path: root/gdb/testsuite/gdb.guile/scm-value.exp

# Copyright (C) 2008-2018 Free Software Foundation, Inc.

# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.

# This file is part of the GDB testsuite.
# It tests the mechanism exposing values to Guile.

load_lib gdb-guile.exp

standard_testfile

set has_argv0 [gdb_has_argv0]

# Build inferior to language specification.
# LANG is one of "c" or "c++".
proc build_inferior {exefile lang} {
    global srcdir subdir srcfile testfile hex

    # Use different names for .o files based on the language.
    # For Fission, the debug info goes in foo.dwo and we don't want,
    # for example, a C++ compile to clobber the dwo of a C compile.
    # ref: http://gcc.gnu.org/wiki/DebugFission
    switch ${lang} {
	"c" { set filename ${testfile}.o }
	"c++" { set filename ${testfile}-cxx.o }
    }
    set objfile [standard_output_file $filename]

    if { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${objfile}" object "debug $lang"] != ""
	 || [gdb_compile "${objfile}" "${exefile}" executable "debug $lang"] != "" } {
	untested "failed to compile in $lang mode"
	return -1
    }
    return 0
}

proc test_value_in_inferior {} {
    global gdb_prompt
    global testfile

    gdb_breakpoint [gdb_get_line_number "break to inspect struct and union"]

    gdb_continue_to_breakpoint "break to inspect struct and union"

    # Just get inferior variable s in the value history, available to guile.
    gdb_test "print s" "= {a = 3, b = 5}" ""

    gdb_scm_test_silent_cmd "gu (define s (history-ref 0))" "set s"

    gdb_test "gu (print (value-field s \"a\"))" \
	"= 3" "access element inside struct using string name"

    # Append value in the value history.
    gdb_scm_test_silent_cmd "gu (define i (history-append! (make-value 42)))" \
	"append 42"

    gdb_test "gu i" "\[0-9\]+"
    gdb_test "gu (history-ref i)" "#<gdb:value 42>"
    gdb_test "p \$" "= 42"

    # Verify the recorded history value survives a gc.
    gdb_test_no_output "guile (gc)"
    gdb_test "p \$\$" "= 42"

    # Make sure 'history-append!' rejects non-value objects.
    gdb_test "gu (history-append! 123)" \
	"ERROR:.* Wrong type argument.*" "history-append! type error"

    # Test dereferencing the argv pointer.

    # Just get inferior variable argv the value history, available to guile.
    gdb_test "print argv" "= \\(char \\*\\*\\) 0x.*" ""

    gdb_scm_test_silent_cmd "gu (define argv (history-ref 0))" \
	"set argv"
    gdb_scm_test_silent_cmd "gu (define arg0 (value-dereference argv))" \
	"set arg0"

    # Check that the dereferenced value is sane.
    global has_argv0
    set test "verify dereferenced value"
    if { $has_argv0 } {
	gdb_test_no_output "set print elements unlimited" ""
	gdb_test_no_output "set print repeats unlimited" ""
	gdb_test "gu (print arg0)" "0x.*$testfile\"" $test
    } else {
	unsupported $test
    }

    # Smoke-test value-optimized-out?.
    gdb_test "gu (print (value-optimized-out? arg0))" \
	"= #f" "Test value-optimized-out?"

    # Test address attribute.
    gdb_test "gu (print (value-address arg0))" \
	"= 0x\[\[:xdigit:\]\]+" "Test address attribute"
    # Test address attribute is #f in a non-addressable value.
    gdb_test "gu (print (value-address (make-value 42)))" \
	"= #f" "Test address attribute in non-addressable value"

    # Test displaying a variable that is temporarily at a bad address.
    # But if we can examine what's at memory address 0, then we'll also be
    # able to display it without error.  Don't run the test in that case.
    set can_read_0 [is_address_zero_readable]

    # Test memory error.
    set test "parse_and_eval with memory error"
    if {$can_read_0} {
	untested $test
    } else {
	gdb_test "gu (print (parse-and-eval \"*(int*)0\"))" \
	    "ERROR: Cannot access memory at address 0x0.*" $test
    }

    # Test Guile lazy value handling
    set test "memory error and lazy values"
    if {$can_read_0} {
	untested $test
    } else {
	gdb_test_no_output "gu (define inval (parse-and-eval \"*(int*)0\"))"
	gdb_test "gu (print (value-lazy? inval))" \
	    "#t"
	gdb_test "gu (define inval2 (value-add inval 1))" \
	    "ERROR: Cannot access memory at address 0x0.*" $test
	gdb_test "gu (value-fetch-lazy! inval))" \
	    "ERROR: Cannot access memory at address 0x0.*" $test
    }
    gdb_test_no_output "gu (define argc-lazy (parse-and-eval \"argc\"))"
    gdb_test_no_output "gu (define argc-notlazy (parse-and-eval \"argc\"))"
    gdb_test_no_output "gu (value-fetch-lazy! argc-notlazy)"
    gdb_test "gu (print (value-lazy? argc-lazy))" "= #t"
    gdb_test "gu (print (value-lazy? argc-notlazy))" "= #f"
    gdb_test "print argc" "= 1" "sanity check argc"
    gdb_test "gu (print (value-lazy? argc-lazy))" "= #t"
    gdb_test_no_output "set argc=2"
    gdb_test "gu (print argc-notlazy)" "= 1"
    gdb_test "gu (print argc-lazy)" "= 2"
    gdb_test "gu (print (value-lazy? argc-lazy))" "= #f"

    # Test string fetches, both partial and whole.
    gdb_test "print st" "\"divide et impera\""
    gdb_scm_test_silent_cmd "gu (define st (history-ref 0))" \
	"inf: get st value from history"
    gdb_test "gu (print (value->string st))" \
	"= divide et impera"  "Test string with no length"
    gdb_test "gu (print (value->string st #:length -1))" \
	"= divide et impera" "Test string (length = -1) is all of the string"
    gdb_test "gu (print (value->string st #:length 6))" \
	"= divide"
    gdb_test "gu (print (string-append \"---\" (value->string st #:length 0) \"---\"))" \
	"= ------" "Test string (length = 0) is empty"
    gdb_test "gu (print (string-length (value->string st #:length 0)))" \
	"= 0" "Test length is 0"

    # Fetch a string that has embedded nulls.
    gdb_test "print nullst" "\"divide\\\\000et\\\\000impera\".*"
    gdb_scm_test_silent_cmd "gu (define nullst (history-ref 0))" \
	"inf: get nullst value from history"
    gdb_test "gu (print (value->string nullst))" \
	"divide" "Test string to first null"
    gdb_scm_test_silent_cmd "gu (set! nullst (value->string nullst #:length 9))" \
	"get string beyond null"
    gdb_test "gu (print nullst)" \
	"= divide\\\\000et"
}

proc test_strings {} {
    gdb_test "gu (make-value \"test\")" "#<gdb:value \"test\">" "make string"

    # Test string conversion errors.
    set save_charset [get_target_charset]
    gdb_test_no_output "set target-charset UTF-8"

    gdb_test_no_output "gu (set-port-conversion-strategy! #f 'error)"
    gdb_test "gu (print (value->string (make-value (string #\\x1234)) #:encoding \"ASCII\"))" \
	"ERROR.*decoding-error.*" \
	"value->string with default #:errors = 'error"

    # There is no 'escape strategy for C->SCM string conversions, but it's
    # still a legitimate value for %default-port-conversion-strategy.
    # GDB handles this by, umm, substituting 'substitute.
    # Use this case to also handle "#:errors #f" which explicitly says
    # "use %default-port-conversion-strategy".
    gdb_test_no_output "gu (set-port-conversion-strategy! #f 'escape)"
    gdb_test "gu (print (value->string (make-value (string #\\x1234)) #:encoding \"ASCII\" #:errors #f))" \
	"= \[?\]{3}" "value->string with default #:errors = 'escape"

    # This is last in the default conversion tests so that
    # %default-port-conversion-strategy ends up with the default value.
    gdb_test_no_output "gu (set-port-conversion-strategy! #f 'substitute)"
    gdb_test "gu (print (value->string (make-value (string #\\x1234)) #:encoding \"ASCII\"))" \
	"= \[?\]{3}" "value->string with default #:errors = 'substitute"

    gdb_test "gu (print (value->string (make-value (string #\\x1234)) #:encoding \"ASCII\" #:errors 'error))" \
	"ERROR.*decoding-error.*" "value->string #:errors 'error"
    gdb_test "gu (print (value->string (make-value (string #\\x1234)) #:encoding \"ASCII\" #:errors 'substitute))" \
	"= \[?\]{3}" "value->string #:errors 'substitute"
    gdb_test "gu (print (value->string (make-value \"abc\") #:errors \"foo\"))" \
	"ERROR.*invalid error kind.*" "bad value for #:errors"

    gdb_test_no_output "set target-charset $save_charset" \
	"restore target-charset"
}

proc test_inferior_function_call {} {
    global gdb_prompt hex decimal

    # Correct inferior call without arguments.
    gdb_test "p/x fp1" "= $hex.*"
    gdb_scm_test_silent_cmd "gu (define fp1 (history-ref 0))" \
	"get fp1 value from history"
    gdb_scm_test_silent_cmd "gu (set! fp1 (value-dereference fp1))" \
	"dereference fp1"
    gdb_test "gu (print (value-call fp1 '()))" \
	"= void"

    # Correct inferior call with arguments.
    gdb_test "p/x fp2" "= $hex.*"
    gdb_scm_test_silent_cmd "gu (define fp2 (history-ref 0))" \
	"get fp2 value from history"
    gdb_scm_test_silent_cmd "gu (set! fp2 (value-dereference fp2))" \
	"dereference fp2"
    gdb_test "gu (print (value-call fp2 (list 10 20)))" \
	"= 30"

    # Incorrect to call an int value.
    gdb_test "p i" "= $decimal.*"
    gdb_scm_test_silent_cmd "gu (define i (history-ref 0))" \
	"inf call: get i value from history"
    gdb_test "gu (print (value-call i '()))" \
	"ERROR: .*: Wrong type argument in position 1 \\(expecting function \\(value of TYPE_CODE_FUNC\\)\\): .*"

    # Incorrect number of arguments.
    gdb_test "p/x fp2" "= $hex.*"
    gdb_scm_test_silent_cmd "gu (define fp3 (history-ref 0))" \
	"get fp3 value from history"
    gdb_scm_test_silent_cmd "gu (set! fp3 (value-dereference fp3))" \
	"dereference fp3"
    gdb_test "gu (print (value-call fp3 (list 10)))" \
	"ERROR: Too few arguments in function call.*"
}

proc test_value_after_death {} {
    # Construct a type while the inferior is still running.
    gdb_scm_test_silent_cmd "gu (define ptrtype (lookup-type \"PTR\"))" \
	"create PTR type"

    # Kill the inferior and remove the symbols.
    gdb_test "kill" "" "kill the inferior" \
	"Kill the program being debugged. .y or n. $" \
	"y"
    gdb_test "file" "" "discard the symbols" \
	"Discard symbol table from.*y or n. $" \
	"y"

    # First do a garbage collect to delete anything unused.  PR 16612.
    gdb_scm_test_silent_cmd "gu (gc)" "garbage collect"

    # Now create a value using that type.  Relies on arg0, created by
    # test_value_in_inferior.
    gdb_scm_test_silent_cmd "gu (define castval (value-cast arg0 (type-pointer ptrtype)))" \
	"cast arg0 to PTR"

    # Make sure the type is deleted.
    gdb_scm_test_silent_cmd "gu (set! ptrtype #f)" \
	"delete PTR type"

    # Now see if the value's type is still valid.
    gdb_test "gu (print (value-type castval))" \
	"= PTR ." "print value's type"
}

# Regression test for invalid subscript operations.  The bug was that
# the type of the value was not being checked before allowing a
# subscript operation to proceed.

proc test_subscript_regression {exefile lang} {
    # Start with a fresh gdb.
    clean_restart ${exefile}

    if ![gdb_guile_runto_main ] {
	fail "can't run to main"
	return
    }

    if {$lang == "c++"} {
	gdb_breakpoint [gdb_get_line_number "break to inspect pointer by reference"]
	gdb_continue_to_breakpoint "break to inspect pointer by reference"

	gdb_scm_test_silent_cmd "print rptr_int" \
	    "Obtain address"
	gdb_scm_test_silent_cmd "gu (define rptr (history-ref 0))" \
	    "set rptr"
	gdb_test "gu (print (value-subscript rptr 0))" \
	    "= 2" "Check pointer passed as reference"

	# Just the most basic test of dynamic_cast -- it is checked in
	# the C++ tests.
	gdb_test "gu (print (value->bool (value-dynamic-cast (parse-and-eval \"base\") (type-pointer (lookup-type \"Derived\")))))" \
	    "= #t"

	# Likewise.
	gdb_test "gu (print (value-dynamic-type (parse-and-eval \"base\")))" \
	    "= Derived \[*\]"
	gdb_test "gu (print (value-dynamic-type (parse-and-eval \"base_ref\")))" \
	    "= Derived \[&\]"
	# A static type case.
	gdb_test "gu (print (value-dynamic-type (parse-and-eval \"5\")))" \
	    "= int"
    }

    gdb_breakpoint [gdb_get_line_number "break to inspect struct and union"]
    gdb_continue_to_breakpoint "break to inspect struct and union"

    gdb_scm_test_silent_cmd "gu (define intv (make-value 1))" \
	"Create int value for subscript test"
    gdb_scm_test_silent_cmd "gu (define stringv (make-value \"foo\"))" \
	"Create string value for subscript test"

    # Try to access an int with a subscript.  This should fail.
    gdb_test "gu (print intv)" \
	"= 1" "Baseline print of an int Guile value"
    gdb_test "gu (print (value-subscript intv 0))" \
	"ERROR: Cannot subscript requested type.*" \
	"Attempt to access an integer with a subscript"

    # Try to access a string with a subscript.  This should pass.
    gdb_test "gu (print stringv)" \
	"= \"foo\"" "Baseline print of a string Guile value"
    gdb_test "gu (print (value-subscript stringv 0))" \
	"= 102 'f'" "Attempt to access a string with a subscript"

    # Try to access an int array via a pointer with a subscript.
    # This should pass.
    gdb_scm_test_silent_cmd "print p" "Build pointer to array"
    gdb_scm_test_silent_cmd "gu (define pointer (history-ref 0))" "set pointer"
    gdb_test "gu (print (value-subscript pointer 0))" \
	"= 1" "Access array via pointer with int subscript"
    gdb_test "gu (print (value-subscript pointer intv))" \
	"= 2" "Access array via pointer with value subscript"

    # Try to access a single dimension array with a subscript to the
    # result.  This should fail.
    gdb_test "gu (print (value-subscript (value-subscript pointer intv) 0))" \
	"ERROR: Cannot subscript requested type.*" \
	"Attempt to access an integer with a subscript 2"

    # Lastly, test subscript access to an array with multiple
    # dimensions.  This should pass.
    gdb_scm_test_silent_cmd "print {\"fu \",\"foo\",\"bar\"}" "Build array"
    gdb_scm_test_silent_cmd "gu (define marray (history-ref 0))" ""
    gdb_test "gu (print (value-subscript (value-subscript marray 1) 2))" \
	"o." "Test multiple subscript"
}

# A few tests of gdb:parse-and-eval.

proc test_parse_and_eval {} {
    gdb_test "gu (print (parse-and-eval \"23\"))" \
	"= 23" "parse-and-eval constant test"
    gdb_test "gu (print (parse-and-eval \"5 + 7\"))" \
	"= 12" "parse-and-eval simple expression test"
    gdb_test "gu (raw-print (parse-and-eval \"5 + 7\"))" \
	"#<gdb:value 12>" "parse-and-eval type test"
}

# Test that values are hashable.
# N.B.: While smobs are hashable, the hash is really non-existent,
# they all get hashed to the same value.  Guile may provide a hash function
# for smobs in a future release.  In the meantime one should use a custom
# hash table that uses gdb:hash-gsmob.

proc test_value_hash {} {
    gdb_test_multiline "Simple Guile value dictionary" \
	"guile" "" \
	"(define one (make-value 1))" "" \
	"(define two (make-value 2))" "" \
	"(define three (make-value 3))" "" \
        "(define vdict (make-hash-table 5))" "" \
	"(hash-set! vdict one \"one str\")" "" \
	"(hash-set! vdict two \"two str\")" "" \
	"(hash-set! vdict three \"three str\")" "" \
	"end"
    gdb_test "gu (print (hash-ref vdict one))" \
	"one str" "Test dictionary hash 1"
    gdb_test "gu (print (hash-ref vdict two))" \
	"two str" "Test dictionary hash 2"
    gdb_test "gu (print (hash-ref vdict three))" \
	"three str" "Test dictionary hash 3"
}

# Build C version of executable.  C++ is built later.
if { [build_inferior "${binfile}" "c"] < 0 } {
    return
}

# Start with a fresh gdb.
clean_restart ${binfile}

# Skip all tests if Guile scripting is not enabled.
if { [skip_guile_tests] } { continue }

gdb_install_guile_utils
gdb_install_guile_module

test_parse_and_eval
test_value_hash

# The following tests require execution.

if ![gdb_guile_runto_main] {
    fail "can't run to main"
    return
}

test_value_in_inferior
test_inferior_function_call
test_strings
test_value_after_death

# Test either C or C++ values. 

test_subscript_regression "${binfile}" "c"

if ![skip_cplus_tests] {
    if { [build_inferior "${binfile}-cxx" "c++"] < 0 } {
	return
    }
    with_test_prefix "c++" {
	test_subscript_regression "${binfile}-cxx" "c++"
    }
}