Re: [PATCH 0/3] checkpatch: Add missing c90 types

[Andy Whitcroft <apw@canonical.com>]

On Thu, Jul 17, 2014 at 08:52:00AM -0700, Joe Perches wrote:
> c90 section "6.7.2 Type Specifiers" says:
>     "type specifiers may occur in any order"
> 
> That means that:
>     short int is the same as int short
>     unsigned short int is the same as int unsigned short
>     etc...
> 
> checkpatch currently parses only a subset of these allowed types.
> 
> For instance: "unsigned short" is a found by checkpatch as a
> specific type, but none of the "signed int" or "int short" variants
> are found.
> 
> Change all the existing types to allow signed and unsigned variants.
> Reorder the existing types array to match longest type first.
> 
> Add another table for the "kernel style misordered" variants.
> 
> Add this misordered table to the findable types.
> 
> Warn when the misordered style is used.
> 
> This improves the "Missing a blank line after declarations" test as
> it depends on the correct parsing of the $Declare variable which
> looks for "$Type $Ident;" (ie: declarations like "int foo;").
> 
> Joe Perches (3):
>   checkpatch: Add short int to c variable types
>   checkpatch: Add signed generic types
>   checkpatch: Add test for native c90 types in unusual order
> 
>  scripts/checkpatch.pl | 61 ++++++++++++++++++++++++++++++++++++++++++++-------
>  1 file changed, 53 insertions(+), 8 deletions(-)

Looks like a sane plan to me.

Acked-by: Andy Whitcroft <apw@canonical.com>

-apw