* [pch-branch] missing binary operator issue.
@ 2002-12-13 4:04 Karel Gardas
2003-01-02 17:59 ` Geoff Keating
0 siblings, 1 reply; 6+ messages in thread
From: Karel Gardas @ 2002-12-13 4:04 UTC (permalink / raw)
To: GCC Mailing List
Hello,
I've updated to todays pch-branch sources and try to compile MICO with
this compiler. Unfortunatelly, when using PCH it fails to compile first
file with:
os-unix.cc:194:12: missing binary operator before token "("
os-unix.cc:243:14: missing binary operator before token "("
../include/mico/code_set_registry.h:5: warning: `CORBA::Codeset::Info
_osf_cs_reg[193]' defined but not used
This file is perfectly compilable w/o using PCH with this compiler.
Problematic lines look:
193
194 #if defined(HAVE_DLOPEN) && defined(HAVE_DYNAMIC)
195
and
242
243 #elif defined(HAVE_SHL_LOAD) && defined(HAVE_DYNAMIC)
244
FYI: Comeau is perfectly ok with these sources even with PCH on, so I hope
they are ok from PCH point of view, but it might be a bug in gcc.
Anything what should I test for you?
Thanks,
Karel
--
Karel Gardas kgardas@objectsecurity.com
ObjectSecurity Ltd. http://www.objectsecurity.com
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [pch-branch] missing binary operator issue.
2002-12-13 4:04 [pch-branch] missing binary operator issue Karel Gardas
@ 2003-01-02 17:59 ` Geoff Keating
2003-01-02 21:26 ` Karel Gardas
0 siblings, 1 reply; 6+ messages in thread
From: Geoff Keating @ 2003-01-02 17:59 UTC (permalink / raw)
To: Karel Gardas; +Cc: gcc
Karel Gardas <kgardas@objectsecurity.com> writes:
> os-unix.cc:194:12: missing binary operator before token "("
> 194 #if defined(HAVE_DLOPEN) && defined(HAVE_DYNAMIC)
This should now be fixed.
--
- Geoffrey Keating <geoffk@geoffk.org>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [pch-branch] missing binary operator issue.
2003-01-02 17:59 ` Geoff Keating
@ 2003-01-02 21:26 ` Karel Gardas
2003-01-02 22:21 ` Andrew Pinski
2003-01-07 19:32 ` Geoff Keating
0 siblings, 2 replies; 6+ messages in thread
From: Karel Gardas @ 2003-01-02 21:26 UTC (permalink / raw)
To: Geoff Keating; +Cc: gcc
On Thu, 2 Jan 2003, Geoff Keating wrote:
> Karel Gardas <kgardas@objectsecurity.com> writes:
>
> > os-unix.cc:194:12: missing binary operator before token "("
>
> > 194 #if defined(HAVE_DLOPEN) && defined(HAVE_DYNAMIC)
>
> This should now be fixed.
>
Great! But now I get:
thinkpad:~/arch/devel/mico-devel/orb$ c++ -I../include -Wall -DFAST_PCH
-fpermissive -DPIC -fPIC -c orb_pch.h
../include/mico/code_set_registry.h:5: warning: `CORBA::Codeset::Info
_osf_cs_reg[193]' defined but not used
thinkpad:~/arch/devel/mico-devel/orb$ make
c++ -I../include -Wall -DFAST_PCH -fpermissive -DPIC -fPIC -c
os-unix.cc -o os-unix.pic.o
../include/mico/code_set_registry.h:5: warning: `CORBA::Codeset::Info
_osf_cs_reg[193]' defined but not used
/tmp/cczg5mr7.s: Assembler messages:
/tmp/cczg5mr7.s:2963: Error: symbol `.LCFI0' is already defined
/tmp/cczg5mr7.s:2965: Error: symbol `.LCFI1' is already defined
/tmp/cczg5mr7.s:2967: Error: symbol `.LCFI2' is already defined
/tmp/cczg5mr7.s:2969: Error: symbol `.LCFI3' is already defined
/tmp/cczg5mr7.s:2973: Error: symbol `.L2' is already defined
/tmp/cczg5mr7.s:2985: Error: symbol `.L4' is already defined
/tmp/cczg5mr7.s:2999: Error: symbol `.L6' is already defined
/tmp/cczg5mr7.s:3033: Error: symbol `.L10' is already defined
/tmp/cczg5mr7.s:3039: Error: symbol `.L11' is already defined
/tmp/cczg5mr7.s:3090: Error: symbol `.L15' is already defined
/tmp/cczg5mr7.s:3104: Error: symbol `.L8' is already defined
/tmp/cczg5mr7.s:3125: Error: symbol `.LCFI4' is already defined
/tmp/cczg5mr7.s:3127: Error: symbol `.LCFI5' is already defined
/tmp/cczg5mr7.s:3129: Error: symbol `.LCFI6' is already defined
/tmp/cczg5mr7.s:3131: Error: symbol `.LCFI7' is already defined
/tmp/cczg5mr7.s:3155: Error: symbol `.LEHB0' is already defined
/tmp/cczg5mr7.s:3157: Error: symbol `.LEHE0' is already defined
/tmp/cczg5mr7.s:3173: Error: symbol `.LEHB1' is already defined
/tmp/cczg5mr7.s:3181: Error: symbol `.LEHE1' is already defined
/tmp/cczg5mr7.s:3214: Error: symbol `.LEHB2' is already defined
/tmp/cczg5mr7.s:3216: Error: symbol `.LEHE2' is already defined
make: *** [os-unix.pic.o] Error 1
thinkpad:~/arch/devel/mico-devel/orb$
BTW: IMHO something strage is with my gcc3.4-pch... the branch compiled
10-20 % slower than version which I tried before, and whole configuration
process seems to be moved to build process. Is this behaviour right, or
have I done something wrong?
I've configured gcc with:
$HOME/cvs/gcc/pch-branch/gcc/configure --prefix=$HOME/usr/local/pch-branch
--enable-shared --enable-threads --enable-languages=c++ --disable-checking
--enable-__cxa_atexit
and build it with:
time make CFLAGS='-O' LIBCFLAGS='-g -O2' LIBCXXFLAGS='-g -O2
-fno-implicit-templates' bootstrap-lean
Unfortunately gcc -v said:
thinkpad:~/arch/devel/mico-devel/orb$ gcc -v
Reading specs from /home/karel/usr/local/pch-branch/lib/gcc-lib/i686-pc-linux-gnu/3.4-pch/specs
Configured with:
Thread model: posix
gcc version 3.4-pch 20021218 (experimental)
thinkpad:~/arch/devel/mico-devel/orb$
Anything what should I test?
Thanks,
Karel
--
Karel Gardas kgardas@objectsecurity.com
ObjectSecurity Ltd. http://www.objectsecurity.com
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [pch-branch] missing binary operator issue.
2003-01-02 21:26 ` Karel Gardas
@ 2003-01-02 22:21 ` Andrew Pinski
2003-01-07 19:32 ` Geoff Keating
1 sibling, 0 replies; 6+ messages in thread
From: Andrew Pinski @ 2003-01-02 22:21 UTC (permalink / raw)
To: Karel Gardas; +Cc: gcc
On Thursday, Jan 2, 2003, at 13:25 US/Pacific, Karel Gardas wrote:
> BTW: IMHO something strage is with my gcc3.4-pch... the branch compiled
> 10-20 % slower than version which I tried before, and whole
> configuration
> process seems to be moved to build process. Is this behaviour right, or
> have I done something wrong?
This is because 3.4 changed how every thing is done with building gcc,
It was done
so to autoconfive the top-level configure and Makefile.
Thanks,
Andrew Pinski
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [pch-branch] missing binary operator issue.
2003-01-02 21:26 ` Karel Gardas
2003-01-02 22:21 ` Andrew Pinski
@ 2003-01-07 19:32 ` Geoff Keating
2003-01-08 16:06 ` Karel Gardas
1 sibling, 1 reply; 6+ messages in thread
From: Geoff Keating @ 2003-01-07 19:32 UTC (permalink / raw)
To: kgardas; +Cc: gcc
> /tmp/cczg5mr7.s:2963: Error: symbol `.LCFI0' is already defined
...
> /tmp/cczg5mr7.s:2973: Error: symbol `.L2' is already defined
This patch might fix this problem. Could you test it?
I know there are other problems with DWARF2 output, so I don't promise
the resulting DWARF2 is correct, but it should at least assemble.
It'd be good if you could provide a small test case that triggers the
problem, I didn't manage to generate one.
--
- Geoffrey Keating <geoffk@geoffk.org>
===File ~/patches/apple/pchbranch-moredwarfcounters.patch===
2003-01-06 Geoffrey Keating <geoffk@apple.com>
* dwarf2out.c (dwarf2out_cfi_label): Move label_num out of
the routine and mark it for PCH.
(gen_internal_sym): Likewise.
Index: dwarf2out.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/dwarf2out.c,v
retrieving revision 1.349.2.17
diff -u -p -u -p -r1.349.2.17 dwarf2out.c
--- dwarf2out.c 24 Dec 2002 02:06:23 -0000 1.349.2.17
+++ dwarf2out.c 7 Jan 2003 01:30:38 -0000
@@ -1,5 +1,5 @@
/* Output Dwarf2 format symbol table information from the GNU C compiler.
- Copyright (C) 1992, 1993, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002
+ Copyright (C) 1992, 1993, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003
Free Software Foundation, Inc.
Contributed by Gary Funck (gary@intrepid.com).
Derived from DWARF 1 implementation of Ron Guilmette (rfg@monkeys.com).
@@ -305,7 +305,11 @@ struct indirect_string_node GTY(())
static GTY ((param_is (struct indirect_string_node))) htab_t debug_str_hash;
+/* Various counters for generating labels. */
+
static GTY(()) int dw2_string_counter;
+static GTY(()) unsigned long dwarf2out_cfi_label_counter;
+static GTY(()) int gen_internal_sym_counter;
#if defined (DWARF2_DEBUGGING_INFO) || defined (DWARF2_UNWIND_INFO)
@@ -561,9 +565,8 @@ char *
dwarf2out_cfi_label ()
{
static char label[20];
- static unsigned long label_num = 0;
- ASM_GENERATE_INTERNAL_LABEL (label, "LCFI", label_num++);
+ ASM_GENERATE_INTERNAL_LABEL (label, "LCFI", dwarf2out_cfi_label_counter++);
ASM_OUTPUT_LABEL (asm_out_file, label);
return label;
}
@@ -5939,9 +5942,8 @@ gen_internal_sym (prefix)
const char *prefix;
{
char buf[256];
- static int label_num;
- ASM_GENERATE_INTERNAL_LABEL (buf, prefix, label_num++);
+ ASM_GENERATE_INTERNAL_LABEL (buf, prefix, gen_internal_sym_counter++);
return xstrdup (buf);
}
============================================================
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [pch-branch] missing binary operator issue.
2003-01-07 19:32 ` Geoff Keating
@ 2003-01-08 16:06 ` Karel Gardas
0 siblings, 0 replies; 6+ messages in thread
From: Karel Gardas @ 2003-01-08 16:06 UTC (permalink / raw)
To: Geoff Keating; +Cc: gcc
On Tue, 7 Jan 2003, Geoff Keating wrote:
>
> > /tmp/cczg5mr7.s:2963: Error: symbol `.LCFI0' is already defined
> ...
> > /tmp/cczg5mr7.s:2973: Error: symbol `.L2' is already defined
>
> This patch might fix this problem. Could you test it?
>
No it doesn't. Still I get:
thinkpad:~/arch/devel/mico-devel/orb$ c++ -I../include -Wall -DFAST_PCH -fpermissive -DPIC -fPIC -c orb_pch.h
../include/mico/code_set_registry.h:5:
warning: `CORBA::Codeset::Info
_osf_cs_reg[193]' defined but not used
thinkpad:~/arch/devel/mico-devel/orb$
thinkpad:~/arch/devel/mico-devel/orb$
make c++ -I../include -Wall -DFAST_PCH -fpermissive -DPIC -fPIC -c
os-unix.cc -o os-unix.pic.o
../include/mico/code_set_registry.h:5: warning: `CORBA::Codeset::Info
_osf_cs_reg[193]' defined but not used
/tmp/cctHJBgN.s: Assembler messages:
/tmp/cctHJBgN.s:2973: Error: symbol `.L2' is already defined
/tmp/cctHJBgN.s:2985: Error: symbol `.L4' is already defined
/tmp/cctHJBgN.s:2999: Error: symbol `.L6' is already defined
/tmp/cctHJBgN.s:3033: Error: symbol `.L10' is already defined
/tmp/cctHJBgN.s:3039: Error: symbol `.L11' is already defined
/tmp/cctHJBgN.s:3090: Error: symbol `.L15' is already defined
/tmp/cctHJBgN.s:3104: Error: symbol `.L8' is already defined
/tmp/cctHJBgN.s:3155: Error: symbol `.LEHB0' is already defined
/tmp/cctHJBgN.s:3157: Error: symbol `.LEHE0' is already defined
/tmp/cctHJBgN.s:3173: Error: symbol `.LEHB1' is already defined
/tmp/cctHJBgN.s:3181: Error: symbol `.LEHE1' is already defined
/tmp/cctHJBgN.s:3214: Error: symbol `.LEHB2' is already defined
/tmp/cctHJBgN.s:3216: Error: symbol `.LEHE2' is already defined
make: *** [os-unix.pic.o] Error 1
> I know there are other problems with DWARF2 output, so I don't promise
> the resulting DWARF2 is correct, but it should at least assemble.
>
> It'd be good if you could provide a small test case that triggers the
> problem, I didn't manage to generate one.
I'll try, but IMHO it's quite hard. Do you have any advice which can help
me with it?
Thanks,
Karel
PS: I've used CVS sources 3 hours old + your patch.
--
Karel Gardas kgardas@objectsecurity.com
ObjectSecurity Ltd. http://www.objectsecurity.com
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2003-01-08 12:21 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-12-13 4:04 [pch-branch] missing binary operator issue Karel Gardas
2003-01-02 17:59 ` Geoff Keating
2003-01-02 21:26 ` Karel Gardas
2003-01-02 22:21 ` Andrew Pinski
2003-01-07 19:32 ` Geoff Keating
2003-01-08 16:06 ` Karel Gardas
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).