Re: GCC build of HEAD failed for native with your patch on 2004-04-19T16:18:21Z.

[Andrew Pinski <pinskia at physics dot uc dot edu>]

On Apr 19, 2004, at 17:36, Richard Kenner wrote:

>     Kenner this has to been caused by your patch,
>
> Why?  There were dozens of patches in there.  Why do you think it
> had to be that one?  I didn't see any build failures when I tested it.

Because there are only four patches since the last time it was able to  
build
and regression test:
ChangeLog entries since last run on 2004-04-19T06:30:05Z:
--- /Users/regress/tbox/changelog_mail/gcc/gcc/ChangeLog	Sun Apr 18  
23:30:52 2004
+++ gcc/gcc/ChangeLog	Mon Apr 19 09:19:06 2004
@@ -1,3 +1,22 @@
+2004-04-19  Andrew PInski  <pinskia@physics.uc.edu>
+
+	* builtins.c (fold_builtin_cabs): Remove fndecl parameter.
+	(fold_builtin): Update caller to match.
+
+	PR bootstrap/15009
+	* bb-reorder.c (fix_up_fall_thru_edges): Init cond_jump.
+
+	PR bootstrap/14999
+	* builtins.c (fold_builtin_cabs): Mark fndecl as unused.
+
+2004-04-19  Richard Kenner  <kenner@vlsi1.ultra.nyu.edu>
+
+	* varasm.c (compare_constant, case VIEW_CONVERT_EXPR): Add case.
+
+	* expmed.c (expand_mult_highpart_adjust): Make OP1 valid for MODE.
+	(expand_mult_highpart_optab): Likewise.
+	(expand_mult_highpart): Make OP1 valid for WIDER_MODE, not MODE.
+
 2004-04-19  Paul Brook  <paul@codesourcery.com>

 	* config/arm/arm.md (fixuns_truncsfsi2, fixuns_truncdfsi2,
--- /Users/regress/tbox/changelog_mail/gcc/gcc/testsuite/ChangeLog	Sun  
Apr 18 13:47:13 2004
+++ gcc/gcc/testsuite/ChangeLog	Mon Apr 19 09:19:26 2004
@@ -1,3 +1,7 @@
+2004-04-19  Andrew Pinski  <pinskia@physics.uc.edu>
+
+	* gcc.c-torture/compile/20040419-1.c: New test.
+
 2004-04-18  Aldy Hernandez  <aldyh@redhat.com>

 	* gcc.dg/vmx/varargs-1.c: Handle when printf outputs \r\n.
--- /Users/regress/tbox/changelog_mail/gcc/libstdc++-v3/ChangeLog	Sun  
Apr 18 04:05:04 2004
+++ gcc/libstdc++-v3/ChangeLog	Mon Apr 19 09:19:39 2004
@@ -1,3 +1,9 @@
+2004-04-19  Paolo Carlini  <pcarlini@suse.de>
+
+	PR libstdc++/15002 (partial)
+	* include/bits/basic_string.h (_M_replace_aux, _M_replace_safe):
+	Special case __n2 == 1, not calling traits_type::assign/copy.
+
 2004-04-17  Benjamin Kosnik  <bkoz@redhat.com>

 	* include/bits/stl_bvector.h: Use _M_impl._M_start.

So it cannot be Paolo's because it does not touch the code which can be  
invoked
in this case.  It cannot be mine as I am just removing an unused  
parameter and
initializing a variable to start with and the seg fault is not faulting  
on an
address of zero.
/Users/regress/tbox/native/build/gcc/gcj  
-B/Users/regress/tbox/native/build/powerpc-apple-darwin7.2.0/libjava/  
-B/Users/regress/tbox/native/build/gcc/ --encoding=UTF-8  
-Wno-deprecated -fclasspath=  
-fbootclasspath=/Users/regress/tbox/native/build/powerpc-apple- 
darwin7.2.0/libjava -g -O2 -MD -MT  
javax/swing/plaf/basic/BasicTabbedPaneUI.lo -MF  
javax/swing/plaf/basic/BasicTabbedPaneUI.d -c  
/Users/regress/tbox/cvs-gcc/gcc/libjava/javax/swing/plaf/basic/ 
BasicTabbedPaneUI.java -fno-common -o  
javax/swing/plaf/basic/.libs/BasicTabbedPaneUI.o
[address=afafafb3 pc=00375464]
/Users/regress/tbox/cvs-gcc/gcc/libjava/javax/swing/plaf/basic/ 
BasicTabbedPaneUI.java: In class  
`javax.swing.plaf.basic.BasicTabbedPaneUI$TabbedPaneScrollLayout':
/Users/regress/tbox/cvs-gcc/gcc/libjava/javax/swing/plaf/basic/ 
BasicTabbedPaneUI.java: In method  
`javax.swing.plaf.basic.BasicTabbedPaneUI$TabbedPaneScrollLayout.layoutC 
ontainer(java.awt.Container)':
/Users/regress/tbox/cvs-gcc/gcc/libjava/javax/swing/plaf/basic/ 
BasicTabbedPaneUI.java:846: internal compiler error: Segmentation Fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
So that leaves only yours.