Re: [PATCH v4 2/4] gdb/cli: add '.' as an argument for 'list' command

[Bruno Larsen <blarsen@redhat.com>]

On 14/07/2023 19:50, Pedro Alves wrote:
> Hi,
>
> Sorry for coming late to the party here, but while trying to catch up on the
> list I spotted a few things that I think should be fixed.  See below.
I forgot to send the email, but I've already pushed this patch... I'll 
fix up the things you pointed in a moment and send a new patch, I just 
have a few questions.
>
> BTW, I think the feature is cool!
>
> On 2023-07-13 11:24, Bruno Larsen via Gdb-patches wrote:
>> Currently, after the user has used the list command once, there is no
>> self-contained way to ask GDB to print the location where the inferior is
>> stopped.  The current best options require either using a separate
>> command to scope out where the inferior is stopped, or using "list *$pc"
>> requiring knowledge of GDB standard registers.  This commit adds a way
>> to do that using '.' as a new argument for the 'list' command.  If the
>> inferior isn't running, the command prints around the main function.
>>
>> Because this necessitated having the inferior running and the test was
>> (seemingly unnecessarily) using printf in a non-essential way and it
>> would make the resulting log harder to read for no benefit, it was
>> replaced by a differen t statement.
>> ---
>>   gdb/NEWS                        |  4 ++++
>>   gdb/cli/cli-cmds.c              | 31 ++++++++++++++++++++++++--
>>   gdb/doc/gdb.texinfo             |  5 +++++
>>   gdb/testsuite/gdb.base/list.exp | 39 +++++++++++++++++++++++++++++++++
>>   gdb/testsuite/gdb.base/list1.c  |  2 +-
>>   5 files changed, 78 insertions(+), 3 deletions(-)
>>
>> diff --git a/gdb/NEWS b/gdb/NEWS
>> index b924834d3d7..eef440a5242 100644
>> --- a/gdb/NEWS
>> +++ b/gdb/NEWS
>> @@ -84,6 +84,10 @@
>>     is 64k.  To print longer strings you should increase
>>     'max-value-size'.
>>   
>> +* The 'list' command now accepts '.' as an argument, which tells GDB to
>> +  print the location where the inferior is stopped.  If the inferior hasn't
>> +  started yet, the command will print around the main function.
> It would be more accurate to say that it's where the current thread is stopped.
>
> Say you run to breakpoint hit in thread 1, and then switch to thread 2, and then do
> "list .".  That will show you the current frame of thread 2, while "where the
> inferior is stopped" could very well be interpreted as "thread 1".

The wording does need changing, I agree. I think using the wording that 
already exists in the documentation is the best way to go: "prints 
around the last solitary line printed as part of displaying a stack 
frame". The wording I originally used (and the improvement you 
suggested) could make it sound like I would always print the lowermost 
frame, when that is not the point I wanted. What I wanted was to re-do 
what the first usage of "list" does without needing to know the lines.

My big question: Can I change the NEWS entry, or is that some taboo 
that's best avoided?

>
>> +
>>   * New commands
>>   
>>   maintenance print record-instruction [ N ]
>> diff --git a/gdb/cli/cli-cmds.c b/gdb/cli/cli-cmds.c
>> index 00977bc2ee3..1c459afdc97 100644
>> --- a/gdb/cli/cli-cmds.c
>> +++ b/gdb/cli/cli-cmds.c
>> @@ -1234,14 +1234,14 @@ list_command (const char *arg, int from_tty)
>>     const char *p;
>>   
>>     /* Pull in the current default source line if necessary.  */
>> -  if (arg == NULL || ((arg[0] == '+' || arg[0] == '-') && arg[1] == '\0'))
>> +  if (arg == NULL || ((arg[0] == '+' || arg[0] == '-' || arg[0] == '.') && arg[1] == '\0'))
>>       {
>>         set_default_source_symtab_and_line ();
>>         symtab_and_line cursal = get_current_source_symtab_and_line ();
>>   
>>         /* If this is the first "list" since we've set the current
>>   	 source line, center the listing around that line.  */
>> -      if (get_first_line_listed () == 0)
>> +      if (get_first_line_listed () == 0 && (arg == nullptr || arg[0] != '.'))
>>   	{
>>   	  list_around_line (arg, cursal);
>>   	}
>> @@ -1263,6 +1263,32 @@ list_command (const char *arg, int from_tty)
>>   	  print_source_lines (cursal.symtab, range, 0);
>>   	}
>>   
>> +      /* "l ." lists the default location again.  */
> Spelling out "list ." would be better for grepping, IMHO.
I followed the convention of the code around it. All used only 'l' 
instead of "list", but I can change that easily enough.
>
>> +      else if (arg[0] == '.')
>> +	{
>> +	  try
>> +	    {
>> +	      /* Find the current line by getting the PC of the currently
>> +		 selected frame, and finding the line associated to it.  */
>> +	      frame_info_ptr frame = get_selected_frame (nullptr);
> So this is actually using the selected frame, not the current frame.  So even
> the "where the thread is stopped" description would be incorrect.  If you really
> wanted to use frame #0 (where the thread stopped), then this should use get_current_frame.
Yeah... that's the confusion I mentioned earlier... I want the selected 
frame, since its what "list" does when you first call it after entering 
a frame.
>
>
>> +	      CORE_ADDR curr_pc = get_frame_pc (frame);
>> +	      cursal = find_pc_line (curr_pc, 0);
>> +	    }
>> +	  catch (const gdb_exception &e)
>> +	    {
>> +		  /* If there was an exception above, it means the inferior
>> +		     is not running, so reset the current source location to
>> +		     the default.  */
> Aww.  Please don't use a try/catch for this...  You can just check target_has_execution.
Oh, I thought target_has_execution would query if the target can execute 
the inferior, rather than looking if the inferior is being executed. 
Will change
>
> Also, if an exception would be good for this (which it isn't), please don't catch
> gdb_exception -- it catches QUIT/Ctrl-C as well, which is most often wrong.  You would
> want gdb_exception_error normally.
Also news to me! I probably won't use it, but its good to know for when 
I need it :)
>> +		  clear_current_source_symtab_and_line ();
>> +		  set_default_source_symtab_and_line ();
>> +		  cursal = get_current_source_symtab_and_line ();
>> +	    }
>> +	  list_around_line (arg, cursal);
>> +	  /* Advance argument so just pressing "enter" after using "list ."
>> +	     will print the following lines instead of the same lines again. */
>> +	  arg++;
>> +	}
>> +
>>         return;
>>       }
>>   
>> @@ -2770,6 +2796,7 @@ and send its output to SHELL_COMMAND."));
>>       = add_com ("list", class_files, list_command, _("\
>>   List specified function or line.\n\
>>   With no argument, lists ten more lines after or around previous listing.\n\
>> +\"list .\" lists ten lines arond where the inferior is stopped.\n\
> arond -> around
>
>>   \"list -\" lists the ten lines before a previous ten-line listing.\n\
>>   One argument specifies a line, and ten lines are listed around that line.\n\
>>   Two arguments with comma between specify starting and ending lines to list.\n\
>> diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
>> index b10c06ae91f..7619efe3de9 100644
>> --- a/gdb/doc/gdb.texinfo
>> +++ b/gdb/doc/gdb.texinfo
>> @@ -9148,6 +9148,11 @@ Stack}), this prints lines centered around that line.
>>   
>>   @item list -
>>   Print lines just before the lines last printed.
>> +
>> +@item list .
>> +Print the lines surrounding the location that is where the inferior
>> +is stopped.  If the inferior is not running, print around the main
>> +function instead.
> This should be clarified as well.
>
>>   @end table
>>   
>>   @cindex @code{list}, how many lines to display
>> diff --git a/gdb/testsuite/gdb.base/list.exp b/gdb/testsuite/gdb.base/list.exp
>> index 18ecd13ac0f..ed178a1dd95 100644
>> --- a/gdb/testsuite/gdb.base/list.exp
>> +++ b/gdb/testsuite/gdb.base/list.exp
>> @@ -400,6 +400,42 @@ proc test_list_invalid_args {} {
>>   	"second use of \"list +INVALID\""
>>   }
>>   
>> +proc test_list_current_location {} {
>> +    global binfile
>> +    # If the first "list" command that GDB runs is "list ." GDB may be
>> +    # unable to recognize that the inferior isn't running, so we should
>> +    # reload the inferior to test that condition.
> I don't understand this comment.  Why would it be unable to do so?
doh, this is a comment form previous iterations. Back when I thought 
"list ." should error, we were entering the "never before listed" if 
clause, instead of finding "list ." which was a problem. Now that the 
behavior is the same, there isn't a reason to restart the inferior.
>
>> +    clean_restart
>> +    gdb_file_cmd ${binfile}
>> +
>> +    # Ensure that we are printing 10 lines
>> +    if {![set_listsize 10]} {
>> +	return
>> +    }
>> +
>> +    # First guarantee that GDB prints around the main function correctly
>> +    gdb_test "list ." \
>> +	"1.*\r\n2\[ \t\]+\r\n3\[ \t\]+int main \[)(\]+.*5\[ \t\]+int x;.*" \
>> +	"list . with inferior not running"
>> +
>> +    if {![runto_main]} {
>> +	warning "couldn't start inferior"
>> +	return
>> +    }
>> +
>> +    # Walk forward some lines
> Missing period.
>
>> +    gdb_test "until 15" ".*15.*foo.*"
>> +
>
>> +    # Test that the correct location is printed and that
>> +    # using just "list" will print the following lines.
>> +    gdb_test "list ." ".*" "list current line after starting"
>> +    gdb_test "list" ".*" "confirm we are printing the following lines"
>> +
>> +    # Test that list . will reset to current location
>> +    gdb_test "list ." ".*" "list around current line again"
>> +    gdb_test " " ".*" "testing repeated invocations with GDB's auto-repeat"
> I don't understand -- these 4 tests match ".*", so how how they ensuring what
> they claim they test?  Looks like WIP?
you are correct, it is WIP... taking time off while developing this 
patch was clearly a bad choice hahaha

-- 
Cheers,
Bruno

>
> Pedro Alves
>
>> +}
>> +
>>   clean_restart
>>   gdb_file_cmd ${binfile}
>>   
>> @@ -519,4 +555,7 @@ test_list "list -" 10 2 "7-8" "5-6"
>>   # the current line.
>>   test_list "list -" 10 1 "7" "6"
>>   
>> +# Test printing the location where the inferior is stopped.
>> +test_list_current_location
>> +
>>   remote_exec build "rm -f list0.h"
>> diff --git a/gdb/testsuite/gdb.base/list1.c b/gdb/testsuite/gdb.base/list1.c
>> index d694495c3fb..9297f958f46 100644
>> --- a/gdb/testsuite/gdb.base/list1.c
>> +++ b/gdb/testsuite/gdb.base/list1.c
>> @@ -7,7 +7,7 @@ void bar (int x)
>>      -
>>      - */
>>   {
>> -    printf ("%d\n", x);
>> +    x++;
>>   
>>       long_line ();
>>   }