[PATCH v3 0/3] kconfig: enhance parsing and operator recognition

[PATCH v3 0/3] kconfig: enhance parsing and operator recognition

[Jan Beulich]

There being at least one case (DEBUG_UART_8250_WORD in
arch/arm/Kconfig.debug) where the utility silently behaves other than
what the programmer expects tells us that (a) we shouldn't silently
ignore unrecognized characters and (b) it's not just me always having
wanted to be able to use comparisons other than (in)equality ones.

1: trivial: sound/soc/intel/Kconfig: use single backslash for line continuation
2: kconfig: don't silently ignore unhandled characters
3: kconfig: allow use of relations other than (in)equality

Signed-off-by: Jan Beulich <jbeulich@suse.com>
---
v3: Split off sound/soc/intel/Kconfig change from patch 1 (now 2).
v2: printf() -> fprintf(stderr, ) upon Paul Bolle's request. Replace
    two instances of double backslash line continuations with single
    ones in sound/soc/intel/Kconfig. Drop stray debugging printf()s
    from patch 2.

Re: [PATCH v3 0/3] kconfig: enhance parsing and operator recognition

[Paul Bolle]

On Tue, 2015-01-20 at 12:51 +0000, Jan Beulich wrote:
> There being at least one case (DEBUG_UART_8250_WORD in
> arch/arm/Kconfig.debug) where the utility silently behaves other than
> what the programmer expects tells us that (a) we shouldn't silently
> ignore unrecognized characters and (b) it's not just me always having
> wanted to be able to use comparisons other than (in)equality ones.
> 
> 1: trivial: sound/soc/intel/Kconfig: use single backslash for line continuation

This turned out to be already fixed by commit 34084a436703 ("ASoC:
intel: Remove superfluous backslash in Kconfig"), which was added in
v4.0-rc1.

> 2: kconfig: don't silently ignore unhandled characters
> 3: kconfig: allow use of relations other than (in)equality
> 
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

What's the status on 2: and 3:? Is everyone waiting on someone else to
do something? (One thing I remember is that Michal wanted the changes to
the _shipped files separately. Is that correct?)

For my part I still think 2: should go in.

And for 3: I would like to see related changes to Kconfig files
submitted too. But, of course, whether that is actually needed is mostly
Michal's call.

> v3: Split off sound/soc/intel/Kconfig change from patch 1 (now 2).
> v2: printf() -> fprintf(stderr, ) upon Paul Bolle's request. Replace
>     two instances of double backslash line continuations with single
>     ones in sound/soc/intel/Kconfig. Drop stray debugging printf()s
>     from patch 2.


Paul Bolle

Re: [PATCH v3 0/3] kconfig: enhance parsing and operator recognition

[Jan Beulich]

>>> On 12.03.15 at 13:34, <pebolle@tiscali.nl> wrote:
> On Tue, 2015-01-20 at 12:51 +0000, Jan Beulich wrote:
>> 2: kconfig: don't silently ignore unhandled characters
>> 3: kconfig: allow use of relations other than (in)equality
>> 
>> Signed-off-by: Jan Beulich <jbeulich@suse.com>
> 
> What's the status on 2: and 3:? Is everyone waiting on someone else to
> do something? (One thing I remember is that Michal wanted the changes to
> the _shipped files separately. Is that correct?)

Since I don't recall any requests for changes mandatory for any
of this to be accepted, I didn't plan on re-submitting unless other
changes would require the patches to be refreshed in non-trivial
ways.

> For my part I still think 2: should go in.
> 
> And for 3: I would like to see related changes to Kconfig files
> submitted too. But, of course, whether that is actually needed is mostly
> Michal's call.

Not sure what related changes to refer to. As it stands, there is
at least one use already in the tree that this change fixes, so I
didn't think there was an immediate further need for actual uses
thereof. And of course I'm hesitant to invest work in using these
new capabilities elsewhere when I can't be certain the base
changes are going to be accepted.

Jan