Re: [xfs-masters] Re: [PATCH 2.6.18-rc4-mm3 2/2] fs/xfs: Converting into generic boolean

Re: [xfs-masters] Re: [PATCH 2.6.18-rc4-mm3 2/2] fs/xfs: Converting into generic boolean

[Nathan Scott]

On Tue, Sep 05, 2006 at 02:47:02PM +0200, Richard Knutsson wrote:
> Just the notion: "your" guys was the ones to make those to boolean(_t), 

Sort of, we actually inherited that type from IRIX where it is
defined in <sys/types.h>.

> and now you seem to want to patch them away because I tried to make them 
> more general.

Nah, I just don't see the value either way, and see it as another
code churn exercise.

> So, is the:
> B_FALSE -> false
> B_TRUE -> true
> ok by you?

Personally, no.  Thats code churn with no value IMO.

> >"int needflush;" is just as readable (some would argue moreso) as
> >"bool needflush;" and thats pretty much the level of use in XFS -
> >  
> How are you sure "needflush" is, for example, not a counter?

Well, that would be named "flushcount" or some such thing.  And you
would be able to tell that it was a counter by the way its used in
the surrounding code.

This discussion really isn't going anywhere useful; I think you need
to accept that not everyone sees value in a boolean type. :)

cheers.

-- 
Nathan

Re: [xfs-masters] Re: [PATCH 2.6.18-rc4-mm3 2/2] fs/xfs: Converting into generic boolean

[Richard Knutsson]

Nathan Scott wrote:

>On Tue, Sep 05, 2006 at 02:47:02PM +0200, Richard Knutsson wrote:
>  
>
>>Just the notion: "your" guys was the ones to make those to boolean(_t), 
>>    
>>
>
>Sort of, we actually inherited that type from IRIX where it is
>defined in <sys/types.h>.
>  
>
Oh, ok

>>>"int needflush;" is just as readable (some would argue moreso) as
>>>"bool needflush;" and thats pretty much the level of use in XFS -
>>> 
>>>      
>>>
>>How are you sure "needflush" is, for example, not a counter?
>>    
>>
>
>Well, that would be named "flushcount" or some such thing.  And you
>would be able to tell that it was a counter by the way its used in
>the surrounding code.
>  
>
True, thinking more of when you have a quick look at the headers, but 
"flushcount" would be a more logical name in such a case.

>This discussion really isn't going anywhere useful; I think you need
>to accept that not everyone sees value in a boolean type. :)
>  
>
Well, can you blame me for trying? ;)
But the more important thing is to clean up the boolean-type and 
FALSE/TRUE mess in the kernel.

>cheers.
>  
>
Thank you for your time and happy coding :)