[PATCH]PR target/10129 ADA bootstrap failure on Darwin

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

The problem is that the Ada front end is marking
a variable as common but is also having a DECL_INITIAL which
confuses the Darwin back-end.

This patch unconfuses the Darwin back-end.

I now get back to the original bootstrap failure of:
stage1/xgcc -Bstage1/ 
-B/Volumes/UFS_Partition/pinskia/fsf-clean/powerpc-apple-darwin6.6/bin/ 
-c -g -O2 -gnatpg -gnata -I- -I. -Iada 
-I/Volumes/UFS_Partition/pinskia/src/fsf/gcc-clean/src/gcc/ada 
/Volumes/UFS_Partition/pinskia/src/fsf/gcc-clean/src/gcc/ada/ali.adb -o 
ada/ali.o
/var/tmp//cc1BemnQ.s:10438:non-relocatable subtraction expression, 
"L00000000149$pb" minus "LSJR780"
/var/tmp//cc1BemnQ.s:10438:symbol: "L00000000149$pb" can't be undefined 
in a subtraction expression
/var/tmp//cc1BemnQ.s:10437:non-relocatable subtraction expression, 
"L00000000149$pb" minus "LSJR780"
/var/tmp//cc1BemnQ.s:10437:symbol: "L00000000149$pb" can't be undefined 
in a subtraction expression
/var/tmp//cc1BemnQ.s:unknown:Undefined local symbol L00000000149$pb

This is a bug in restoring the PIC register.
So I am getting closer to bootstraping with Ada, will look into this 
other problem later today.


ChangeLog:
2003-07-02  Andrew Pinski  <pinskia@physics.uc.edu>

	* config/darwin.c (darwin_encode_section_info): When the decl has
	a DECL_INITIAL, it is only defined also when it is not a common.


Patch:

Attachment: temp1.diff
Description: Binary data