Re: Re: [PATCH 2/5] perf: Reorder parameters of strglobmatch

[Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>]

(2013/05/21 18:19), Arnaldo Carvalho de Melo wrote:
> Em Fri, May 17, 2013 at 11:21:11AM +0900, Masami Hiramatsu escreveu:
>> (2013/05/16 23:55), Steven Rostedt wrote:
>>> I'm a bit confused to the rational here. Can you explain in more detail
>>> to why this patch is actually needed?
>  
>> Yes, actually, this patch is not needed from the viewpoint of execution,
>> but less misuse for future use of the strglobmatch, I think.
>  
>> For example, glob(3) has the pattern parameter as the first one,
>>
>>        int glob(const char *pattern, int flags,
>>                 int (*errfunc) (const char *epath, int eerrno),
>>                 glob_t *pglob);
>  
>> regexec(3) also has the compiled regexp at the first parameter,
>  
>>        int regexec(const regex_t *preg, const char *string, size_t nmatch,
>>                    regmatch_t pmatch[], int eflags);
>  
>> Thus, I think a new user of strglobmatch() may guess that the first
>> parameter should be the glob pattern.
>  
>> So, this patch is not technically needed, but from the viewpoint of coding
>> naturally, it should be changed, IMHO.
> 
> So I suggest you state this in the changeset comment :-)

Agreed, I must update it ...

Thank you,
-- 
Masami HIRAMATSU
IT Management Research Dept. Linux Technology Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt@hitachi.com