RFR: JDK-8033292 configure MAKE=foo causes configure to fail

Dave Pointon dpointo8 at linux.vnet.ibm.com
Tue Feb 4 12:40:01 UTC 2014


On Fri, 2014-01-31 at 10:03 +0100, Magnus Ihse Bursie wrote:
> Bug: https://bugs.openjdk.java.net/browse/JDK-8033292
> 
> With the fix for JDK-8031759, I added a check for unknown variables 
> given on the configure command line.
> 
> Unfortunately, this check required that all known variables used by 
> configure "register" themselves; if this is not done, then the variable 
> is considered unknown. And not all cases of such variable usages was 
> checked by me at that time. :-(
> 
> A prominent example is MAKE.
> 
> While I think the way forward is to properly register all such uses (and 
> probably unify tool detection even more across the line), at the moment, 
> this check should be a warning, not an error.
> 
> This fix is too small for a webrev; please see inlined patch.
> 
> Patch:
> diff --git a/common/autoconf/basics.m4 b/common/autoconf/basics.m4
> --- a/common/autoconf/basics.m4
> +++ b/common/autoconf/basics.m4
> @@ -250,7 +250,7 @@
>     if test "x$CONFIGURE_OVERRIDDEN_VARIABLES" != x; then
>       # Replace the separating ! with spaces before presenting for end user.
>       unknown_variables=${CONFIGURE_OVERRIDDEN_VARIABLES//!/ }
> -    AC_MSG_ERROR([The following variables are unknown to configure: 
> $unknown_variables])
> +    AC_MSG_WARN([The following variables might be unknown to configure: 
> $unknown_variables])
>     fi
>   ])
> 
> 
> /Magnus

Looks good to this non-reviewer.

-- 
Dave Pointon FIAP MBCS

Now I saw, tho' too late, the folly of beginning a work before we count
the cost and before we we judge rightly of our strength to go thro'
with it - Robinson Crusoe




More information about the build-dev mailing list