[KJ] what's the deal with using real "boolean" variables?

[KJ] what's the deal with using real "boolean" variables?

[Robert P. J. Day]

  is there a consensus on how to use actual C99 boolean variables in
the kernel source? C99 defines the unsigned integer type "_Bool" and,
from that point on, it's all downhill:

$ grep -ri "typedef.*bool" .
./drivers/scsi/BusLogic.h:typedef bool boolean;
./drivers/scsi/pci2000.h:typedef        BOOL    *PBOOL;
./drivers/telephony/ixj.h:typedef __u8 BOOL;
./drivers/block/DAC960.h:typedef bool boolean;
./drivers/video/riva/riva_hw.h:typedef int Bool;
./drivers/video/sis/vgatypes.h:typedef unsigned int BOOLEAN;
./arch/parisc/math-emu/float.h:typedef int boolean;
./include/acpi/actypes.h:typedef unsigned char BOOLEAN;
./include/linux/types.h:typedef _Bool                   bool;
./include/linux/efi.h:typedef u8 efi_bool_t;
... snip ...
./fs/xfs/xfs_types.h:typedef enum { B_FALSE,B_TRUE }    boolean_t;

  it seems like everyone has their own idea of what constitutes a
"boolean" variable.  is there a "coding style" suggestion for this?

rday
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] what's the deal with using real "boolean" variables?

[Robert P. J. Day]

On Wed, 10 Jan 2007, Richard Knutsson wrote:

> Robert P. J. Day wrote:
> >   is there a consensus on how to use actual C99 boolean variables in
> > the kernel source? C99 defines the unsigned integer type "_Bool" and,
> > from that point on, it's all downhill:
> >
> <snip>
> > ./include/linux/types.h:typedef _Bool                   bool;
> >
> This is the one that was added to replace the rest.

so the approved way to define boolean variables is with "bool," then?
ok, that makes sense.  i was asking since that would be the obvious
return type of something like "is_power_of_2()", of course.

> There is also 'true' and 'false' defined (enum'ed) in include/linux/stddef.h.
>
> A few file-systems, including ntfs, has been converted to this one.

ah, yes, i see that now.  it might be worth submitting a short patch
to update "CodingStyle" to mention all of this, if this is in fact the
consensus.

rday
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] what's the deal with using real "boolean" variables?

[Richard Knutsson]

Robert P. J. Day wrote:
>   is there a consensus on how to use actual C99 boolean variables in
> the kernel source? C99 defines the unsigned integer type "_Bool" and,
> from that point on, it's all downhill:
>   
<snip>
> ./include/linux/types.h:typedef _Bool                   bool;
>   
This is the one that was added to replace the rest. I have not done too 
much work on this for a while but should take it up soon again (but of 
course, any help is appreciated).

There is also 'true' and 'false' defined (enum'ed) in 
include/linux/stddef.h.

A few file-systems, including ntfs, has been converted to this one.

/Richard Knutsson

_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] what's the deal with using real "boolean" variables?

[Richard Knutsson]

Robert P. J. Day wrote:
> On Wed, 10 Jan 2007, Richard Knutsson wrote:
>
>   
>> Robert P. J. Day wrote:
>>     
>>>   is there a consensus on how to use actual C99 boolean variables in
>>> the kernel source? C99 defines the unsigned integer type "_Bool" and,
>>> from that point on, it's all downhill:
>>>
>>>       
>> <snip>
>>     
>>> ./include/linux/types.h:typedef _Bool                   bool;
>>>
>>>       
>> This is the one that was added to replace the rest.
>>     
>
> so the approved way to define boolean variables is with "bool," then?
> ok, that makes sense.  i was asking since that would be the obvious
> return type of something like "is_power_of_2()", of course.
>   
Sounds good
>   
>> There is also 'true' and 'false' defined (enum'ed) in include/linux/stddef.h.
>>
>> A few file-systems, including ntfs, has been converted to this one.
>>     
>
> ah, yes, i see that now.  it might be worth submitting a short patch
> to update "CodingStyle" to mention all of this, if this is in fact the
> consensus.
>   
Linus and Andrew added some functions whom used bool :)

I started on such a patch for "CodingStyle" but never got it to a 
"ready-to-submit"-state.

/Richard Knutsson

_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] what's the deal with using real "boolean" variables?

[Alexey Dobriyan]

On 1/10/07, Robert P. J. Day <rpjday@mindspring.com> wrote:
> On Wed, 10 Jan 2007, Richard Knutsson wrote:
>
> > Robert P. J. Day wrote:
> > >   is there a consensus on how to use actual C99 boolean variables in
> > > the kernel source? C99 defines the unsigned integer type "_Bool" and,
> > > from that point on, it's all downhill:
> > >
> > <snip>
> > > ./include/linux/types.h:typedef _Bool                   bool;
> > >
> > This is the one that was added to replace the rest.
>
> so the approved way to define boolean variables is with "bool," then?
> ok, that makes sense.  i was asking since that would be the obvious
> return type of something like "is_power_of_2()", of course.
>
> > There is also 'true' and 'false' defined (enum'ed) in
> include/linux/stddef.h.
> >
> > A few file-systems, including ntfs, has been converted to this one.
>
> ah, yes, i see that now.  it might be worth submitting a short patch
> to update "CodingStyle" to mention all of this, if this is in fact the
> consensus.

Before you'll decide whether consesus was reached, read thread
when TRUE and FALSE consolidation patch was first introduced.

Bonus points for getting a warning on this code:

void f(void)
{
        bool foo = 42;
}
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] what's the deal with using real "boolean" variables?

[Richard Knutsson]

Alexey Dobriyan wrote:
> On 1/10/07, Robert P. J. Day <rpjday@mindspring.com> wrote:
>> On Wed, 10 Jan 2007, Richard Knutsson wrote:
>>
>> > Robert P. J. Day wrote:
>> > > is there a consensus on how to use actual C99 boolean variables in
>> > > the kernel source? C99 defines the unsigned integer type "_Bool" 
>> and,
>> > > from that point on, it's all downhill:
>> > >
>> > <snip>
>> > > ./include/linux/types.h:typedef _Bool bool;
>> > >
>> > This is the one that was added to replace the rest.
>>
>> so the approved way to define boolean variables is with "bool," then?
>> ok, that makes sense. i was asking since that would be the obvious
>> return type of something like "is_power_of_2()", of course.
>>
>> > There is also 'true' and 'false' defined (enum'ed) in
>> include/linux/stddef.h.
>> >
>> > A few file-systems, including ntfs, has been converted to this one.
>>
>> ah, yes, i see that now. it might be worth submitting a short patch
>> to update "CodingStyle" to mention all of this, if this is in fact the
>> consensus.
>
> Before you'll decide whether consesus was reached, read thread
> when TRUE and FALSE consolidation patch was first introduced.
You mean the one started by akpm on March 16'th? There were much about 
people who wanted FALSE/TRUE converted to 0/1. But since there is those 
who wants booleans then a generic one is preferable, don't you think?
>
> Bonus points for getting a warning on this code:
>
> void f(void)
> {
> bool foo = 42;
> }
Hint: false is 0, true is !false (so that statement is actually correct. 
But if you check "foo", you will find the value 1 :) )
But we have potential bugs because of some "self-made" booleans. For 
example:

typedef boolean_t int;

#define FALSE 0

#define TRUE  1

...

boolean_t foo = 42;

	if (foo =TRUE) {

	...


/Richard Knutsson

_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] what's the deal with using real "boolean" variables?

[Bernd Petrovitsch]

On Wed, 2007-01-10 at 11:17 +0100, Richard Knutsson wrote:
[...]
> 	if (foo =TRUE) {

This is IMHO ugly for any definition and implementation of boolean in C.
Write it as 
        if (foo) {
Otherwise why don't we write
        if ((foo = TRUE) = TRUE) {
?

	Bernd
-- 
Firmix Software GmbH                   http://www.firmix.at/
mobil: +43 664 4416156                 fax: +43 1 7890849-55
          Embedded Linux Development and Services

_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] what's the deal with using real "boolean" variables?

[walter harms]

hi,
i took a dive into the 'c reference manual' and found that the header stdbool.h
defines (#define) bool/true/false. It is permited to undefine and redefine these
defines. It defines also the macro "__bool_true_false_are_defined".
The best ways seems to emulate this way and provide a kernelbased stdbool.h if
"__bool_true_false_are_defined" does not exist.

my 2 cents,
walter



Richard Knutsson wrote:
> Alexey Dobriyan wrote:
>> On 1/10/07, Robert P. J. Day <rpjday@mindspring.com> wrote:
>>> On Wed, 10 Jan 2007, Richard Knutsson wrote:
>>>
>>>> Robert P. J. Day wrote:
>>>>> is there a consensus on how to use actual C99 boolean variables in
>>>>> the kernel source? C99 defines the unsigned integer type "_Bool" 
>>> and,
>>>>> from that point on, it's all downhill:
>>>>>
>>>> <snip>
>>>>> ./include/linux/types.h:typedef _Bool bool;
>>>>>
>>>> This is the one that was added to replace the rest.
>>> so the approved way to define boolean variables is with "bool," then?
>>> ok, that makes sense. i was asking since that would be the obvious
>>> return type of something like "is_power_of_2()", of course.
>>>
>>>> There is also 'true' and 'false' defined (enum'ed) in
>>> include/linux/stddef.h.
>>>> A few file-systems, including ntfs, has been converted to this one.
>>> ah, yes, i see that now. it might be worth submitting a short patch
>>> to update "CodingStyle" to mention all of this, if this is in fact the
>>> consensus.
>> Before you'll decide whether consesus was reached, read thread
>> when TRUE and FALSE consolidation patch was first introduced.
> You mean the one started by akpm on March 16'th? There were much about 
> people who wanted FALSE/TRUE converted to 0/1. But since there is those 
> who wants booleans then a generic one is preferable, don't you think?
>> Bonus points for getting a warning on this code:
>>
>> void f(void)
>> {
>> bool foo = 42;
>> }
> Hint: false is 0, true is !false (so that statement is actually correct. 
> But if you check "foo", you will find the value 1 :) )
> But we have potential bugs because of some "self-made" booleans. For 
> example:
> 
> typedef boolean_t int;
> 
> #define FALSE 0
> 
> #define TRUE  1
> 
> ...
> 
> boolean_t foo = 42;
> 
> 	if (foo =TRUE) {
> 
> 	...
> 
> 
> /Richard Knutsson
> 
> _______________________________________________
> Kernel-janitors mailing list
> Kernel-janitors@lists.osdl.org
> https://lists.osdl.org/mailman/listinfo/kernel-janitors
> 
> 
> 
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] what's the deal with using real "boolean" variables?

[Richard Knutsson]

Bernd Petrovitsch wrote:
> On Wed, 2007-01-10 at 11:17 +0100, Richard Knutsson wrote:
> [...]
>   
>> 	if (foo =TRUE) {
>>     
>
> This is IMHO ugly for any definition and implementation of boolean in C.
> Write it as 
>         if (foo) {
> Otherwise why don't we write
>         if ((foo = TRUE) = TRUE) {
> ?
>
> 	Bernd
>   
It sure is! But a quick check:
linux-2.6]$ grep -Er "\=\= *(TRUE|true)" * | wc -l
70
linux-2.6]$ grep -Er "\=\= *(FALSE|false)" * | wc -l
118
suggest it is used (saw a few false-positivs).

This is something that should be mentioned in "CodingStyle", but have 
not got there yet. Care to make a draft/patch?

/Richard Knutsson
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] what's the deal with using real "boolean" variables?

[Richard Knutsson]

walter harms wrote:
> hi,
> i took a dive into the 'c reference manual' and found that the header stdbool.h
> defines (#define) bool/true/false. It is permited to undefine and redefine these
> defines. It defines also the macro "__bool_true_false_are_defined".
> The best ways seems to emulate this way and provide a kernelbased stdbool.h if
> "__bool_true_false_are_defined" does not exist.
>   
Why "__bool_true_false_are_defined" when we _know_ it is defined? Since 
2.6.15 (I believe), only gcc >= 3.2 is supported, and from version 3, 
gcc supports _Boole.
I first thought to implement it as stdbool.h, but the consensus seems to 
be: types in linux/types.h and false/true found a home in linux/stddef.h.

/Richard Knutsson

PS
Please don't top-post. :)
DS
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors

Re: [KJ] what's the deal with using real "boolean" variables?

[Robert P. J. Day]

On Wed, 10 Jan 2007, Richard Knutsson wrote:

> Bernd Petrovitsch wrote:
> > On Wed, 2007-01-10 at 11:17 +0100, Richard Knutsson wrote:
> > [...]
> >
> >> 	if (foo =TRUE) {
> >>
> >
> > This is IMHO ugly for any definition and implementation of boolean in C.
> > Write it as
> >         if (foo) {
> > Otherwise why don't we write
> >         if ((foo = TRUE) = TRUE) {
> > ?
> >
> > 	Bernd
> >
> It sure is! But a quick check:
> linux-2.6]$ grep -Er "\=\= *(TRUE|true)" * | wc -l
> 70
> linux-2.6]$ grep -Er "\=\= *(FALSE|false)" * | wc -l
> 118
> suggest it is used (saw a few false-positivs).
>
> This is something that should be mentioned in "CodingStyle", but
> have not got there yet. Care to make a draft/patch?

oh, god ... what have i started *this* time?  :-P

rday
_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors