Re: [PATCH 0/9] powerpc: delete duplicated words

[Joe Perches <joe@perches.com>]

On 2020-07-26 12:08, Randy Dunlap wrote:
> On 7/26/20 10:49 AM, Joe Perches wrote:
>> On Sun, 2020-07-26 at 10:23 -0700, Randy Dunlap wrote:
>>> On 7/26/20 7:29 AM, Christophe Leroy wrote:
>>>> Randy Dunlap <rdunlap@infradead.org> a écrit :
>>>> 
>>>>> Drop duplicated words in arch/powerpc/ header files.
>>>> 
>>>> How did you detect them ? Do you have some script for tgat, or you 
>>>> just read all comments ?
>>> 
>>> Yes, it's a script that finds lots of false positives, so I have to 
>>> check
>>> each and every one of them for validity.
>> 
>> And it's a lot of work too. (thanks Randy)
>> 
>> It could be something like:
>> 
>> $ grep-2.5.4 -nrP --include=*.[ch] '\b([A-Z]?[a-z]{2,}\b)[ \t]*(?:\n[ 
>> \t]*\*[ \t]*|)\1\b' * | \
>>   grep -vP '\b(?:struct|enum|union)\s+([A-Z]?[a-z]{2,})\s+\*?\s*\1\b' 
>> | \
>>   grep -vP '\blong\s+long\b' | \
>>   grep -vP '\b([A-Z]?[a-z]{2,})(?:\t+| {2,})\1\b'
> 
> Hi Joe,

Hi Randy

> (what is grep-2.5.4 ?)

It's the last version of grep that allowed spanning multiple lines.

That's to find the comment second lines that start with *

> It looks like you tried a few iterations of this -- since it drops 
> things
> like "long long".  There are lots of data types that are repeated & 
> valid.
> And many struct names, like "struct kref kref", "struct completion 
> completion",
> and "struct mutex mutex".  I handle (ignore) those manually

that's the first exclude pattern.