r180982 - Attempt to un-break the gdb buildbot.

[Eric Christopher]

Yes, the commented code.

And yeah the style guide says to prefer preincrement. :)

-eric

On Tue, May 7, 2013 at 3:38 PM, Adrian Prantl <aprantl at apple.com> wrote:
> Oh! You were talking about the old, commented out code. All this time I thought the second comment referred to the quoted block below it.
>
> As for the increment -- I found the preincrement slightly more ugly than the increment assignment. No strong feelings about it, though. Feel free to change it.
>
> thanks,
> Adrian
>
> On May 7, 2013, at 3:29 PM, Eric Christopher <echristo at gmail.com> wrote:
>
>> Was more in the "incrementing and testing in an if statement" is
>> somewhat error prone. I mean, why not just increment and then check?
>>
>> Also why no preincrements for the increments by 1?
>>
>> -eric
>>
>>
>> On Tue, May 7, 2013 at 3:24 PM, Adrian Prantl <aprantl at apple.com> wrote:
>>> I agree about the commented out code, but what’s the story about the side effects? -- isEvaluatable is const, right?
>>>
>>> On May 7, 2013, at 3:19 PM, Eric Christopher <echristo at gmail.com> wrote:
>>>
>>>>> -    if (++NumStopPoints == 1)
>>>>> -      FirstStopPoint = Loc;
>>>>> +    //if (++NumStopPoints == 1)
>>>>> +      LastStopPoint = Loc;
>>>>>
>>>>
>>>> Eh? Commented out code? Also, side effects in an if conditional? Ew.
>>>>
>>>>> +  NumReturnExprs += 1;
>>>>>  if (RV == 0 || RV->isEvaluatable(getContext()))
>>>>> -    ++NumSimpleReturnExprs;
>>>>> +    NumSimpleReturnExprs += 1;
>>>>>
>>>>
>>>> Eh?
>>>>
>>>> -eric
>