[PATCH] Allow strings with printf/eval

[PATCH] Allow strings with printf/eval

[Keith Seitz]

If a user attempts to use a string with either `printf' or `eval'
commands, gdb returns an error (inferior not running):

(gdb) printf "%s\n", "hello"
evaluation of this expression requires the target program to be active

However, the parser can certainly handle this case:

(gdb) p "hello"
$1 = "hello"

This discrepancy occurs because printf_c_string does not handle
this specific case.  The passed-in value that we are attempting to print
as a string is TYPE_CODE_ARRAY but it's lval type is not_lval.

printf_c_string will only attempt to print a string from the value's
contents when !TYPE_CODE_PTR, lval is lval_internalvar, and the value's
type is considered a string type:

  if (value->type ()->code () != TYPE_CODE_PTR
      && value->lval () == lval_internalvar
      && c_is_string_type_p (value->type ()))
    {
      ...
    }

Otherwise, it attempts to read the value of the string from the target's
memory (which is what actually generates the "evaluation of this ..."
error message).
---
 gdb/printcmd.c                       |  4 ++--
 gdb/testsuite/gdb.base/eval.exp      | 16 ++++++++++++++++
 gdb/testsuite/gdb.base/printcmds.exp | 25 +++++++++++++++++++++++++
 3 files changed, 43 insertions(+), 2 deletions(-)

diff --git a/gdb/printcmd.c b/gdb/printcmd.c
index dd92e31d31b..7eb28da5e24 100644
--- a/gdb/printcmd.c
+++ b/gdb/printcmd.c
@@ -2447,8 +2447,8 @@ printf_c_string (struct ui_file *stream, const char *format,
 {
   const gdb_byte *str;
 
-  if (value->type ()->code () != TYPE_CODE_PTR
-      && value->lval () == lval_internalvar
+  if (((value->type ()->code () != TYPE_CODE_PTR && value->lval () == lval_internalvar)
+       || (value->type ()->code () == TYPE_CODE_ARRAY && value->lval () == not_lval))
       && c_is_string_type_p (value->type ()))
     {
       size_t len = value->type ()->length ();
diff --git a/gdb/testsuite/gdb.base/eval.exp b/gdb/testsuite/gdb.base/eval.exp
index ead053ca712..8a31c9ea944 100644
--- a/gdb/testsuite/gdb.base/eval.exp
+++ b/gdb/testsuite/gdb.base/eval.exp
@@ -20,3 +20,19 @@ gdb_test_no_output "set \$a = 10" "initialize \$a."
 
 gdb_test "eval \"echo %d\\n\", \$a++" "10" "first eval."
 gdb_test "eval \"echo %d\\n\", \$a*2" "22" "second eval."
+
+gdb_test_no_output "set var \$hi = \"Hi\""
+gdb_test {eval "echo %s, %s%c\n", $hi, "mom", '!'} "Hi, mom!"
+
+set listsize_value -1
+gdb_test_multiple "show listsize" "get listsize value" {
+    -re "Number of source lines gdb will list by default is (\[0-9\]+)\.\r\n$gdb_prompt $" {
+	set listsize_value $expect_out(1,string)
+    }
+}
+
+if {$listsize_value > -1} {
+    gdb_test {eval "echo %s\n", $_gdb_setting_str("listsize")} $listsize_value
+} else {
+    send_log "warning: could not get listsize\n"
+}
diff --git a/gdb/testsuite/gdb.base/printcmds.exp b/gdb/testsuite/gdb.base/printcmds.exp
index 21a2cad458c..db57769c303 100644
--- a/gdb/testsuite/gdb.base/printcmds.exp
+++ b/gdb/testsuite/gdb.base/printcmds.exp
@@ -957,6 +957,30 @@ proc test_printf_with_dfp {} {
     gdb_test "printf \"%Hf %Hf\\n\",1.2df,1.3df" "1.2 1.3"
 }
 
+# Test printf with strings.
+proc test_printf_with_strings {} {
+    global gdb_prompt
+
+    gdb_test {printf "I ate a %s yesterday; it was very %s!\n", "clock", "time-consuming"} \
+	"I ate a clock yesterday; it was very time-consuming!"
+    gdb_test_no_output "set var \$hello = \"Hello\""
+
+    gdb_test {printf "%s, %s%c\n", $hello, "world", '!'} "Hello, world!"
+
+    set listsize_value -1
+    gdb_test_multiple "show listsize" "get listsize value" {
+	-re "Number of source lines gdb will list by default is (\[0-9\]+)\.\r\n$gdb_prompt $" {
+	    set listsize_value $expect_out(1,string)
+	}
+    }
+
+    if {$listsize_value > -1} {
+	gdb_test {printf "%s\n", $_gdb_setting_str("listsize")} $listsize_value
+    } else {
+	send_log "warning: could not get listsize\n"
+    }
+}
+
 proc test_print_symbol {} {
     gdb_test_no_output "set print symbol on"
 
@@ -1171,6 +1195,7 @@ test_print_array_constants
 test_print_enums
 test_printf
 test_printf_with_dfp
+test_printf_with_strings
 test_print_symbol
 test_repeat_bytes
 test_radices
-- 
2.39.2

Re: [PATCH] Allow strings with printf/eval

[Tom Tromey]

>>>>> "Keith" == Keith Seitz via Gdb-patches <gdb-patches@sourceware.org> writes:

Keith> (gdb) printf "%s\n", "hello"
Keith> evaluation of this expression requires the target program to be active

This is https://sourceware.org/bugzilla/show_bug.cgi?id=13098

Keith> -  if (value->type ()->code () != TYPE_CODE_PTR
Keith> -      && value->lval () == lval_internalvar
Keith> +  if (((value->type ()->code () != TYPE_CODE_PTR && value->lval () == lval_internalvar)
Keith> +       || (value->type ()->code () == TYPE_CODE_ARRAY && value->lval () == not_lval))

I wonder if lval check is necessary here.

Tom

Re: [PATCH] Allow strings with printf/eval

[Keith Seitz]

On 4/24/23 10:12, Tom Tromey wrote:
>>>>>> "Keith" == Keith Seitz via Gdb-patches <gdb-patches@sourceware.org> writes:
> 
> Keith> (gdb) printf "%s\n", "hello"
> Keith> evaluation of this expression requires the target program to be active
> 
> This is https://sourceware.org/bugzilla/show_bug.cgi?id=13098

Thank you for that bz; I will add a reference in the commit log.

> Keith> -  if (value->type ()->code () != TYPE_CODE_PTR
> Keith> -      && value->lval () == lval_internalvar
> Keith> +  if (((value->type ()->code () != TYPE_CODE_PTR && value->lval () == lval_internalvar)
> Keith> +       || (value->type ()->code () == TYPE_CODE_ARRAY && value->lval () == not_lval))
> 
> I wonder if lval check is necessary here.

It does not appear to be, but I erred on the overly conservative side when
implementing the change. Regression testing shows no differences when
removing that lval check.

I've removed it locally.

Keith

Re: [PATCH] Allow strings with printf/eval

[Tom Tromey]

>>>>> "Keith" == Keith Seitz <keiths@redhat.com> writes:

Keith> On 4/24/23 10:12, Tom Tromey wrote:
>>>>>>> "Keith" == Keith Seitz via Gdb-patches <gdb-patches@sourceware.org> writes:
Keith> (gdb) printf "%s\n", "hello"
Keith> evaluation of this expression requires the target program to be active
>> This is https://sourceware.org/bugzilla/show_bug.cgi?id=13098

Keith> Thank you for that bz; I will add a reference in the commit log.

Keith> -  if (value->type ()->code () != TYPE_CODE_PTR
Keith> -      && value->lval () == lval_internalvar
Keith> +  if (((value->type ()->code () != TYPE_CODE_PTR && value->lval () == lval_internalvar)
Keith> +       || (value->type ()->code () == TYPE_CODE_ARRAY && value->lval () == not_lval))
>> I wonder if lval check is necessary here.

Keith> It does not appear to be, but I erred on the overly conservative side when
Keith> implementing the change. Regression testing shows no differences when
Keith> removing that lval check.

Keith> I've removed it locally.

I think with those 2 small changes, the patch is ok.  Thank you.

Tom

Re: [PATCH] Allow strings with printf/eval

[Keith Seitz]

On 4/25/23 13:21, Tom Tromey wrote:
> 
> I think with those 2 small changes, the patch is ok.  Thank you.

Thank you for the review. I have pushed this.

Keith