Turnaround upgrade in mi2

Turnaround upgrade in mi2

[Sean Chen]

Hi,

I compared the two versions of MI (mi1 and mi2) but found no obvious
syntax change except some new commands. I think, personally, if there
is no turnaround upgrade in gdb 6.0, why did we need to accumulate the
version number to mi2? You know, there are also many new commands in
gdb 7.0, but we don’t accumulate the mi version number to mi3 in gdb
7.0.

So I am prone to think I must have missed the turnaround upgrade in
mi2. Could somebody help to clarify? Thanks.

-- 
Best Regards,
Sean Chen

Re: Turnaround upgrade in mi2

[Vladimir Prus]

Sean Chen wrote:

> Hi,
> 
> I compared the two versions of MI (mi1 and mi2) but found no obvious
> syntax change except some new commands. I think, personally, if there
> is no turnaround upgrade in gdb 6.0, 

What is "turnaround upgrade"?

> why did we need to accumulate the
> version number to mi2? You know, there are also many new commands in
> gdb 7.0, but we don’t accumulate the mi version number to mi3 in gdb
> 7.0.
> 
> So I am prone to think I must have missed the turnaround upgrade in
> mi2. Could somebody help to clarify? Thanks.

I am afraid I don't know what are the differences between mi1 and mi2 and I am
not sure it's worth investigating. Is there a specific problem you are
having?

- Volodya


 

Re: Turnaround upgrade in mi2

[Dave Korn]

Vladimir Prus wrote:
> Sean Chen wrote:
> 
>> Hi,
>>
>> I compared the two versions of MI (mi1 and mi2) but found no obvious
>> syntax change except some new commands. I think, personally, if there
>> is no turnaround upgrade in gdb 6.0, 
> 
> What is "turnaround upgrade"?

  I'm inferring it means any kind of back-compat break.

> 
>> why did we need to accumulate the
>> version number to mi2? You know, there are also many new commands in
>> gdb 7.0, but we don’t accumulate the mi version number to mi3 in gdb
>> 7.0.
>>
>> So I am prone to think I must have missed the turnaround upgrade in
>> mi2. Could somebody help to clarify? Thanks.
> 
> I am afraid I don't know what are the differences between mi1 and mi2 and I am
> not sure it's worth investigating. 

  I have a vague memory that the difference is how asynchronous output from
the inferior is presented in the output stream, and that the difference was
enough to make it impossible to unambiguously parse an mi stream without
knowing which one you were dealing with in advance.  Sorry, I have no
reference, but I'm sure there was a sane reason, I think it was to do with the
output format and its parseability, and so it wouldn't have been just because
new commands were added.

    cheers,
      DaveK

Re: Turnaround upgrade in mi2

[Daniel Jacobowitz]

On Sun, Jan 10, 2010 at 06:34:18PM +0000, Dave Korn wrote:
>   I have a vague memory that the difference is how asynchronous output from
> the inferior is presented in the output stream, and that the difference was
> enough to make it impossible to unambiguously parse an mi stream without
> knowing which one you were dealing with in advance.  Sorry, I have no
> reference, but I'm sure there was a sane reason, I think it was to do with the
> output format and its parseability, and so it wouldn't have been just because
> new commands were added.

Yes.  There were also some fixes to existing commands, like using a
tuple instead of a list.  It had nothing to do with new commands.

-- 
Daniel Jacobowitz
CodeSourcery