Patch to support spaces in filenames & paths

Patch to support spaces in filenames & paths

[Jon Beniston]

[-- Attachment #1: Type: text/plain, Size: 366 bytes --]

Hi,

Attached is a patch that adds support for spaces in filenames or paths for
CLI commands such as restore, by supporting quotes. E.g:

restore "path with spaces/test.x" 

2008-12-02 Jon Beniston <jon@beniston.com>

        * cli/cli-dump.c: (scan_filename_with_cleanup) Add support for
quoted 
        strings to allow spaces in paths and filenames.

Cheers,
Jon

[-- Attachment #2: quoted_filenames.patch --]
[-- Type: application/octet-stream, Size: 1555 bytes --]

Index: cli-dump.c
===================================================================
RCS file: /home/CVSROOT/insight/gdb/cli/cli-dump.c,v
retrieving revision 1.1.1.4
diff -c -p -r1.1.1.4 cli-dump.c
*** cli-dump.c	22 Nov 2008 18:47:15 -0000	1.1.1.4
--- cli-dump.c	2 Dec 2008 12:55:36 -0000
*************** scan_filename_with_cleanup (char **cmd, 
*** 98,111 ****
      }
    else
      {
-       /* FIXME: should parse a possibly quoted string.  */
-       char *end;
- 
        (*cmd) = skip_spaces (*cmd);
!       end = *cmd + strcspn (*cmd, " \t");
!       filename = savestring ((*cmd), end - (*cmd));
        make_cleanup (xfree, filename);
-       (*cmd) = skip_spaces (end);
      }
    gdb_assert (filename != NULL);
  
--- 98,125 ----
      }
    else
      {
        (*cmd) = skip_spaces (*cmd);
!       if ((*(*cmd)) == '"')
!         {
!           size_t skip;
!           
!           (*cmd)++;
!           skip = strcspn ((*cmd), "\"");
!           filename = savestring ((*cmd), skip);
!           if (strlen((*cmd)) == skip)
!             error (_("Unterminated string or character constant."));
!           else  
!             (*cmd) = skip_spaces ((*cmd) + skip + 1);  
!         }  
!       else
!         {   
!           char *end;
! 
!           end = *cmd + strcspn (*cmd, " \t");
!           filename = savestring ((*cmd), end - (*cmd));
!           (*cmd) = skip_spaces (end);
!         }  
        make_cleanup (xfree, filename);
      }
    gdb_assert (filename != NULL);
  

Re: Patch to support spaces in filenames & paths

[Michael Snyder]

Jon Beniston wrote:
> Hi,
> 
> Attached is a patch that adds support for spaces in filenames or paths for
> CLI commands such as restore, by supporting quotes. E.g:
> 
> restore "path with spaces/test.x" 
> 
> 2008-12-02 Jon Beniston <jon@beniston.com>
> 
>         * cli/cli-dump.c: (scan_filename_with_cleanup) Add support for
> quoted 
>         strings to allow spaces in paths and filenames.
> 

Fascinating -- now we have two overlapping patches submitted on
the same day, both mentioning "restore" as one of the commands
affected.

Denis, is it possible that Jon's patch will serve in place of
the filename portion of your patch?  And that you could then
resubmit your patch with just the other portions?

Jon's patch:
http://sourceware.org/ml/gdb-patches/2008-12/msg00032.html

Denis' patch:
http://sourceware.org/ml/gdb-patches/2008-12/msg00029.html

Re: Patch to support spaces in filenames & paths

[Daniel Jacobowitz]

On Tue, Dec 02, 2008 at 01:17:22PM -0800, Michael Snyder wrote:
> Denis, is it possible that Jon's patch will serve in place of
> the filename portion of your patch?  And that you could then
> resubmit your patch with just the other portions?
>
> Jon's patch:
> http://sourceware.org/ml/gdb-patches/2008-12/msg00032.html
>
> Denis' patch:
> http://sourceware.org/ml/gdb-patches/2008-12/msg00029.html

I have not looked at the patches in depth but I encourage Denis's
approach - uniform parsing is a Very Good Thing and buildargv is what
we use elsewhere.

-- 
Daniel Jacobowitz
CodeSourcery

Re: Patch to support spaces in filenames & paths

[Pedro Alves]

On Tuesday 02 December 2008 23:37:38, Daniel Jacobowitz wrote:
> On Tue, Dec 02, 2008 at 01:17:22PM -0800, Michael Snyder wrote:
> > Denis, is it possible that Jon's patch will serve in place of
> > the filename portion of your patch?  And that you could then
> > resubmit your patch with just the other portions?
> >
> > Jon's patch:
> > http://sourceware.org/ml/gdb-patches/2008-12/msg00032.html
> >
> > Denis' patch:
> > http://sourceware.org/ml/gdb-patches/2008-12/msg00029.html
> 
> I have not looked at the patches in depth but I encourage Denis's
> approach - uniform parsing is a Very Good Thing and buildargv is what
> we use elsewhere.

Same here, but I think it should be using the new gdb_buildargv.

-- 
Pedro Alves

Re: Patch to support spaces in filenames & paths

[Doug Evans]

On Tue, Dec 2, 2008 at 3:37 PM, Daniel Jacobowitz <drow@false.org> wrote:
> On Tue, Dec 02, 2008 at 01:17:22PM -0800, Michael Snyder wrote:
>> Denis, is it possible that Jon's patch will serve in place of
>> the filename portion of your patch?  And that you could then
>> resubmit your patch with just the other portions?
>>
>> Jon's patch:
>> http://sourceware.org/ml/gdb-patches/2008-12/msg00032.html
>>
>> Denis' patch:
>> http://sourceware.org/ml/gdb-patches/2008-12/msg00029.html
>
> I have not looked at the patches in depth but I encourage Denis's
> approach - uniform parsing is a Very Good Thing and buildargv is what
> we use elsewhere.

While perhaps not applicable in Denis' case (since the command accepts
"a b c" instead of "a, b, c" (though I wonder if it could accept
both), for completeness' sake there is also parse_to_comma_and_eval.

(gdb) printf "%d %d %d\n", 1 + 1, 2 + 2, 3 + 3
2 4 6

With buildargv it'd be

(gdb) printf "%d %d %d\n" "1 + 1" "2 + 2" "3 + 3"

and that just doesn't sit right. :-)

Re: Patch to support spaces in filenames & paths

[Denis PILAT]

Doug Evans wrote:
> On Tue, Dec 2, 2008 at 3:37 PM, Daniel Jacobowitz <drow@false.org> wrote:
>   
>> On Tue, Dec 02, 2008 at 01:17:22PM -0800, Michael Snyder wrote:
>>     
>>> Denis, is it possible that Jon's patch will serve in place of
>>> the filename portion of your patch?  And that you could then
>>> resubmit your patch with just the other portions?
>>>
>>> Jon's patch:
>>> http://sourceware.org/ml/gdb-patches/2008-12/msg00032.html
>>>
>>> Denis' patch:
>>> http://sourceware.org/ml/gdb-patches/2008-12/msg00029.html
>>>       
>> I have not looked at the patches in depth but I encourage Denis's
>> approach - uniform parsing is a Very Good Thing and buildargv is what
>> we use elsewhere.
>>     
>
> While perhaps not applicable in Denis' case (since the command accepts
> "a b c" instead of "a, b, c" (though I wonder if it could accept
> both), for completeness' sake there is also parse_to_comma_and_eval.
>
> (gdb) printf "%d %d %d\n", 1 + 1, 2 + 2, 3 + 3
> 2 4 6
>
> With buildargv it'd be
>
> (gdb) printf "%d %d %d\n" "1 + 1" "2 + 2" "3 + 3"
>
> and that just doesn't sit right. :-)
>
>   
Actual implementation of append/dump/restore does not accept spaces at 
all, except in the last argument, and I think it's just a side effect.
To me comma must not be considered as a separator since dump command 
accepts expression and the calculation of START , END address or OFFSET 
is often a function call like sizeof (), but can be more than that like 
a function that takes more than one parameters, "max(a,b)" or whatever.

I'm fine with using gdb_buildargv, that would simplify my patch 
(http://sourceware.org/ml/gdb-patches/2008-12/msg00029.html).

-- 
Denis

Re: Patch to support spaces in filenames & paths

[Daniel Jacobowitz]

On Thu, Dec 04, 2008 at 10:37:39AM +0100, Denis PILAT wrote:
> Actual implementation of append/dump/restore does not accept spaces at  
> all, except in the last argument, and I think it's just a side effect.
> To me comma must not be considered as a separator since dump command  
> accepts expression and the calculation of START , END address or OFFSET  
> is often a function call like sizeof (), but can be more than that like a 
> function that takes more than one parameters, "max(a,b)" or whatever.

I don't think it applies here, but take a peek at the function Doug
mentioned - it does handle nested commas.  It's not perfect, but it's
pretty good.

-- 
Daniel Jacobowitz
CodeSourcery

Re: Patch to support spaces in filenames & paths

[Denis PILAT]

I made some experiments with parse_to_comma_and_eval, but I think I must 
be sure to understand what you need before going further in my 
implementation, and summarize the situation
My original proposal 
(http://sourceware.org/ml/gdb-patches/2008-12/msg00029.html) allows 
spaces in dump/restore/append command arguments, having the constraint 
for user to quote its arguments that contains spaces, but that doesn't 
break actual gdb behavior.

Then Doug said

While perhaps not applicable in Denis' case (since the command accepts

"a b c" instead of "a, b, c" (though I wonder if it could accept

both), for completeness' sake there is also parse_to_comma_and_eval.


In the current implementation, space is considered as a separator for 
all arguments, except for the last that can contains any spaces and will 
be evaluated entirely.
Accepting "a, b, c" as 3 argument consists in making comma the separator 
for arguments, and that will break existing front-ends, it's an API change.

The patch I have proposed accepts "a, b, c" but considers it's a single 
argument, like it could be for any expression with coma (ie "max(a, b)").

Please enlighten me on what kind of syntax you'd like 
dump/restore/append to accept, knowing that a, b and c could also be 
complex expressions with comma and spaces as well.

Thanks for your feedback
-- 
Denis



Daniel Jacobowitz wrote:
> On Thu, Dec 04, 2008 at 10:37:39AM +0100, Denis PILAT wrote:
>   
>> Actual implementation of append/dump/restore does not accept spaces at  
>> all, except in the last argument, and I think it's just a side effect.
>> To me comma must not be considered as a separator since dump command  
>> accepts expression and the calculation of START , END address or OFFSET  
>> is often a function call like sizeof (), but can be more than that like a 
>> function that takes more than one parameters, "max(a,b)" or whatever.
>>     
>
> I don't think it applies here, but take a peek at the function Doug
> mentioned - it does handle nested commas.  It's not perfect, but it's
> pretty good.
>
>   

Re: Patch to support spaces in filenames & paths

[Daniel Jacobowitz]

On Wed, Dec 10, 2008 at 04:31:48PM +0100, Denis PILAT wrote:
> I made some experiments with parse_to_comma_and_eval, but I think I must  
> be sure to understand what you need before going further in my  
> implementation, and summarize the situation

parse_to_comma_and_eval is only required for functions that take more
than one expression.  This doesn't, so it doesn't apply.

I see that buildargv is going to be a problem if the last argument is
a free-form expression.  I'm not entirely sure what to do about that.
Maybe a GDB-local version that extracts one argument string at a time
and returns that, plus a pointer to the next unparsed bit of the
command line.

The important thing, in my opinion, is to use the same handling of
quotes and spaces that buildargv does for consistency with other
commands, like "file".

Maybe we can solve this later and just use buildargv.  Compare with
add_symbol_file_command which takes an address.

-- 
Daniel Jacobowitz
CodeSourcery

Re: Patch to support spaces in filenames & paths

[Denis PILAT]

[-- Attachment #1: Type: text/plain, Size: 1929 bytes --]


Daniel Jacobowitz wrote:
> On Wed, Dec 10, 2008 at 04:31:48PM +0100, Denis PILAT wrote:
>   
>> I made some experiments with parse_to_comma_and_eval, but I think I must  
>> be sure to understand what you need before going further in my  
>> implementation, and summarize the situation
>>     
>
> parse_to_comma_and_eval is only required for functions that take more
> than one expression.  This doesn't, so it doesn't apply.
>
> I see that buildargv is going to be a problem if the last argument is
> a free-form expression.  I'm not entirely sure what to do about that.
> Maybe a GDB-local version that extracts one argument string at a time
> and returns that, plus a pointer to the next unparsed bit of the
> command line.
>
> The important thing, in my opinion, is to use the same handling of
> quotes and spaces that buildargv does for consistency with other
> commands, like "file".
>
> Maybe we can solve this later and just use buildargv.  Compare with
> add_symbol_file_command which takes an address.
>
>   
Hi Daniel,

I'm fine with what you said.
I read your comments and produce a new patch that uses gdb_buildargv (as 
Pedro suggested), inspired from add_symbol_file_command as well.
That fixes the problem of last argument that could be a free form. Now 
all the last arguments, are concatenate into a single one.
It avoids as well any possible regression since it behaves exactly the 
same as the previous gdb, except that it accepts arguments with spaces 
like file command does.
And it consistent with most of commands that accept spaces in their 
arguments.

I also removed lot of dead code in cli/cli-dump.c and fopen_with_cleanup 
is now used only locally, so removed from cli-dump.h.

Dump restore and append commands are impacted by this patch.

As the "binary" option was missing from the help message, I add it as 
well into help.exp in a separate patch.

Patch are attached.
Ok for commit ?

Denis



[-- Attachment #2: cli-dump.patch --]
[-- Type: text/plain, Size: 9404 bytes --]

2009-01-05  Denis Pilat  <denis.pilat@st.com>

	* cli/cli-dump.h (scan_filename_with_cleanup, skip_spaces,
	fopen_with_cleanup, parse_and_eval_with_error,
	scan_expression_with_cleanup): Remove.
	* cli/cli-dump.c: Likewise.
	(dump_memory_to_file): use of gdb_buildargv for argument parsing.
	Concatenate last arguments on the command line.
	(dump_value_to_file, restore_command): Likewise.
	(_initialize_cli_dump): add "binary" option for the help message of 
	restore command.


Index: cli/cli-dump.c
===================================================================
RCS file: /cvs/src/src/gdb/cli/cli-dump.c,v
retrieving revision 1.30
diff -u -p -r1.30 cli-dump.c
--- cli/cli-dump.c	3 Jan 2009 05:57:54 -0000	1.30
+++ cli/cli-dump.c	5 Jan 2009 14:14:36 -0000
@@ -33,75 +33,6 @@
 
 #define XMALLOC(TYPE) ((TYPE*) xmalloc (sizeof (TYPE)))
 
-
-char *
-skip_spaces (char *chp)
-{
-  if (chp == NULL)
-    return NULL;
-  while (isspace (*chp))
-    chp++;
-  return chp;
-}
-
-char *
-scan_expression_with_cleanup (char **cmd, const char *def)
-{
-  if ((*cmd) == NULL || (**cmd) == '\0')
-    {
-      char *exp = xstrdup (def);
-      make_cleanup (xfree, exp);
-      return exp;
-    }
-  else
-    {
-      char *exp;
-      char *end;
-
-      end = (*cmd) + strcspn (*cmd, " \t");
-      exp = savestring ((*cmd), end - (*cmd));
-      make_cleanup (xfree, exp);
-      (*cmd) = skip_spaces (end);
-      return exp;
-    }
-}
-
-
-char *
-scan_filename_with_cleanup (char **cmd, const char *defname)
-{
-  char *filename;
-  char *fullname;
-
-  /* FIXME: Need to get the ``/a(ppend)'' flag from somewhere.  */
-
-  /* File.  */
-  if ((*cmd) == NULL)
-    {
-      if (defname == NULL)
-	error (_("Missing filename."));
-      filename = xstrdup (defname);
-      make_cleanup (xfree, filename);
-    }
-  else
-    {
-      /* FIXME: should parse a possibly quoted string.  */
-      char *end;
-
-      (*cmd) = skip_spaces (*cmd);
-      end = *cmd + strcspn (*cmd, " \t");
-      filename = savestring ((*cmd), end - (*cmd));
-      make_cleanup (xfree, filename);
-      (*cmd) = skip_spaces (end);
-    }
-  gdb_assert (filename != NULL);
-
-  fullname = tilde_expand (filename);
-  make_cleanup (xfree, fullname);
-  
-  return fullname;
-}
-
 FILE *
 fopen_with_cleanup (const char *filename, const char *mode)
 {
@@ -222,24 +153,35 @@ dump_memory_to_file (char *cmd, char *mo
   char *lo_exp;
   char *hi_exp;
   int len;
+  char ** argv;
+  int argcnt;
 
-  /* Open the file.  */
-  filename = scan_filename_with_cleanup (&cmd, NULL);
+  argv = gdb_buildargv (cmd);
+  make_cleanup_freeargv (argv);
 
+  /* Open the file.  */
+  if (argv == NULL || argv[0] == NULL)
+    error (_("Missing filename."));
+  filename = tilde_expand (argv[0]);
+  make_cleanup (xfree, filename);
+  
   /* Find the low address.  */
-  if (cmd == NULL || *cmd == '\0')
+  lo_exp = argv[1];
+  if (lo_exp == NULL)
     error (_("Missing start address."));
-  lo_exp = scan_expression_with_cleanup (&cmd, NULL);
 
   /* Find the second address - rest of line.  */
-  if (cmd == NULL || *cmd == '\0')
+  hi_exp = argv[2];
+  if (hi_exp == NULL)
     error (_("Missing stop address."));
-  hi_exp = cmd;
+  for (argcnt = 3; argv[argcnt] != NULL; argcnt++)
+    hi_exp = concat (hi_exp, " ", argv[argcnt], (char *)NULL);
 
   lo = parse_and_eval_address (lo_exp);
   hi = parse_and_eval_address (hi_exp);
   if (hi <= lo)
     error (_("Invalid memory address range (start >= end)."));
+
   count = hi - lo;
 
   /* FIXME: Should use read_memory_partial() and a magic blocking
@@ -273,14 +215,28 @@ dump_value_to_file (char *cmd, char *mod
   struct cleanup *old_cleanups = make_cleanup (null_cleanup, NULL);
   struct value *val;
   char *filename;
+  char ** argv;
+  char *strvalue;
+  int argcnt;
 
-  /* Open the file.  */
-  filename = scan_filename_with_cleanup (&cmd, NULL);
+  argv = gdb_buildargv (cmd);
+  make_cleanup_freeargv (argv);
 
-  /* Find the value.  */
-  if (cmd == NULL || *cmd == '\0')
+  /* Open the file.  */
+  if (argv == NULL || argv[0] == NULL)
+    error (_("Missing filename."));
+  filename = tilde_expand (argv[0]);
+  make_cleanup (xfree, filename);
+
+  /* Find the value - rest of line.  */
+  strvalue = argv[1];
+  if (strvalue == NULL)
     error (_("No value to %s."), *mode == 'a' ? "append" : "dump");
-  val = parse_and_eval (cmd);
+
+  for (argcnt = 2; argv[argcnt] != NULL; argcnt++)
+    strvalue = concat (strvalue, " ", argv[argcnt], (char *)NULL);
+
+  val = parse_and_eval (strvalue);
   if (val == NULL)
     error (_("Invalid expression."));
 
@@ -559,8 +515,14 @@ restore_command (char *args, int from_tt
   char *filename;
   struct callback_data data;
   bfd *ibfd;
-  int binary_flag = 0;
-
+  char **argv;
+  int binary_flag;
+  char *arg;
+  int arg_num;
+  char *end_addr = NULL;
+  filename = NULL;
+  binary_flag = 0;
+  
   if (!target_has_execution)
     noprocess ();
 
@@ -568,43 +530,54 @@ restore_command (char *args, int from_tt
   data.load_start  = 0;
   data.load_end    = 0;
 
-  /* Parse the input arguments.  First is filename (required). */
-  filename = scan_filename_with_cleanup (&args, NULL);
-  if (args != NULL && *args != '\0')
-    {
-      char *binary_string = "binary";
+  argv = gdb_buildargv (args);
+  make_cleanup_freeargv (argv);
 
-      /* Look for optional "binary" flag.  */
-      if (strncmp (args, binary_string, strlen (binary_string)) == 0)
-	{
-	  binary_flag = 1;
-	  args += strlen (binary_string);
-	  args = skip_spaces (args);
-	}
-      /* Parse offset (optional). */
-      if (args != NULL && *args != '\0')
-      data.load_offset = 
-	parse_and_eval_long (scan_expression_with_cleanup (&args, NULL));
-      if (args != NULL && *args != '\0')
-	{
-	  /* Parse start address (optional). */
-	  data.load_start = 
-	    parse_and_eval_long (scan_expression_with_cleanup (&args, NULL));
-	  if (args != NULL && *args != '\0')
-	    {
-	      /* Parse end address (optional). */
-	      data.load_end = parse_and_eval_long (args);
-	      if (data.load_end <= data.load_start)
-		error (_("Start must be less than end."));
-	    }
-	}
+  /* At least filename argument is required.  */
+  if (argv == NULL || argv[0] == NULL)
+    error (_("Missing filename as the first argument."));
+
+  /* The first argument is the file name.  */
+  filename = tilde_expand (argv[0]);
+  make_cleanup (xfree, filename);
+
+  /* Look for optionals arguments from here: "binary", offset, 
+   start and end address.  */
+  if (argv[1] != NULL && strcmp (argv[1], "binary") == 0) 
+        binary_flag = 1;
+
+  arg_num = 1 + binary_flag;
+  for (arg = argv[arg_num]; arg != NULL; arg = argv[++arg_num])
+    {
+      /* parse offset argument (optional).  */
+      if (arg_num == 1 + binary_flag)
+          data.load_offset = parse_and_eval_long (arg);
+      
+      /* parse start address argument (optional).  */
+      else if (arg_num == 2 + binary_flag)
+        data.load_start = parse_and_eval_long (arg);
+  
+      /* Rest of line is considered as end address optional argument.  */
+      else if (arg_num == 3 + binary_flag )
+        end_addr = arg;
+      else if (arg_num > 3 + binary_flag)
+        end_addr = concat (end_addr, " ", arg, (char *)NULL);
+    }      
+
+  /* If end address argument is given, parse it and test it's greater than 
+     the start address.  */
+  if (end_addr != NULL)
+    {
+      data.load_end = parse_and_eval_long (end_addr);
+      if (data.load_end <= data.load_start)
+        error (_("Start must be less than end."));
     }
 
   if (info_verbose)
     printf_filtered ("Restore file %s offset 0x%lx start 0x%lx end 0x%lx\n",
-		     filename, (unsigned long) data.load_offset, 
-		     (unsigned long) data.load_start, 
-		     (unsigned long) data.load_end);
+                     filename, (unsigned long) data.load_offset, 
+                     (unsigned long) data.load_start, 
+                     (unsigned long) data.load_end);
 
   if (binary_flag)
     {
@@ -775,7 +748,8 @@ to the specified FILE in raw target orde
 
   c = add_com ("restore", class_vars, restore_command, _("\
 Restore the contents of FILE to target memory.\n\
-Arguments are FILE OFFSET START END where all except FILE are optional.\n\
+Arguments are FILE \"binary\" OFFSET START END\n\
+where all except FILE are optional.\n\
 OFFSET will be added to the base address of the file (default zero).\n\
 If START and END are given, only the file contents within that range\n\
 (file relative) will be restored to target memory."));
Index: cli/cli-dump.h
===================================================================
RCS file: /cvs/src/src/gdb/cli/cli-dump.h,v
retrieving revision 1.7
diff -u -p -r1.7 cli-dump.h
--- cli/cli-dump.h	3 Jan 2009 05:57:54 -0000	1.7
+++ cli/cli-dump.h	5 Jan 2009 14:14:36 -0000
@@ -24,15 +24,4 @@ extern void add_dump_command (char *name
 			      void (*func) (char *args, char *mode),
 			      char *descr);
 
-/* Utilities for doing the dump.  */
-extern char *scan_filename_with_cleanup (char **cmd, const char *defname);
-
-extern char *scan_expression_with_cleanup (char **cmd, const char *defname);
-
-extern FILE *fopen_with_cleanup (const char *filename, const char *mode);
-
-extern char *skip_spaces (char *inp);
-
-extern struct value *parse_and_eval_with_error (char *exp, const char *fmt, ...) ATTR_FORMAT (printf, 2, 3);
-
 #endif

[-- Attachment #3: help.exp.patch --]
[-- Type: text/plain, Size: 1738 bytes --]

2009-01-05  Denis Pilat  <denis.pilat@st.com>

	* gdb.base/help.exp: add "binary" option for restore command.

Index: gdb.base/help.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.base/help.exp,v
retrieving revision 1.31
diff -u -p -r1.31 help.exp
--- gdb.base/help.exp	16 Nov 2008 18:03:25 -0000	1.31
+++ gdb.base/help.exp	5 Jan 2009 14:02:44 -0000
@@ -355,7 +355,7 @@ gdb_test "help run" "Start debugged prog
 # test help rbreak
 gdb_test "help rbreak" "Set a breakpoint for all functions matching REGEXP\." "help rbreak"
 # test help restore
-gdb_test "help restore" "Restore the contents of FILE to target memory\.\[\r\n\]+Arguments are FILE OFFSET START END where all except FILE are optional\.\[\r\n\]+OFFSET will be added to the base address of the file \\(default zero\\)\.\[\r\n\]+If START and END are given, only the file contents within that range\[\r\n\]+\\(file relative\\) will be restored to target memory\."
+gdb_test "help restore" "Restore the contents of FILE to target memory\.\[\r\n\]+Arguments are FILE \"binary\" OFFSET START END\[\r\n\]+where all except FILE are optional\.\[\r\n\]+OFFSET will be added to the base address of the file \\(default zero\\)\.\[\r\n\]+If START and END are given, only the file contents within that range\[\r\n\]+\\(file relative\\) will be restored to target memory\."
 # test help return
 gdb_test "help return" "Make selected stack frame return to its caller\.\[\r\n\]+Control remains in the debugger, but when you continue\[\r\n\]+execution will resume in the frame above the one now selected\.\[\r\n\]+If an argument is given, it is an expression for the value to return\." "help return"
 # test help reverse-search