Re: [PING][PATCH] Align natural-format register values to the same column

[Simon Marchi <simon.marchi@polymtl.ca>]

On 2017-07-23 21:57, Ruslan Kabatsayev wrote:
>> If we are going to do the alignment by hand using spaces (which seems 
>> like a good idea, as it gives more control), I'd rather not keep the 
>> tab as an historic artifact.  How many tests would have to be fixed, 
>> and how difficult would it be?
>> 
> I'm not sure how many, because I can't possibly run tests for all the
> architectures. Also, there are many tests with intermittent failures
> (racy), which make counting harder even for a single arch.
> 
> At first I did try to fix things for i386, but after fixing several
> dozens of lines of expected output I realized that it'll not fix
> anything for ARM etc., so I instead went with keeping the tab.

Ok, I see.  If you are willing to fix what you can for x86, we can then 
ask for others to contribute testing for other architectures.  You can 
also make the changes for other architectures (grepping for "info reg" 
in the testsuite), and say that you weren't able to test the changes for 
architecture X and Y.  The changes should be relatively straightforward.

Otherwise, if other maintainers are fine with leaving the tab here, I'm 
not strongly against it.  I would also be fine with a two-step process 
(accept this patch and remote the tab/fix the tests in another patch).

>>>> Currently, commands such as "info reg", "info all-reg", as well as 
>>>> register
>>>> window in the TUI print badly aligned columns, like here:
>>>> 
>>>> eax            0x1      1
>>>> ecx            0xffffd3e0       -11296
>>>> edx            0xffffd404       -11260
>>>> ebx            0xf7fa5ff4       -134586380
>>>> esp            0xffffd390       0xffffd390
>>>> ebp            0xffffd3c8       0xffffd3c8
>>>> esi            0x0      0
>>>> edi            0x0      0
>>>> eip            0x8048b60        0x8048b60 <main+16>
>>>> eflags         0x286    [ PF SF IF ]
>>>> cs             0x23     35
>>>> ss             0x2b     43
>>>> ds             0x2b     43
>>>> es             0x2b     43
>>>> fs             0x0      0
>>>> gs             0x63     99
>>>> 
>>>> After this patch, these commands print the third column values 
>>>> consistently
>>>> aligned one under another, provided the second column is not too 
>>>> long.
>>>> Originally, the third column was (attempted to be) aligned using a 
>>>> simple tab
>>>> character. Lots of tests in the test suite check for it, so this 
>>>> patch retains
>>>> the tab in the output after the second column. This allows these 
>>>> tests to
>>>> continue working unchanged. What is different is that now the tab 
>>>> may be
>>>> followed by several spaces, which complete the task of aligning the 
>>>> third
>>>> column when the sole tab doesn't work well.
>>>> 
>>>> gdb/ChangeLog:
>>>> 
>>>>     * infcmd.c (default_print_one_register_info): Align natural-
>>>>     format column value consistently one after another.
>>>> ---
>>>>  gdb/infcmd.c |   27 ++++++++++++++++++---------
>>>>  1 file changed, 18 insertions(+), 9 deletions(-)
>>>> 
>>>> diff --git a/gdb/infcmd.c b/gdb/infcmd.c
>>>> index defa7b0..5de4e68 100644
>>>> --- a/gdb/infcmd.c
>>>> +++ b/gdb/infcmd.c
>>>> @@ -2280,9 +2280,10 @@ default_print_one_register_info (struct 
>>>> ui_file *file,
>>>>  {
>>>>    struct type *regtype = value_type (val);
>>>>    int print_raw_format;
>>>> +  string_file format_stream;
>>>> 
>>>> -  fputs_filtered (name, file);
>>>> -  print_spaces_filtered (15 - strlen (name), file);
>>>> +  format_stream.puts (name);
>>>> +  format_stream.puts (n_spaces (15 - strlen (name)));
>>>> 
>>>>    print_raw_format = (value_entirely_available (val)
>>>>                && !value_optimized_out (val));
>>>> @@ -2301,14 +2302,18 @@ default_print_one_register_info (struct 
>>>> ui_file *file,
>>>> 
>>>>        val_print (regtype,
>>>>           value_embedded_offset (val), 0,
>>>> -         file, 0, val, &opts, current_language);
>>>> +         &format_stream, 0, val, &opts, current_language);
>>>> 
>>>>        if (print_raw_format)
>>>>      {
>>>> -      fprintf_filtered (file, "\t(raw ");
>>>> -      print_hex_chars (file, valaddr, TYPE_LENGTH (regtype), 
>>>> byte_order,
>>>> +      const int size_with_tab = format_stream.size () / 8 * 8 + 8;
>>>> +      format_stream.putc ('\t');
>>>> +      if (size_with_tab < 32)
>>>> +          format_stream.puts (n_spaces (32 - size_with_tab));
>> 
>> Could you extract that constant (32) to a const variable with some 
>> meaningful name?  If we ever change the width, I predict that we'll 
>> forget to update the instances of the same 32 down there.
> Yeah, I can do this. But should the magic 15 above be named as well
> (as 32 is the second "tab stop" while 15 is the first one)?

I think it would nice too, while at it.

Thanks!

Simon