Re: [PATCH 1/4] Change field separator in gdbarch.sh

[Pedro Alves <palves at redhat dot com>]

On 04/19/2017 05:03 AM, Simon Marchi wrote:
> On 2017-04-18 16:20, Pedro Alves wrote:
>> On 04/16/2017 03:14 PM, Simon Marchi wrote:
>>> --- a/gdb/gdbarch.h
>>> +++ b/gdb/gdbarch.h
>>> @@ -129,7 +129,7 @@ extern void set_gdbarch_bits_big_endian (struct
>>> gdbarch *gdbarch, int bits_big_e
>>>
>>>  /* Number of bits in a char or unsigned char for the target machine.
>>>     Just like CHAR_BIT in <limits.h> but describes the target machine.
>>> -   v:TARGET_CHAR_BIT:int:char_bit::::8 * sizeof (char):8::0:
>>> +   v;TARGET_CHAR_BIT;int;char_bit;;;;8 * sizeof (char);8;;0;
>>
>> BTW, do you know why does this bit appear in the .h file in the
>> first place?
> 
> It's in gdbarch.h because gdbarch.sh thinks it's part of the comment
> above short_bit, so it gets copied verbatim to the .h.  

Ah, it's commented out in gdbarch.sh.  I hadn't looked, and confused it
with the gdbarch hook you had added before.  But this is something much
older than that.

> I don't know why
> it's there though.  It seems like it's been there for a really long
> time, but has always been commented out.  I had not seen it before doing
> this patch, and didn't worry too much about it.
> 
> Perhaps it was seen as a possible extension in case someone would bother
> adding support for targets with non-8-bit-bytes?  

Yeah, that looks like it.


> If so, the addressable_memory_unit_size method kind of took that place.

Guess we should just delete that commented out hook.

> 
> Any better guess?

No.

Thanks,
Pedro Alves