970929: MY patch list

970929: MY patch list

[Chip Salzenberg]

(I mailed this once, but haven't seen it; apologies if it's a dup for you.)

H.J. still can't duplicate my exception problem.  I've applied the
following patches; is any of them a Bad Idea when debugging exceptions?

#00:
    Thu Oct  2 14:28:30 1997  David S. Miller  <davem@tanya.rutgers.edu>
	* explow.c (optimize_save_area_alloca): New function for targets
	where SETJMP_VIA_SAVE_AREA is true.
	(allocate_dynamic_stack_space): On SETJMP_VIA_SAVE_AREA targets,
	compute the amount of stack space needed should we find later that
	setjmp is never called by this function, stuff rtl for this inside
	a REG_NOTE of the final SET of stack_pointer_rtx.
	* toplev.c (rest_of_compilation): If SETJMP_VIA_SAVE_AREA and
	current_function_calls_alloca, call optimize_save_area_alloca.

#01:
    n/a

#02:
    Tue Sep 30 11:05:47 1997  Richard Henderson  <rth@cygnus.com>
	* i386.h (RETURN_ADDR_RTX): New definition that works for 
	__builtin_return_address(0) and -fomit-frame-pointer.

#03:
    Wed Oct  1 15:37:58 CEST 1997  Thomas Koenig (ig25@rz.uni-karlsruhe.de)
	* cp/parse.y: Remove duplicate definition of PFUNCNAME

#04:
    Wed Oct  1 23:51:06 CEST 1997  Thomas Koenig (ig25@rz.uni-karlsruhe.de)
	* reload.c (decompose): Always initialize val.base.

#05:
    Wed Oct  1 17:47:48 1997  Richard Henderson  <rth@cygnus.com>
	* expr.c (get_inner_reference): Remove the array bias after 
	converting the index to Pmode.

#06:
    Tue Sep 30 08:51:36 1997  Jason Merrill  <jason@yorick.cygnus.com>
        * method.c (build_decl_overload_real): Reformat.

#07:
    1997-10-04  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>
	* frame.c (__frame_state_for): Execute the FDE insns until the
	current pc value is strictly bigger than the target pc value.

#08:
    Tue Sep 30 16:50:55 CEST 1997  Thomas Koenig  (ig25@rz.uni-karlsruhe.de)
	* cccp.c: (struct file_buf) change system_header_p to int
          (expand_to_temp_buffer) initialize all members of obuf

#09:
    Fri Oct  3 10:42:37 1997  Mark Mitchell  <mmitchell@usa.net>
	* decl.c (make_typename_type): Do not try to call lookup_field for
	non-aggregate types.

#10:
    Fri Oct  3 11:56:36 1997  Jason Merrill  <jason@yorick.cygnus.com>
	* toplev.c (rest_of_compilation): Defer all non-nested inlines.

#11:
    Alexandre Oliva <oliva@dcc.unicamp.br>
	* libio.h: _G_HAVE_PRINTF_FP's value must be tested, not its
	definition
	* iostream.cc: ditto

#12:
    "... patch by Jeff Law that fixes the suboptimal recognition of
     GIVs due to removal of some special case ARRAY_REF code from
     expand_expr (see [Toon Moene's] mail dd 14th of September)"
	* loop.c

-- 
Chip Salzenberg          - a.k.a. -           <chip@pobox.com>
        "He's Mr. Big of 'Big And Tall' fame."  // MST3K

Re: 970929: MY patch list

[H.J. Lu]

> 
> (I mailed this once, but haven't seen it; apologies if it's a dup for you.)
> 
> H.J. still can't duplicate my exception problem.  I've applied the
> following patches; is any of them a Bad Idea when debugging exceptions?
> 

Since you have indicated that the same asm output which works for
me doesn't work for you, so the problem may come from somewhere
else.

I always compile glibc with -fno-exceptions and install egcs in
/usr. They may make a difference. Also I use binutils 2.8.1.0.15.

BTW, can you send me

# gcc -v ......

to show what are used?

Thanks.

H.J.

Re: 970929: MY patch list

[chip]

According to H.J. Lu:
> I always compile glibc with -fno-exceptions and install egcs in
> /usr. They may make a difference.

I don't know how my glibc is built, unfortunately, it's the debian
package.

> Also I use binutils 2.8.1.0.15.

All I know about my binutils is that they're based on 2.8.1.

This isn't very helpful, I know.

> BTW, can you send me
> # gcc -v ......
> to show what are used?

$ sg++ -v tt.cc
Reading specs from /usr/local/lib/gcc-lib/i486-pc-linux-gnu/egcs-2.90.11/specs
gcc version egcs-2.90.11 970929 (gcc2-970802 experimental)
 /usr/local/lib/gcc-lib/i486-pc-linux-gnu/egcs-2.90.11/cpp -lang-c++ -v -undef -D__GNUC__=2 -D__GNUG__=2 -D__cplusplus -D__GNUC_MINOR__=90 -D__ELF__ -Dunix -Dlinux -D__ELF__ -D__unix__ -D__linux__ -D__unix -D__linux -Asystem(posix) -D__EXCEPTIONS -Di386 -Di486 -Asystem(unix) -Acpu(i386) -Amachine(i386) -D__i386__ -D__i486__ -Asystem(unix) -Acpu(i386) -Amachine(i386) tt.cc /tmp/cca27651.ii
GNU CPP version egcs-2.90.11 970929 (gcc2-970802 experimental) (i386 Linux/ELF)
#include "..." search starts here:
#include <...> search starts here:
 /usr/local/include/g++
 /usr/local/include
 /usr/local/i486-pc-linux-gnu/include
 /usr/local/lib/gcc-lib/i486-pc-linux-gnu/egcs-2.90.11/include
 /usr/include
End of search list.
 /usr/local/lib/gcc-lib/i486-pc-linux-gnu/egcs-2.90.11/cc1plus /tmp/cca27651.ii -quiet -dumpbase tt.cc -version -o /tmp/cca27651.s
GNU C++ version egcs-2.90.11 970929 (gcc2-970802 experimental) (i486-pc-linux-gnu) compiled by GNU C version egcs-2.90.11 970929 (gcc2-970802 experimental).
 as -V -Qy -o /tmp/cca276511.o /tmp/cca27651.s
GNU assembler version 970607 (i486-linux), using BFD version 2.8.1
 /usr/local/lib/gcc-lib/i486-pc-linux-gnu/egcs-2.90.11/ld -m elf_i386 -dynamic-linker /lib/ld-linux.so.2 /usr/lib/crt1.o /usr/lib/crti.o /usr/local/lib/gcc-lib/i486-pc-linux-gnu/egcs-2.90.11/crtbegin.o -L/usr/local/lib/gcc-lib/i486-pc-linux-gnu/egcs-2.90.11 -L/usr/local/i486-pc-linux-gnu/lib -L/usr/local/lib /tmp/cca276511.o -lstdc++ -lm -lgcc -lc -lgcc /usr/local/lib/gcc-lib/i486-pc-linux-gnu/egcs-2.90.11/crtend.o /usr/lib/crtn.o

Re: 970929: MY patch list

[Michael Alan Dorman]

chip@atlantic.net writes:
> According to H.J. Lu:
> > I always compile glibc with -fno-exceptions and install egcs in
> > /usr. They may make a difference.
> I don't know how my glibc is built, unfortunately, it's the debian
> package.

It's built using whatever flags glibc's configure happens to
like---which, since HJ implies that it's not the default, suggests it
doesn't use -fno-exceptions.

> > Also I use binutils 2.8.1.0.15.
> All I know about my binutils is that they're based on 2.8.1.
> This isn't very helpful, I know.

Ah, but you have me around. :-)

We're using a 2.8.1.0.4.  I'll see if I can't get Galen to consider an
update.

Mike.

Re: 970929: MY patch list

[H.J. Lu]

> 
> According to H.J. Lu:
> > I always compile glibc with -fno-exceptions and install egcs in
> > /usr. They may make a difference.
> 
> I don't know how my glibc is built, unfortunately, it's the debian
> package.

Please check with debian to see which compiler was used to compile
glibc. If egcs or gcc snapshot was used, -fno-exceptions should
be used. Otherwise, it is looking for trouble since EH is very
buggy. I always use -fno-exceptions to compile glibc.


H.J.