[PATCH][au1000] Remove useless EXTRA_CFLAGS

[PATCH][au1000] Remove useless EXTRA_CFLAGS

[Florian Fainelli]

Remove the EXTRA_CFLAGS because it is useless (code compiles without warnings).

Signed-off-by: Florian Fainelli <florian.fainelli@telecomint.eu>
---
diff --git a/arch/mips/au1000/common/Makefile b/arch/mips/au1000/common/Makefile
index 90e2d7a..4c35525 100644
--- a/arch/mips/au1000/common/Makefile
+++ b/arch/mips/au1000/common/Makefile
@@ -12,5 +12,3 @@ obj-y += prom.o irq.o puts.o time.o reset.o \
 
 obj-$(CONFIG_KGDB)		+= dbg_io.o
 obj-$(CONFIG_PCI)		+= pci.o
-
-EXTRA_CFLAGS += -Werror

Re: [PATCH][au1000] Remove useless EXTRA_CFLAGS

[Manuel Lauss]

Florian Fainelli schrieb:
> Remove the EXTRA_CFLAGS because it is useless (code compiles without warnings).

then where's the harm in leaving it in? It only forces you to think twice
(or apply tons of casts) when working with the au1x code.

>  obj-$(CONFIG_PCI)		+= pci.o
> -
> -EXTRA_CFLAGS += -Werror

MfG,
	Manuel Lauss

Re: [PATCH][au1000] Remove useless EXTRA_CFLAGS

[Florian Fainelli]

Hi Manuel,

Le jeudi 25 octobre 2007, Manuel Lauss a écrit :
> then where's the harm in leaving it in? It only forces you to think twice
> (or apply tons of casts) when working with the au1x code.

II could argue the same way, why not :p ?

If your code has warning, you are very likely not to like the compiler stop on 
it, but rather warn. Also if there are casts or any special conversion, 
sparse checking will do a better job as far as I can tell.

Re: [PATCH][au1000] Remove useless EXTRA_CFLAGS

[Manuel Lauss]

Florian Fainelli schrieb:
> Le jeudi 25 octobre 2007, Manuel Lauss a écrit :
>> then where's the harm in leaving it in? It only forces you to think twice
>> (or apply tons of casts) when working with the au1x code.
> 
> II could argue the same way, why not :p ?

Hehe ;)

> If your code has warning, you are very likely not to like the compiler stop on 
> it, but rather warn. Also if there are casts or any special conversion, 
> sparse checking will do a better job as far as I can tell.
> 

I tripped over it once or twice; but in the end I don't really whether
this stays or goes (not that it matters anyway)

Thanks,
	Manuel Lauss

Re: [PATCH][au1000] Remove useless EXTRA_CFLAGS

[Ralf Baechle]

On Thu, Oct 25, 2007 at 09:08:42PM +0200, Florian Fainelli wrote:

> Remove the EXTRA_CFLAGS because it is useless (code compiles without warnings).

And to ensure it will stay that way I'll keep -Werror.  It seems only
little PITAs like this keep everybody on their toes :-)

  Ralf

Re: [PATCH][au1000] Remove useless EXTRA_CFLAGS

[Maciej W. Rozycki]

On Mon, 29 Oct 2007, Ralf Baechle wrote:

> And to ensure it will stay that way I'll keep -Werror.  It seems only
> little PITAs like this keep everybody on their toes :-)

 Yeah...  If only GCC had no bugs and always had a clue of what to warn 
about...

  Maciej

Re: [PATCH][au1000] Remove useless EXTRA_CFLAGS

[Ralf Baechle]

On Wed, Nov 07, 2007 at 12:41:29PM +0000, Maciej W. Rozycki wrote:

> > And to ensure it will stay that way I'll keep -Werror.  It seems only
> > little PITAs like this keep everybody on their toes :-)
> 
>  Yeah...  If only GCC had no bugs and always had a clue of what to warn 
> about...

The least of all problems.

As of 2.6.20-rc2-git2 we had 137 warnings in MIPS specific code of a total
of 218 for the kernel and 44 for modules - that's a insane 52%.  And people
happily added more crappy code because they didn't not even _notice_ the
warnings some of which indeed were bugs.

Compare to 2.6.23 - 13 warnings for MIPS specific code of a total of 48
warnings and another 23 warnings for all the modules.  For a codebase
which overall had grown by half a million lines between those releases.

  Ralf

Re: [PATCH][au1000] Remove useless EXTRA_CFLAGS

[Maciej W. Rozycki]

On Thu, 8 Nov 2007, Ralf Baechle wrote:

> >  Yeah...  If only GCC had no bugs and always had a clue of what to warn 
> > about...
> 
> The least of all problems.

 It depends for whom I suppose. ;-)  It is a pain when you have to stretch 
your imagination to rewrite a piece of source code GCC warns about unduly 
and both keep it readable and not make the generated binary worse...  
Especially where configuration-specific macros are involved.

> As of 2.6.20-rc2-git2 we had 137 warnings in MIPS specific code of a total
> of 218 for the kernel and 44 for modules - that's a insane 52%.  And people
> happily added more crappy code because they didn't not even _notice_ the
> warnings some of which indeed were bugs.

 Well, adding no warnings should be the rule #1 and I can understand it is 
easier for you to enforce it by the means of -Werror. ;-)

  Maciej