PR35503 - warn for restrict pointer

[Prathamesh Kulkarni]

On 20 September 2016 at 08:57, Martin Sebor <msebor@gmail.com> wrote:
>> and used
>> pp_format for formatting arg_positions by adding specifier %I (name
>> chosen arbitrarily).
>> Does that look OK ?
>
>
> diff --git a/gcc/pretty-print.c b/gcc/pretty-print.c
> index a39815e..e8bd1ef 100644
> --- a/gcc/pretty-print.c
> +++ b/gcc/pretty-print.c
> @@ -610,6 +610,23 @@ pp_format (pretty_printer *pp, text_info *text)
>               (pp, *text->args_ptr, precision, unsigned, "u");
>           break;
>
> +       case 'I':
>
> The comment just above pp_format that lists the format specifiers
> understood by the function should be updated.  There probably (I
> hope) is more formal documentation of the format specifiers
> somewhere else that should be updated as well.
>
> That said, since this specifier formats a vec<int>, it seems that
> it might be useful to be able to format vectors of other elements,
> such as short and long.  With that in mind, would adding a new V
> length modifier instead be a more regular way to extend the pretty
> printer to these types?  The V length modifier would have to be
> accepted in conjunction with other length modifiers (h, l, etc
> and type specifiers (d, i, o, etc.).  E.g, "%hVu" would accept
> a vec<unsigned short>*  as an argument and format it as a sequence
> of decimal numbers, and "%lVx" would do the same for vec<long>*
> formatting them in hex.
Hi,
Thanks everyone for the suggestions.

This patch does the following:
a) adds warning_at_rich_loc_n(). Somehow I missed David's clear explanation of
warning_at_rich_loc_n earlier :(

b) clears TREE_VISITED flag in C/C++ FE after we are done with
warn_for_restrict().
I assumed that TREE_VISITED would be treated as a "scratch" flag and any pass
wishing to use it must set it first, so didn't bother clearing at first.

c) It appears '%I' is already used for some format specifier in
c-family/c-format.c ?
I changed name to %Z instead. Martin suggested a better idea above to
implement %Z as a length modifier so we can extend it to print vec of other
types, like %hZx would print vec<short> formatted in hex.
I am not sure though if it would be a simple fix, and left it as a FIXME
in the patch adding %Z just to print vec<int>, maybe we could revisit it later ?
Could someone please take a look at the change to c-format.c, I am not sure
if I have added that correctly.

Thanks,
Prathamesh
>
> Martin
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pr35503-9.diff
Type: text/x-patch
Size: 13921 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20160920/a676c513/attachment.bin>