Patchwork PATCH RFA: libstdc++ lives in src/.libs

login
register
mail settings
Submitter Ian Taylor
Date Oct. 18, 2010, 8:08 p.m.
Message ID <AANLkTiniynJdbRegwzuAtXiOURWnkc4k97AFpMj4+E8X@mail.gmail.com>
Download mbox | patch
Permalink /patch/68236/
State New
Headers show

Comments

Ian Taylor - Oct. 18, 2010, 8:08 p.m.
Ping for build maintainers.

Ian

On Thu, Oct 7, 2010 at 4:54 PM, Ian Lance Taylor <iant@google.com> wrote:
> In the top level Makefile.def file, the lib_path of a target module
> gives the directory where the library can be found at run time.  This is
> used to set LD_LIBRARY_PATH (or equivalent) when bootstrapping.  The
> value for libstdc++-v3 is .libs, but actually the library is built in
> src/.libs.  This patch corrects this error.
>
> Bootstrapped on x86_64-unknown-linux-gnu.  Passed the libstdc++-v3 and
> g++ testsuites.
>
> OK for mainline?
>
> Ian
>
>
> 2010-10-07  Ian Lance Taylor  <iant@google.com>
>
>        * Makefile.def (target_modules): Set lib_path to src/.libs for
>        libstdc++-v3 module.
>        * Makefile.tpl: Fix typo in TARGET_LIB_PATH comment.
>        * Makefile.in: Rebuild.
>
>
>
Ralf Wildenhues - Oct. 20, 2010, 5:24 a.m.
Hello Ian,

* Ian Lance Taylor wrote on Mon, Oct 18, 2010 at 10:08:21PM CEST:
> Ping for build maintainers.

I cannot approve this but I think this can be considered as obvious.
One question is why it didn't show up earlier.

Cheers,
Ralf

> On Thu, Oct 7, 2010 at 4:54 PM, Ian Lance Taylor <iant@google.com> wrote:
> >        * Makefile.def (target_modules): Set lib_path to src/.libs for
> >        libstdc++-v3 module.
> >        * Makefile.tpl: Fix typo in TARGET_LIB_PATH comment.
> >        * Makefile.in: Rebuild.

> --- Makefile.def	(revision 165058)
> +++ Makefile.def	(working copy)
> @@ -149,7 +149,7 @@ host_modules= { module= lto-plugin; boot
>  
>  target_modules = { module= libstdc++-v3;
>  		   bootstrap=true;
> -		   lib_path=.libs;
> +		   lib_path=src/.libs;
>  		   raw_cxx=true; };
>  target_modules = { module= libmudflap; lib_path=.libs; };
>  target_modules = { module= libssp; lib_path=.libs; };
Paolo Bonzini - Oct. 20, 2010, 6:26 a.m.
On 10/20/2010 07:24 AM, Ralf Wildenhues wrote:
> Hello Ian,
>
> * Ian Lance Taylor wrote on Mon, Oct 18, 2010 at 10:08:21PM CEST:
>> Ping for build maintainers.
>
> I cannot approve this but I think this can be considered as obvious.

Indeed.

Paolo
Ian Taylor - Oct. 20, 2010, 2:40 p.m.
Ralf Wildenhues <Ralf.Wildenhues@gmx.de> writes:

> * Ian Lance Taylor wrote on Mon, Oct 18, 2010 at 10:08:21PM CEST:
>> Ping for build maintainers.
>
> I cannot approve this but I think this can be considered as obvious.
> One question is why it didn't show up earlier.

Thanks.  Committed.  I think it didn't show up earlier simply because
there are no gcc programs that are written in C++ and run during the
build.  The testsuite uses other mechanisms to set LD_LIBRARY_PATH.

Ian

Patch

Index: Makefile.def
===================================================================
--- Makefile.def	(revision 165058)
+++ Makefile.def	(working copy)
@@ -149,7 +149,7 @@  host_modules= { module= lto-plugin; boot
 
 target_modules = { module= libstdc++-v3;
 		   bootstrap=true;
-		   lib_path=.libs;
+		   lib_path=src/.libs;
 		   raw_cxx=true; };
 target_modules = { module= libmudflap; lib_path=.libs; };
 target_modules = { module= libssp; lib_path=.libs; };
Index: Makefile.tpl
===================================================================
--- Makefile.tpl	(revision 165058)
+++ Makefile.tpl	(working copy)
@@ -513,7 +513,7 @@  all:
 ###
 
 # This is the list of directories that may be needed in RPATH_ENVVAR
-# so that prorgams built for the target machine work.
+# so that programs built for the target machine work.
 TARGET_LIB_PATH = [+ FOR target_modules +][+
   IF lib_path +]$(TARGET_LIB_PATH_[+module+])[+ ENDIF lib_path +][+
   ENDFOR target_modules +]$(HOST_LIB_PATH_gcc)