Re: Expected behavior for -cs and -ncs

Ian Carvalho wrote:
> david ingamells wrote:
>> John E Hein wrote:
>>> david ingamells wrote at 19:42 +0200 on Jul 23, 2008:
>>>  > This works in 2.2.10:
>>>  >  > $ indent -npro -ncs
>>>  >  > unsigned char *cp = (unsigned char *)ptr;
>>>  > int
>>>  > foo ()
>>>  > {
>>>  >   return (int)n;
>>>  > }
>>>  >  > $ indent -npro -cs
>>>  >  > unsigned char *cp = (unsigned char *) ptr;
>>>  > int
>>>  > foo ()
>>>  > {
>>>  >   return (int) n;
>>>  > }
>>>  >  >  > I know of no update that fixed this in 2.2.10, so I am 
>>> surprised that it  > doesn't work in 2.2.9.
>>>  > I have added this to the regression test now, so it will be 
>>> checked in  > the future.
>>>
>>> Seems okay in my local test here on 2.2.9, too...
>>>
>>>
>>> echo '(unsigned char *)ptr;' | gindent -npro -cs
>>> (unsigned char *) ptr;
>>>
>>> echo '(unsigned char *)ptr;' | gindent -npro -ncs
>>> (unsigned char *)ptr;
>>>
>>> However, --no-space-after-cast fails:
>>>
>>>
>>> echo '(unsigned char *)ptr;' | gindent -npro --space-after-cast
>>> (unsigned char *) ptr;
>>>
>>> echo '(unsigned char *)ptr;' | gindent -npro --no-space-after-cast
>>> indent: unknown option "no-space-after-cast"
>>>
>>>
>>>   
>> A quick check in the code explains why:
>>
>> {"no-space-after-casts",                        "ncs"},
>>
>> And the manual says the same. I think what is best is that I update 
>> the manual to remove the trailing "s" and recognise both forms in the 
>> code.
>> That way users will move over to the right one and existing users 
>> won't get a nasty surprise.
>>
> On seeing your responses, went through the code I am trying to 
> indent.  Looks like the example I have given is flawed.
>
> When the cast happens with a typedef'd variable then it does not seem 
> to work.
>
> Using this:
> --- start ---
> typedef unsigned char   uchar;
>
> unsigned char *cp = (unsigned char *)ptr;
> uchar *ucp = (uchar *)ptr;
> --- end --
> as the source
>
> $ indent -npro -ncs
> I get
> --- start ---
> typedef unsigned char   uchar;
>
> unsigned char *cp = (unsigned char *)ptr;
> uchar *ucp = (uchar *) ptr;
> --- end ---
>
> and for
>
> $ indent -npro -cs
> I get
> --- start ---
> typedef unsigned char   uchar;
>
> unsigned char *cp = (unsigned char *) ptr;
> uchar *ucp = (uchar *) ptr;
> --- end ---
>
> Thus, the space is present in both cases if the cast is to a typedef'd 
> type.
>
> Is this the expected behavior?
>
> Thanks
>
>
>
Yes that is expected. You need  -Tuchar