Re: Bug in AX_APPEND_LINK_FLAGS

On Wednesday 01 January 2014 12:34:20 Peter Simons wrote:
> Hi Mike,
> 
> you submitted the following patch to AX_APPEND_LINK_FLAGS a while ago:
> 
>  
> http://git.savannah.gnu.org/cgit/autoconf-archive.git/commit/m4/ax_append_
> link_flags.m4?id=7d5e9315f7b93843f7d65699eebfec742323a483
> 
> Now, Peter De Wachter reports the following phenomenon:
>   > If you do something like:
>   >     AX_APPEND_LINK_FLAGS([--foo])
>   > 
>   > configure prints a bogus extra line:
>   >     checking whether the linker accepts ... yes
>   >     checking whether the linker accepts --foo... yes
>   > 
>   > This bug is introduced by commit 7d5e9315: serial 3 (the current
>   > version) has this bug, serial 2 not.
> 
> Do you, by any chance, have an idea how to explain and/or remedy this
> behavior?

blah, it's because i misused the AC_REQUIRE helper.  it not only makes sure 
the func is available, but then also calls it.  i don't know of a way to get 
the desired behavior (an m4 error when the macro isn't available) other than 
AC_REQUIRE.

so the options are:
 - have AX_APPEND_FLAG do nothing when $1 is not set
 - revert my change
-mike

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.17 (GNU/Linux)
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=RsET
-----END PGP SIGNATURE-----