Patchwork Ping ... Re: [Patch, test-suite, plugins] include build_dir/intl to enable finding libintl.h on systems that build it.

login
register
mail settings
Submitter IainS
Date July 11, 2010, 7:02 p.m.
Message ID <C7B2EF24-D619-4667-8B0D-925934EADE98@sandoe-acoustics.co.uk>
Download mbox | patch
Permalink /patch/58537/
State New
Headers show

Comments

IainS - July 11, 2010, 7:02 p.m.
Hi,

This has been hanging around my local tree for  a couple of months..
.. and Jack and I realized that it hadn't been approved when we were  
considering making the (approved) back-port of the plug-in enabler to  
4.5.

In Janis' absence would one of the global reviewers please take a look?
thanks
Iain

On 12 May 2010, at 17:26, IainS wrote:

> Hi,
>
> On  *-apple-darwin9 most of the plugin tests fail with errors like:
> " /GCC/gcc-live-trunk/gcc/testsuite/../../gcc/intl.h:31:21: fatal  
> error: libintl.h: No such file or directory"
>
> libintl.h is built and present in ${gcc_objdir}/intl.
>
> the patch below adds that directory to the plugins testsuite includes,
> OK for trunk?
>
> Iain
>
> gcc/testsuite/Changelog:
>
> 	* lib/plugin-support.exp (plugin-test-execute):  Add ${gcc_objdir}/ 
> intl to includes.
>

refreshed diff:

      set optstr "$includes $extra_flags -DIN_GCC -fPIC -shared"
Mike Stump - July 11, 2010, 8:11 p.m.
On Jul 11, 2010, at 12:02 PM, IainS wrote:
> This has been hanging around my local tree for  a couple of months..
> .. and Jack and I realized that it hadn't been approved when we were considering making the (approved) back-port of the plug-in enabler to 4.5.
> 
> In Janis' absence would one of the global reviewers please take a look?

The important issues are the spelling, and that looks good to me, and the ordering...  having this last should be very safe.  I'd normally just check this in as obvious as it makes testing a new feature better on all systems that don't have an installed intl.h.
IainS - July 16, 2010, 10:42 a.m.
On 11 Jul 2010, at 21:11, Mike Stump wrote:

> On Jul 11, 2010, at 12:02 PM, IainS wrote:
>> This has been hanging around my local tree for  a couple of months..
>> .. and Jack and I realized that it hadn't been approved when we  
>> were considering making the (approved) back-port of the plug-in  
>> enabler to 4.5.
>>
>> In Janis' absence would one of the global reviewers please take a  
>> look?
>
> The important issues are the spelling, and that looks good to me,  
> and the ordering...  having this last should be very safe.  I'd  
> normally just check this in as obvious as it makes testing a new  
> feature better on all systems that don't have an installed intl.h.

checked in as, r162259
Iain

Patch

Index: gcc/testsuite/lib/plugin-support.exp
===================================================================
--- gcc/testsuite/lib/plugin-support.exp	(revision 162059)
+++ gcc/testsuite/lib/plugin-support.exp	(working copy)
@@ -84,7 +84,7 @@  proc plugin-test-execute { plugin_src plugin_tests
      set gcc_objdir "$objdir/../../.."
      set includes "-I. -I${srcdir} -I${gcc_srcdir}/gcc -I$ 
{gcc_objdir}/gcc \
                    -I${gcc_srcdir}/include -I${gcc_srcdir}/libcpp/ 
include \
-                  $GMPINC"
+                  $GMPINC -I${gcc_objdir}/intl "