Re: [PATCH 2.6.18-rc4-mm2] fs/jfs: Conversion to generic boolean

[Dave Kleikamp <shaggy@austin.ibm.com>]

On Tue, 2006-08-29 at 01:33 +0200, Richard Knutsson wrote:
> Dave Kleikamp wrote:
> 
> >On Mon, 2006-08-28 at 22:42 +0200, Richard Knutsson wrote:
> >
> >  
> >
> >>Just why is it, that when there is a change to make locally defined 
> >>booleans into a more generic one, it is converted into integers? ;)
> >>    
> >>
> >
> >I just see this as an opportunity to make jfs more closely fit the
> >coding style of the mainline kernel.
> >  
> >
> That is what I am trying to do, making bool as accepted as any other 
> integer. No more, no less.

Okay.  My initial impression is that you were just offended by the
ugliness of having so many different definitions of true, false, and
boolean types.

> 
> >  
> >
> >>But seriously, what is gained by removing them, other then less 
> >>understandable code? (Not talking about FALSE -> 0, but boolean_t -> int)
> >>    
> >>
> >
> >I don't feel strongly one way or another about the use of boolean_t, but
> >under fs/, the only code that uses that type is in fs/jfs and fs/xfs,
> >which are both ported from other operating systems.  Using ints for
> >boolean values does seem to be the accepted practice in the kernel.
> >  
> >
> Yes it is, but I am (for now) trying to convert those who uses some sort 
> of boolean to the generic one (in fs/ for now). Right now the ntfs/- and 
> partitions/-conversion seem to have thumbs up, in -mm.
> 
> >  
> >
> >>I can understand if authors disprove making an integer into a boolean, 
> >>but here it already were booleans.
> >>But hey, you are the maintainer ;)
> >>    
> >>
> >
> >I could be persuaded to leave the declarations as boolean_t or even
> >making them bool, but right now I'm leaning toward making them int for
> >consistency.
> >  
> >
> A root-beer maybe?

heh

> What do you say, can you hold on it for a while (can't be urgent, can 
> it?) and see how the conversion go. Will take time for it during this 
> week(end) and if the result is that almost no maintainer wants it, then...
> Just seem strange to having a boolean function but declaring it integer, 
> for (in my knowledge) no reason.

Sounds good to me.  I think I'll go ahead and kill the use of TRUE and
FALSE, but hold off on the type change for now.

-- 
David Kleikamp
IBM Linux Technology Center

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/