build of egcs-1.0.1 on mips-sgi-irix6.3

[Andrew V. Klein]

Summary:
========

 A build of egcs-1.0.1 on mips-sgi-irix6.3 went reasonably well.  The
 `-lm' flag caused problems for the Irix 6.3 linker; after removing
 the flag the build proceeded OK.  Almost all of the runtests were OK.


Build details:
==============

 I'd installed the source under /local/egcs-1.0.1, and then set up an
 object directory /local/egcs-objdir.  I'd asked for shared libraries
 in the configure.

 The first time through the build failed when it tried to create a
 shared library for libstdc++ (presumably executed in
 /local/egcs-objdir/libraries/mabi=64/libstdc++):

   /local/egcs-objdir/gcc/xgcc -B/local/egcs-objdir/gcc/
     -g -O2 -fno-implicit-templates
     -mabi=64  -shared -o libstdc++.so.2.8.0 `cat piclist` -lm

   ld: FATAL 9: I/O error (-lm): No such file or directory
   collect2: ld returned 32 exit status
   make[4]: *** [libstdc++.so.2.8.0] Error 1
   make[4]: Leaving directory `/local/egcs-objdir/libraries/mabi=64/libstdc++'

 Note that if the command is executed by hand without the
 -lm option, it completes without error.

 Note also that just a bit earlier the build had no problem with this
 command (presumably executed in
 /local/egcs-objdir/libraries/libstdc++)

   /local/egcs-objdir/gcc/xgcc -B/local/egcs-objdir/gcc/
     -g -O2 -fno-implicit-templates
     -shared -o libstdc++.so.2.8.0 `cat piclist` -lm


 The /local/egcs-objdir/libraries/mabi=64/libstdc++/Makefile said
 this:
 . . .

 SHFLAGS = 
 SHDEPS  =

 STAMP   = bigstmp

 LIBS    = $(ARLIB)

 #### package, host, target, and site dependent Makefile fragments come in here.
 # Warning: this fragment is automatically generated
 # Following fragment copied from ../../../../egcs-1.0.1/libstdc++/config/irix5.ml
 # We don't need -fpic on IRIX, so let's install both the shared and
 # non-shared versions.

 LIBS     = $(ARLIB) $(SHLIB) $(SHLINK)
 DEPLIBS  = ../$(SHLIB)
 SHDEPS   = -lm
 . . .

 Since the build machine was an Irix 6.3, I pulled out the `-lm' from
 here, and from libraries/mabi=64/libstdc++/testsuite/Makefile and
 from libraries/mabi=64/libstdc++/tests/Makefile.  Then I crossed my
 fingers, restarted the build, and eventually got to the end without
 fatal error.


test details:
=============

 The runtests reported this (edited) summary:

		=== libio tests ===
		=== libio Summary ===
 # of expected passes		40

		=== libstdc++ tests ===
		=== libstdc++ Summary ===
 # of expected passes		30

		=== gcc tests ===
 FAIL: gcc.c-torture/execute/complex-5.c execution,  -O0 
 FAIL: gcc.c-torture/execute/complex-5.c execution,  -O1 
 FAIL: gcc.c-torture/execute/complex-5.c execution,  -O2 
 FAIL: gcc.c-torture/execute/complex-5.c execution,  -O2
            -fomit-frame-pointer -finline-functions 

		=== gcc Summary ===
 # of expected passes		4879
 # of unexpected failures	4
 # of expected failures		5
 # of unsupported tests		7
 /local/egcs-objdir/gcc/xgcc
   version egcs-2.90.23 980102 (egcs-1.0.1 release)


		=== g++ tests ===
 Running /local/egcs-1.0.1/gcc/testsuite/g++.old-deja/old-deja.exp ...
 XPASS: g++.jason/destruct3.C - (test for bogus messages, line 38)
 XPASS: g++.mike/dyncast1.C  Execution test
 XPASS: g++.mike/dyncast2.C  Execution test

		=== g++ Summary ===
 # of expected passes		3396
 # of unexpected successes	3
 # of expected failures		84
 # of untested testcases		6
 /local/egcs-objdir/gcc/testsuite/../xgcc
   version egcs-2.90.23 980102 (egcs-1.0.1 release)


		=== g77 tests ===
		=== g77 Summary ===
 # of expected passes		132
 /local/egcs-objdir/gcc/g77
   version egcs-2.90.23 980102 (egcs-1.0.1 release)

-- 

Andrew V. Klein
Motorola, Inc., IL75-3N12, 1475 W. Shure Drive, Arlington Heights, IL 60004
tel.: 847-632-6827     email: avk@cig.mot.com
fax.: 847-435-9017