[alsa-devel] [PATCH] alsamixer: define _GNU_SOURCE to get exp10

[Jaroslav Kysela]

Dne 17.1.2019 v 01:16 makepost at firemail.cc napsal(a):
> These flags vary between systems, on the Gentoo where I have access the 
> Makefile generated from e1aa6d4 contains:
> 
> $ ack '^(AM|CURSES|NCURSESW)_CFLAGS' alsamixer/Makefile
> CURSES_CFLAGS = -I/usr/include/ncursesw
> NCURSESW_CFLAGS = -I/usr/include/ncursesw
> AM_CFLAGS = -I/usr/include/ncursesw  -DCURSESINC="<ncurses.h>"
> 
> Files in /usr/include/ncursesw don't mention _GNU_SOURCE so I'm baffled 
> where GCC gets it from.
> 
> Indeed the local alsamixer/Makefile.am is a more suitable place to put 
> the flag. A newbie question follows, should I somehow update the patch 
> with your edit and reply to this message, or did you generate the diff 
> from an already edited version?

Applied to the repository:

http://git.alsa-project.org/?p=alsa-utils.git;a=commitdiff;h=116488e5f2f1b897084bd151381ee254e1cc177d

				Thanks,
					Jaroslav

> 
> On 2019-01-16 14:22, Jaroslav Kysela wrote:
>> Dne 16.1.2019 v 02:35 makepost at firemail.cc napsal(a):
>>> Fixes Master, Headphones and Speaker stuck at -8 percentage after
>>> building with Clang 7.0.1 and getting warned about implicit 
>>> declaration
>>> of exp10, which is behind _GNU_SOURCE as a non-standard feature. 
>>> Thanks
>>> Takashi Iwai for the CFLAGS suggestion on the mailing list. GCC build 
>>> is
>>> not affected, so not adding a compiler check to the configure script.
>>>
>>> uClibc-ng has got exp10 since 1.0.12, so the fallback macro is no 
>>> longer
>>> needed. However, alsa-utils relies on gettext so might need further
>>> patches to actually work on uClibc systems.
>>>
>>> Signed-off-by: makepost <makepost at firemail.cc>
>>>
>>> diff --git a/alsamixer/volume_mapping.c b/alsamixer/volume_mapping.c
>>> index 94bd0fe..48cfbe2 100644
>>> --- a/alsamixer/volume_mapping.c
>>> +++ b/alsamixer/volume_mapping.c
>>> @@ -36,11 +36,6 @@
>>>   #include <stdbool.h>
>>>   #include "volume_mapping.h"
>>>
>>> -#ifdef __UCLIBC__
>>> -/* 10^x = 10^(log e^x) = (e^x)^log10 = e^(x * log 10) */
>>> -#define exp10(x) (exp((x) * log(10)))
>>> -#endif /* __UCLIBC__ */
>>> -
>>>   #define MAX_LINEAR_DB_SCALE    24
>>>
>>>   static inline bool use_linear_dB_scale(long dBmin, long dBmax)
>>> diff --git a/configure.ac b/configure.ac
>>> index 7938996..0d54942 100644
>>> --- a/configure.ac
>>> +++ b/configure.ac
>>> @@ -350,6 +350,8 @@ if test x$alsamixer = xtrue; then
>>>
>>>
>>>     AC_MSG_RESULT([$CURSESLIB])
>>> +
>>> +  CFLAGS="$CFLAGS -D_GNU_SOURCE"
>>
>> It would be propably better to put this to the local makefile for 
>> alsamixer:
>>
>> diff --git a/alsamixer/Makefile.am b/alsamixer/Makefile.am
>> index 259b29e..249e8f8 100644
>> --- a/alsamixer/Makefile.am
>> +++ b/alsamixer/Makefile.am
>> @@ -1,4 +1,4 @@
>> -AM_CFLAGS = @CURSES_CFLAGS@ -DCURSESINC="@CURSESINC@"
>> +AM_CFLAGS = -D_GNU_SOURCE @CURSES_CFLAGS@ -DCURSESINC="@CURSESINC@"
>>  LDADD = @CURSESLIB@
>>
>>  bin_PROGRAMS = alsamixer
>>
>> The reason, why the problem is not visible with the standard
>> gcc/glibc/ncurses is that this definition is activated by ncurses:
>>
>> $ find alsamixer -type f -exec grep -H GNU_SOURCE {} \;
>> alsamixer/Makefile:CURSES_CFLAGS = -D_GNU_SOURCE -D_DEFAULT_SOURCE
>> alsamixer/Makefile:NCURSESW_CFLAGS = -D_GNU_SOURCE -D_DEFAULT_SOURCE
>> alsamixer/Makefile:AM_CFLAGS = -D_GNU_SOURCE -D_DEFAULT_SOURCE
>> -DCURSESINC="<ncurses.h>"
>>
>> 					Jaroslav


-- 
Jaroslav Kysela <perex at perex.cz>