[vtk-developers] Some interesting warnings from the clang compiler

[Bill Lorensen]

Sean,

It should be:

 = {space} -
NOT
 =-

Some compilers still treat =- as -=.

It takes an old guy to remember when =- was the preferred way.

Bill

On Thu, Jul 29, 2010 at 5:35 PM, Sean McBride <sean at rogue-research.com> wrote:
> Bill,
>
> The code is currently:
>
>  case VTK_PARSER_UNARY_MINUS:
>   this->Stack[stackPosition] =- this->Stack[stackPosition];
>   break;
>
> Are you saying that it should indeed be "-="?  Otherwise, there's no
> error, only confusing whitespace.
>
> Sean
>
>
> On Thu, 29 Jul 2010 17:27:26 -0400, Bill Lorensen said:
>
>>That's a great warning. Looking at the code, obviously an error. Nice
>>catch clang... I think that code has been wrong since it was written
>>over 11 years ago. Guess it was never tested. Nor has 35% of vtk.
>>
>>Bill
>>
>>On Thu, Jul 29, 2010 at 9:26 AM, Michael Halle <mhalle at bwh.harvard.edu>
> wrote:
>>> Originally, C supported syntax like =- as a synonym for -=, but
>>deprecated it because it was error prone (it relied on whitespace for
>>disambiguation). I think that's what the warning means. Just put spaces
>>around the equals sign.
>>>
>>> -Mike
>>>
>>>
>>>
>>> On Jul 29, 2010, at 8:56 AM, David Gobbi <david.gobbi at gmail.com> wrote:
>>>
>>>> The clang compiler has found some interesting warnings.  Can someone
>>>> look into the vtkFunctionParser one?  I'll deal with the Tcl wrapper
>>>> ones.
>>>>
>>>> vtk-git/Common/vtkFunctionParser.cxx:555:36: warning: use of unary
>>>> operator that may be intended as compound assignment (-=)
>>>>         this->Stack[stackPosition] =- this->Stack[stackPosition];
>>>>
>>>> vtk-git-clang/Common/vtkObjectTcl.cxx:153:27: warning: conversion
>>>> specifies type 'unsigned short' but the argument has type 'int'
>>>> [-Wformat]
>>>>     sprintf(tempResult,"%hu",temp20);
>>>> [ditto for several other tcl-wrapped objects]
>>>>
>>>> Full list here:
>>>>
>>>> http://www.cdash.org/CDash/viewBuildError.php?type=1&buildid=678985
>
>
>