On 29 May 2013 21:09, Igor Stasenko <[email protected]> wrote:
> On 29 May 2013 20:55, Frank Shearar <[email protected]> wrote:
>> On 29 May 2013 19:43, Esteban A. Maringolo <[email protected]> wrote:
>>> I see no harm in that.
>>>
>>> It might be misleading,
>>
>> Well, that's _exactly_ the harm in that.
>>
> so don't write it like that :)

Sure, but that's the inverse of the problem. The problem isn't that
_you_ can write a comment anywhere. It's that when you're reviewing my
code _I_ can put a comment anywhere, and make your day that little bit
worse. One answer to that is a lint rule, I suppose, and then you
reject my submission because I did some crazy nonsense with comments.

> I actually like the fact that i can put comment anywhere i want, and
> it doesn't disrupts the parser.
> Of course, it turns to be a problem, when you think that something
> must be represented by a single token
> but it cannot, because it split by comment.

Exactly, which makes parsing that little bit worse.

> But to me it is much more important that i can put comment anywhere i
> like, without thinking "oh.. it may not compile".

I suppose that's fine for hacking around. I just wouldn't want to
review code with these kinds of comments in it. QA's hard enough
already.

frank

>> frank
>>
>>> but who in his/her sane mind would write like that?
>>
>>> Apparently the parser removes all the comments before doing anything else.
>>>
>>> Esteban A. Maringolo
>>>
>>>
>>> 2013/5/29 Nicolas Cellier <[email protected]>
>>>>
>>>> Ah, nice, we can write in french ;) 221" millions "355" mille "799
>>>>
>>>>
>>>> 2013/5/29 Gisela Decuzzi <[email protected]>
>>>>>
>>>>>
>>>>> 2013/5/29 Frank Shearar <[email protected]>
>>>>>>
>>>>>> How does 1"sdads"2 evaluate? To 12?
>>>>>
>>>>>
>>>>> Yes the same... 1"sdads"2 is interpreted as 12 and all estrange
>>>>> combinations...
>>>>>
>>>>> I think this shouldn't be valid but since there are tests verifying the
>>>>> behavior maybe there was a reason (more than the scanner we are using is
>>>>> skipping all the comments without letting the parser decide)
>>>>>
>>>>
>>>
>>
>
>
>
> --
> Best regards,
> Igor Stasenko.
>

Reply via email to