 259             -re "GDB can't use `.*' as its host character set\\.\[\r\n]+${gdb_prompt} $" {
 260                 # Well, then why does its `charsets' entry say it can?
 261                 fail $testname
 262             }
 263             -re "${gdb_prompt} $" {
 264                 pass $testname
 265             }
 266             timeout {
 267                 fail "$testname (timeout)"
 268             }
 269         }
 270
 271         # Check that the command actually had its intended effect:
 272         # $host_charset should now be the host character set.
 273         send_gdb "show charset\n"
 274         set result [parse_show_charset_output "parse `show charset' after `set host-charset $host_charset'"]
 275         if {! [string compare [lindex $result 0] $host_charset]} {
 276             pass "check effect of `set host-charset $host_charset'"
 277         } else {
 278             fail "check effect of `set host-charset $host_charset'"
 279         }
 280
 281         # Now try setting every possible target character set,
 282         # given that host charset.
 283         foreach target_charset [all_charset_names] {
 284             set testname "try `set target-charset $target_charset'"
 285             send_gdb "set target-charset $target_charset\n"
 286             gdb_expect {
 287                 -re "GDB doesn't know of any character set named.*\[\r\n]+${gdb_prompt} $" {
 288                     fail "$testname (didn't recognize name)"
 289                 }
 290                 -re "GDB can't convert from the .* character set to .*\\.\[\r\n\]+${gdb_prompt} $" {
 291                     # This is a serious problem.  GDB should be able to convert
 292                     # between any arbitrary pair of character sets.
 293                     fail "$testname (can't convert)"
 294                 }
 295                 -re "${gdb_prompt} $" {
 296                     pass $testname
 297                 }
 298                 timeout {
 299                     fail "$testname (timeout)"
 300                 }
 301             }
 302
 303             # Check that the command actually had its intended effect:
 304             # $target_charset should now be the target charset.
 305             send_gdb "show charset\n"
 306             set result [parse_show_charset_output "parse `show charset' after `set target-charset $target_charset'"]
 307             if {! [string compare $result [list $host_charset $target_charset]]} {
 308                 pass "check effect of `set target-charset $target_charset'"
 309             } else {
 310                 fail "check effect of `set target-charset $target_charset'"
 311             }
 312
 313             # Test handling of characters in the host charset which
 314             # can't be translated into the target charset.  \xA2 is
 315             # `cent' in ISO-8859-1, which has no equivalent in ASCII.
 316             #
 317             # On some systems, the pseudo-tty through which we
 318             # communicate with GDB insists on stripping the high bit
 319             # from input characters, meaning that `cent' turns into
 320             # `"'.  Since ISO-8859-1 and ASCII are identical in the
 321             # lower 128 characters, it's tough to see how we can test
 322             # this behavior on such systems, so we just xfail it.
 323             #
 324             # Note: the \x16 (Control-V) is an escape to allow \xA2 to
 325             # get past readline.
 326             if {! [string compare $host_charset iso-8859-1] && ! [string compare $target_charset ascii]} {
 327
 328                 set testname "untranslatable character in character literal"
 329                 send_gdb "print '\x16\xA2'\n"
 330                 gdb_expect {
 331                     -re "There is no character corresponding to .* in the target character set .*\\.\[\r\n\]+$gdb_prompt $" {
 332                         pass $testname
 333                     }
 334                     -re " = 34 '\"'\[\r\n\]+$gdb_prompt $" {
 335                         xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
 336                     }
 337                     -re "$gdb_prompt $" {
 338                         fail $testname
 339                     }
 340                     timeout {
 341                         fail "$testname (timeout)"
 342                     }
 343                 }
 344
 345                 set testname "untranslatable character in string literal"
 346                 # If the PTTY zeros bit seven, then this turns into
 347                 #   print """
 348                 # which gets us a syntax error.  We don't care.
 349                 send_gdb "print \"\x16\xA2\"\n"
 350                 gdb_expect {
 351                     -re "There is no character corresponding to .* in the target character set .*\\.\[\r\n\]+$gdb_prompt $" {
 352                         pass $testname
 353                     }
 354                     -re "Unterminated string in expression.\[\r\n\]+$gdb_prompt $" {
 355                         xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
 356                     }
 357                     -re "$gdb_prompt $" {
 358                         fail $testname
 359                     }
 360                     timeout {
 361                         fail "$testname (timeout)"
 362                     }
 363                 }
 364
 365                 set testname "untranslatable characters in backslash escape"
 366                 send_gdb "print '\\\x16\xA2'\n"
 367                 gdb_expect {
 368                     -re "The escape sequence .* is equivalent to plain .*, which has no equivalent\[\r\n\]+in the .* character set\\.\[\r\n\]+$gdb_prompt $" {
 369                         pass $testname
 370                     }
 371                     -re " = 34 '\"'\[\r\n\]+$gdb_prompt $" {
 372                         xfail "$testname (DejaGNU's pseudo-tty strips eighth bit)"
 373                     }
 374                     -re "$gdb_prompt $" {
 375                         fail $testname
 376                     }
 377                     timeout {
 378                         fail "$testname (timeout)"
 379                     }
 380                 }
 381             }
 382         }
 383     }
 384 }
 385
 386
 387 # Set the host character set to plain ASCII, and try actually printing
 388 # some strings in various target character sets.  We need to run the
 389 # test program to the point at which the strings have been
 390 # initialized.
 391 gdb_test "break ${srcfile}:[gdb_get_line_number "all strings initialized"]" \
 392          ".*Breakpoint.* at .*" \
 393          "set breakpoint after all strings have been initialized"
 394 gdb_run_cmd
 395 gdb_expect {
 396     -re "Breakpoint.*all strings initialized.*$gdb_prompt $" {
 397         pass "run until all strings have been initialized"
 398     }
 399     -re "$gdb_prompt $" {
 400         fail "run until all strings have been initialized"
 401     }
 402     timeout {
 403         fail "run until all strings have been initialized (timeout)"
 404     }
 405 }
 406
 407
 408 gdb_test "set host-charset ASCII" ""
 409 foreach target_charset [all_charset_names] {
 410     send_gdb "set target-charset $target_charset\n"
 411     gdb_expect {
 412         -re "$gdb_prompt $" {
 413             pass "set target-charset $target_charset"
 414         }
 415         timeout {
 416             fail "set target-charset $target_charset (timeout)"
 417         }
 418     }
 419
 420     # Try printing the null character.  There seems to be a bug in
 421     # gdb_test that requires us to use gdb_expect here.
 422     send_gdb "print '\\0'\n"
 423     gdb_expect {
 424         -re "\\\$${decimal} = 0 '\\\\0'\[\r\n\]+$gdb_prompt $" {
 425             pass "print the null character in ${target_charset}"
 426         }
 427         -re "$gdb_prompt $" {
 428             fail "print the null character in ${target_charset}"
 429         }
 430         timeout {
 431             fail "print the null character in ${target_charset} (timeout)"
 432         }
 433     }
 434
 435     # Compute the name of the variable in the test program that holds
 436     # a string in $target_charset.  The variable's name is the
 437     # character set's name, in lower-case, with all non-identifier
 438     # characters replaced with '_', with "_string" stuck on the end.
 439     set var_name [string tolower "${target_charset}_string"]
 440     regsub -all -- "\[^a-z0-9_\]" $var_name "_" var_name
 441
 442     # Compute a regexp matching the results we expect.  This is static,
 443     # but it's easier than writing it out.
 444     regsub -all "." "abfnrtv" "(\\\\&|x)" escapes
 445     set uppercase "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
 446     set lowercase "abcdefghijklmnopqrstuvwxyz"
 447     set digits "0123456789"
 448     set octal_escape "\\\\\[0-9\]\[0-9\]\[0-9\]"
 449
 450     send_gdb "print $var_name\n"
 451     # ${escapes}${uppercase}${lowercase}${digits}${octal}${octal}
 452     gdb_expect {
 453         -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 $" {
 454             pass "print string in $target_charset"
 455         }
 456         -re "$gdb_prompt $" {
 457             fail "print string in $target_charset"
 458         }
 459         timeout {
 460             fail "print string in $target_charset (timeout)"
 461         }
 462     }
 463
 464     # Try entering a character literal, and see if it comes back unchanged.
 465     gdb_test "print 'A'" \
 466              " = \[0-9-\]+ 'A'" \
 467              "parse character literal in ${target_charset}"
 468
 469     # Check that the character literal was encoded correctly.
 470     gdb_test "print 'A' == $var_name\[7\]" \
 471              " = 1" \
 472              "check value of parsed character literal in ${target_charset}"
 473
 474     # Try entering a string literal, and see if it comes back unchanged.
 475     gdb_test "print \"abcdefABCDEF012345\"" \
 476              " = \"abcdefABCDEF012345\"" \
 477              "parse string literal in ${target_charset}"
 478
 479     # Check that the string literal was encoded correctly.
 480     gdb_test "print \"q\"\[0\] == $var_name\[49\]" \
 481              " = 1" \
 482              "check value of parsed string literal in ${target_charset}"
 483
 484     # Test handling of characters in the target charset which
 485     # can't be translated into the host charset.
 486     if {! [string compare $target_charset iso-8859-1]} {
 487         gdb_test "print iso_8859_1_string\[69\]" \
 488                  " = \[0-9-\]+ '\\\\242'" \
 489                  "print character with no equivalent in host character set"
 490         gdb_test "print iso_8859_1_string + 70" \
 491                  " = ${hex} \"\\\\242.*\"" \
 492                  "print string with no equivalent in host character set"
 493     }
 494
 495     # Make sure that we don't apply the ISO-8859-1 `print_literally'
 496     # function to ASCII.
 497     if {! [string compare $target_charset ascii]} {
 498         gdb_test "print iso_8859_1_string\[69\]" \
 499                  " = \[0-9-\]+ '\\\\242'" \
 500                  "print ASCII unprintable character"
 501         gdb_test "print iso_8859_1_string + 70" \
 502                  " = ${hex} \"\\\\242.*\"" \
 503                  "print ASCII unprintable string"
 504     }
 505
 506     # Try printing characters with backslash escape equivalents.
 507     set escapees {a b f n r t v}
 508     for {set i 0} {$i < [llength $escapees]} {incr i} {
 509         set escape [lindex $escapees $i]
 510         send_gdb "print $var_name\[$i\]\n"
 511         set have_escape 1
 512         gdb_expect {
 513             -re "= \[0-9-\]+ '\\\\${escape}'\[\r\n\]+$gdb_prompt $" {
 514                 pass "try printing '\\${escape}' in ${target_charset}"
 515             }
 516             -re "= \[0-9-\]+ 'x'\[\r\n\]+$gdb_prompt $" {
 517                 xfail "try printing '\\${escape}' in ${target_charset} (no such escape)"
 518                 set have_escape 0
 519             }
 520             -re "$gdb_prompt $" {
 521                 fail "try printing '\\${escape}' in ${target_charset}"
 522             }
 523             timeout {
 524                 fail "try printing '\\${escape}' in ${target_charset} (timeout)"
 525             }
 526         }
 527
 528         if {$have_escape} {
 529
 530             # Try parsing a backslash escape in a character literal.
 531             gdb_test "print '\\${escape}' == $var_name\[$i\]" \
 532                      " = 1" \
 533                      "check value of '\\${escape}' in ${target_charset}"
 534
 535             # Try parsing a backslash escape in a string literal.
 536             gdb_test "print \"\\${escape}\"\[0\] == $var_name\[$i\]" \
 537                      " = 1" \
 538                      "check value of \"\\${escape}\" in ${target_charset}"
 539         }
 540     }
 541
 542     # Try printing a character escape that doesn't exist.  We should
 543     # get the unescaped character, in the target character set.
 544     gdb_test "print '\\q'" " = \[0-9-\]+ 'q'" \
 545              "print escape that doesn't exist in $target_charset"
 546     gdb_test "print '\\q' == $var_name\[49\]" " = 1" \
 547              "check value of escape that doesn't exist in $target_charset"
 548 }
 549
 550 gdb_exit