Tech Insider					   Technology and Trends


			   USENET Archives


Electronic mail:			      WorldWideWeb:
   tech-insider@outlook.com		         http://tech-insider.org/

Path: gmdzi!unido!mcsun!uunet!bu.edu!purdue!tut.cis.ohio-state.edu!
V6550C.NRL.NAVY.MIL!YOUNGDALE
From: YOUN...@V6550C.NRL.NAVY.MIL (Eric Youngdale)
Newsgroups: gnu.utils.bug
Subject: Misc bugs
Message-ID: <9003072203.AA10212@aeneas.MIT.EDU>
Date: 7 Mar 90 21:06:00 GMT
Sender: dae...@tut.cis.ohio-state.edu
Distribution: gnu
Organization: GNUs Not Usenet
Lines: 80
Posted: Wed Mar  7 22:06:00 1990


	I have been fooling around with various portions of the GNU-C, GNU-AS,
GNU-C++ and BISON compilers, and I have found that there are some rough edges
in some of the source distributions. (We are a VMS V5.1 site).

	1)  If you try to compile a C program with GNU-C that only consists
of data declarations, with no functions, the .OBJ file that is generated
has a problem with it such that the linker will gag on it.  I discovered it
trying to build a fresh version of BISON that the docs for G++ claim is needed
for proper operation.  (this occurs for all versions of GNU-AS up to 1.34)
There is a message  to the effect:

%LINK-W-UDEFPSC, attempt to reference undefined psect number 65535.
	in module GRAM file $DISKU:[ERIC.GNU.BISON-DIST]GRAM.OBJ;9

	2) When I tried to build GNU-AS, version 1.35, I found that there is
a deficiency in the "MAKE-GAS.COM" file, and the differences(VMS) are
diff [.gas-1.35]make-gas1.com [.gas-1.35]make-gas.com
************
File $DISKU:[ERIC.GNU.GAS-1.35]MAKE-GAS1.COM;3
   38   $ gcc/debug/define=("VMS", "error=as_fatal") vax.c
   39   $ gcc/debug/define="VMS" atof-vax.c
   40   $ gcc/debug/define=("VMS", "error=as_fatal") vms.c
   41   $ Link:
******
File $DISKU:[ERIC.GNU.GAS-1.35]MAKE-GAS.COM;1
   35   $ gcc/debug/define="VMS" vax.c
   36   $ gcc/debug/define="VMS" atof-vax.c
   37   $ gcc/debug/define="VMS" vms.c
   38   $ Link:
************

There are problems with the file VMS.C, and I can provide further details
upon request (It will not compile, some of the XAB structures are not defined
right.  It appears that this was an intermediate version of the file,
that was not fully debugged.).


	3) I build the GCC compiler (using an older version of the GCC
compiler), and found that the CPP is unable to write any output files.  The
file is created, but is zero bytes in length.  (This is version 1.37.1)
The error message is:
gcc-cpp: I/O error on output


	4) I tried to build the g++ compiler, and for two of the modules, the
the compiler crashed.  Here I was using an older CPP (see above), and the 
GCC-CC1 generated from 1.37.1.  The crash information is:

%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=7FF454E4,
 PC=0004AC05, PSL=03C00000
%TRACE-F-TRACEBACK, symbolic stack dump follows
module name     routine name                     line       rel PC    abs PC
LOCAL-ALLOC     block_alloc                       636      000005A5  0004AC05
                                                           000B2506  000B2506
LOCAL-ALLOC     block_alloc                       576      00000460  0004AAC0
LOCAL-ALLOC     local_alloc                       306      00000294  0004A652
TOPLEV          rest_of_compilation              1587      00000534  0001054A
C-DECL          finish_function                  3853      00000056  0001D0F2
C-PARSE_TAB     yyparse                           248      00000614  00011DD6
TOPLEV          compile_file                     1118      00000689  0000FAE9
TOPLEV          main                             1995      000005B6  00010F1A

This occurs when I compile the modules CPLUS-DECL and CPLUS-EXCEPT from the 
1.37.1 distribution of C++, with the same traceback for both modules.
I was using the configuration files generated from building GNU-C, since I do
not have unix, and config.g++ looked pretty similar to config.gcc. (Please let
me know if this is an incorrect assumption.)  These two modules compile
when I use the older version of GNU-C (1.35.9).  I cannot say if they compile
correctly, but the compiler does not bomb.

	Please let me know if there is any further information that you require.
------------------------------------------------------------------
Eric Youngdale          INTERNET:  YOUN...@V6550C.NRL.NAVY.MIL
Code 6551		    SPAN:  11.14 (or 11277::)
Naval Research Lab	 ATT-NET:  (202) 767-3276
Washington, DC  20375  FLAME-NET:  NLA0:
	Disclaimer:  The opinions expressed herein are my own and not
necessarily those of my employer.
------------------------------------------------------------------

			   USENET Archives


Notice
******

The materials and information included in this website may only be used
for purposes such as criticism, review, private study, scholarship, or 
research.


Electronic mail:			      WorldWideWeb:
   tech-insider@outlook.com		         http://tech-insider.org/