public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
* -frepo not working?
@ 1998-12-14 11:13 Chad Gatesman
  1998-12-16  0:41 ` Alexandre Oliva
  0 siblings, 1 reply; 15+ messages in thread
From: Chad Gatesman @ 1998-12-14 11:13 UTC (permalink / raw)
  To: egcs, egcs-bugs

I have had a problem with the -frepo option on egcs ever since
egcs-1.1.  This has forced me to use egcs-1.0.3 to this very day.

The problem I see is that when I compile with -frepo, I get unresolved
symbols to what looks like all my templates when it does the final
link.  Why would it not be instantiating my templates?

Has anyone else experienced this problem?  Has something changed that
you have to do differently to use the -frepo option on 1.1 and 1.1.1?

--
 Chad Gatesman
 Software Engineer
 Bristol Technology
 http://www.bristol.com/



^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1998-12-14 11:13 -frepo not working? Chad Gatesman
@ 1998-12-16  0:41 ` Alexandre Oliva
  1998-12-16  9:55   ` Chad Gatesman
  0 siblings, 1 reply; 15+ messages in thread
From: Alexandre Oliva @ 1998-12-16  0:41 UTC (permalink / raw)
  To: Chad Gatesman; +Cc: egcs, egcs-bugs

On Dec 14, 1998, Chad Gatesman <chadg@redrose.net> wrote:

> The problem I see is that when I compile with -frepo, I get unresolved
> symbols to what looks like all my templates when it does the final
> link.  Why would it not be instantiating my templates?

There were bugs in older releases, that may have been fixed in 1.1.1
or not.  Will you please post a code snippet that demonstrates the
problem so that we can verify if the bug you have found has been fixed 
or not?

-- 
Alexandre Oliva  http://www.dcc.unicamp.br/~oliva  aoliva@{acm.org}
oliva@{dcc.unicamp.br,gnu.org,egcs.cygnus.com,samba.org}
Universidade Estadual de Campinas, SP, Brasil

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1998-12-16  0:41 ` Alexandre Oliva
@ 1998-12-16  9:55   ` Chad Gatesman
  1998-12-16 11:07     ` Lee Iverson
  1998-12-17  8:06     ` Horst von Brand
  0 siblings, 2 replies; 15+ messages in thread
From: Chad Gatesman @ 1998-12-16  9:55 UTC (permalink / raw)
  To: Alexandre Oliva; +Cc: egcs, egcs-bugs

Alexandre Oliva wrote:

> On Dec 14, 1998, Chad Gatesman <chadg@redrose.net> wrote:
>
> > The problem I see is that when I compile with -frepo, I get unresolved
> > symbols to what looks like all my templates when it does the final
> > link.  Why would it not be instantiating my templates?
>
> There were bugs in older releases, that may have been fixed in 1.1.1
> or not.  Will you please post a code snippet that demonstrates the
> problem so that we can verify if the bug you have found has been fixed
> or not?

This can be reproduced very easily.  I will show you an example using
egcs-1.1.1:

_________________________________

// repo-fail.cc
#include <vector>

void main()
{
    vector<int> v;

    v.push_back(1);
}

_________________________________


chad:/tmp> g++ -c -frepo repo-fail.cc
chad:/tmp> g++ -o repo-fail repo-fail.o
repo-fail.o: In function `vector<int, __default_alloc_template<false, 0>
>::push_back(int const &)':
repo-fail.o(.vector<int, __default_alloc_template<0, 0>
>::gnu.linkonce.t.push_back(int const &)+0x35): undefined reference to
`vector<int, __default_alloc_template<false, 0> >::insert_aux(int *, int
const &)'
collect2: ld returned 1 exit status



--
Chad Gatesman
Software Engineer
Bristol Technology
http://www.bristol.com/


^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1998-12-16  9:55   ` Chad Gatesman
@ 1998-12-16 11:07     ` Lee Iverson
  1998-12-16 11:20       ` Chad Gatesman
  1998-12-17  8:06     ` Horst von Brand
  1 sibling, 1 reply; 15+ messages in thread
From: Lee Iverson @ 1998-12-16 11:07 UTC (permalink / raw)
  To: Chad Gatesman; +Cc: Alexandre Oliva, egcs, egcs-bugs

In message < 367801D4.F6C06EC3@redrose.net > you write:
> 
> This can be reproduced very easily.  I will show you an example using
> egcs-1.1.1:
> 
> _________________________________
> 
> // repo-fail.cc
> #include <vector>
> 
> void main()
> {
>     vector<int> v;
> 
>     v.push_back(1);
> }
> 
> _________________________________
> 
> 
> chad:/tmp> g++ -c -frepo repo-fail.cc
> chad:/tmp> g++ -o repo-fail repo-fail.o
> repo-fail.o: In function `vector<int, __default_alloc_template<false, 0>
> >::push_back(int const &)':
> repo-fail.o(.vector<int, __default_alloc_template<0, 0>
> >::gnu.linkonce.t.push_back(int const &)+0x35): undefined reference to
> `vector<int, __default_alloc_template<false, 0> >::insert_aux(int *, int
> const &)'
> collect2: ld returned 1 exit status
> 

Bzzzt...  Close but...  -frepo is a *linker* command.  Try adding it
to the second line 'g++ -frepo -o repo-fail repo-fail.o'

-------------------------------------------------------------------------------
Lee Iverson     		SRI International
leei@ai.sri.com			333 Ravenswood Ave., Menlo Park CA 94025
http://www.ai.sri.com/~leei/	(650) 859-3307

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1998-12-16 11:07     ` Lee Iverson
@ 1998-12-16 11:20       ` Chad Gatesman
  1998-12-16 11:33         ` Lee Iverson
  0 siblings, 1 reply; 15+ messages in thread
From: Chad Gatesman @ 1998-12-16 11:20 UTC (permalink / raw)
  To: Lee Iverson; +Cc: Alexandre Oliva, egcs, egcs-bugs

Lee Iverson wrote:

> In message < 367801D4.F6C06EC3@redrose.net > you write:
> >
> > This can be reproduced very easily.  I will show you an example using
> > egcs-1.1.1:
> >
> > _________________________________
> >
> > // repo-fail.cc
> > #include <vector>
> >
> > void main()
> > {
> >     vector<int> v;
> >
> >     v.push_back(1);
> > }
> >
> > _________________________________
> >
> >
> > chad:/tmp> g++ -c -frepo repo-fail.cc
> > chad:/tmp> g++ -o repo-fail repo-fail.o
> > repo-fail.o: In function `vector<int, __default_alloc_template<false, 0>
> > >::push_back(int const &)':
> > repo-fail.o(.vector<int, __default_alloc_template<0, 0>
> > >::gnu.linkonce.t.push_back(int const &)+0x35): undefined reference to
> > `vector<int, __default_alloc_template<false, 0> >::insert_aux(int *, int
> > const &)'
> > collect2: ld returned 1 exit status
> >
>
> Bzzzt...  Close but...  -frepo is a *linker* command.  Try adding it
> to the second line 'g++ -frepo -o repo-fail repo-fail.o'

Umm...No.
This problem still exists if you add -frepo to the link line as well.

--
 Chad Gatesman
 Software Engineer
 Bristol Technology
 http://www.bristol.com/



^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1998-12-16 11:20       ` Chad Gatesman
@ 1998-12-16 11:33         ` Lee Iverson
  1998-12-16 12:04           ` Chad Gatesman
  0 siblings, 1 reply; 15+ messages in thread
From: Lee Iverson @ 1998-12-16 11:33 UTC (permalink / raw)
  To: Chad Gatesman; +Cc: Alexandre Oliva, egcs, egcs-bugs

In message < 98Dec16.141926est.26886@gateway.bristol.com > you write:
> Lee Iverson wrote:
> 
> >
> > Bzzzt...  Close but...  -frepo is a *linker* command.  Try adding it
> > to the second line 'g++ -frepo -o repo-fail repo-fail.o'
> 
> Umm...No.
> This problem still exists if you add -frepo to the link line as well.
> 

Perhaps I was a bit terse.  After compiling with -frepo -c, you must
then provide -frepo to the *linker* line as well:

% g++ -c -frepo repo-fail.cc
% g++ -frepo -o repo-fail repo-fail.o
collect: recompiling repo-fail.cc
collect: relinking
ld: WARNING 84: /usr/lib/libm.so is not used for resolving any symbol.
ld: WARNING 84: /opt/gnu/gcc/egcs-o32/lib/gcc-lib/mips-sgi-irix5/egcs-2.92.21/libgcc.a is not used for resolving any symbol.
ld: WARNING 84: /usr/lib/libm.so is not used for resolving any symbol.
ld: WARNING 84: /opt/gnu/gcc/egcs-o32/lib/gcc-lib/mips-sgi-irix5/egcs-2.92.21/libgcc.a is not used for resolving any symbol.
% 

Notice the collect: lines!  That is -frepo working.

-------------------------------------------------------------------------------
Lee Iverson     		SRI International
leei@ai.sri.com			333 Ravenswood Ave., Menlo Park CA 94025
http://www.ai.sri.com/~leei/	(650) 859-3307

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1998-12-16 11:33         ` Lee Iverson
@ 1998-12-16 12:04           ` Chad Gatesman
  1998-12-16 12:25             ` H.J. Lu
  0 siblings, 1 reply; 15+ messages in thread
From: Chad Gatesman @ 1998-12-16 12:04 UTC (permalink / raw)
  To: Lee Iverson; +Cc: Alexandre Oliva, egcs, egcs-bugs

Lee Iverson wrote:

> In message < 98Dec16.141926est.26886@gateway.bristol.com > you write:
> > Lee Iverson wrote:
> >
> > >
> > > Bzzzt...  Close but...  -frepo is a *linker* command.  Try adding it
> > > to the second line 'g++ -frepo -o repo-fail repo-fail.o'
> >
> > Umm...No.
> > This problem still exists if you add -frepo to the link line as well.
> >
>
> Perhaps I was a bit terse.  After compiling with -frepo -c, you must
> then provide -frepo to the *linker* line as well:
>
> % g++ -c -frepo repo-fail.cc
> % g++ -frepo -o repo-fail repo-fail.o
> collect: recompiling repo-fail.cc
> collect: relinking
> ld: WARNING 84: /usr/lib/libm.so is not used for resolving any symbol.
> ld: WARNING 84: /opt/gnu/gcc/egcs-o32/lib/gcc-lib/mips-sgi-irix5/egcs-2.92.21/libgcc.a is not used for resolving any symbol.
> ld: WARNING 84: /usr/lib/libm.so is not used for resolving any symbol.
> ld: WARNING 84: /opt/gnu/gcc/egcs-o32/lib/gcc-lib/mips-sgi-irix5/egcs-2.92.21/libgcc.a is not used for resolving any symbol.
> %
>
> Notice the collect: lines!  That is -frepo working.

Yes.  I understood you, and I am saying it still does not work for me if I put -frepo on the link line.  Here are the versions
I am using.  Could it possibly be a problem there, or maybe a platform specific bug?  I am using Linux 2.1.129 on an Intel
Pentium using libc.so.5.4.46.


chad:/tmp> g++ -v
Reading specs from /usr/local/lib/gcc-lib/i586-pc-linux-gnulibc1/pgcc-2.91.60/specs
gcc version pgcc-2.91.60 19981201 (egcs-1.1.1 release)
chad:/tmp> ld -v
GNU ld version 2.9.1 (with BFD 2.9.1)

chad:/tmp> g++ -frepo -c repo-fail.cc
chad:/tmp> g++ -frepo -o repo-fail repo-fail.o
repo-fail.o: In function `vector<int, __default_alloc_template<false, 0> >::push_back(int const &)':
repo-fail.o(.vector<int, __default_alloc_template<0, 0> >::gnu.linkonce.t.push_back(int const &)+0x35): undefined reference to
`vector<int, __default_alloc_template<false, 0> >::insert_aux(int *, int const &)'
collect2: ld returned 1 exit status


The actuall code I am writing uses STL extensively, and I do see collect run in that case, but it eventually craps out with
undefined symbols as well.  And it does work with egcs-1.0.3.


--
 Chad Gatesman
 Software Engineer
 Bristol Technology
 http://www.bristol.com/


^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1998-12-16 12:04           ` Chad Gatesman
@ 1998-12-16 12:25             ` H.J. Lu
  1998-12-16 15:37               ` Chad Gatesman
  1998-12-17  8:21               ` Chad Gatesman
  0 siblings, 2 replies; 15+ messages in thread
From: H.J. Lu @ 1998-12-16 12:25 UTC (permalink / raw)
  To: Chad Gatesman; +Cc: leei, oliva, egcs, egcs-bugs

> Yes.  I understood you, and I am saying it still does not work for me if I put -frepo on the link line.  Here are the versions
> I am using.  Could it possibly be a problem there, or maybe a platform specific bug?  I am using Linux 2.1.129 on an Intel
> Pentium using libc.so.5.4.46.
> 
> 
> chad:/tmp> g++ -v
> Reading specs from /usr/local/lib/gcc-lib/i586-pc-linux-gnulibc1/pgcc-2.91.60/specs
> gcc version pgcc-2.91.60 19981201 (egcs-1.1.1 release)
> chad:/tmp> ld -v
> GNU ld version 2.9.1 (with BFD 2.9.1)

If you use Linux, you should use my binutils. The
current one is 2.9.1.0.19a.

> 
> chad:/tmp> g++ -frepo -c repo-fail.cc
> chad:/tmp> g++ -frepo -o repo-fail repo-fail.o

# uname -a
Linux ocean.lucon.org 2.1.131 #1 SMP Tue Dec 15 19:35:46 PST 1998 i686 unknown
# g++ -frepo -c repo-fail.cc 
# g++ -o repo-fail repo-fail.o
collect: recompiling repo-fail.cc
collect: relinking

BTW, I am using glibc 2.0.7.

H.J.

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1998-12-16 12:25             ` H.J. Lu
@ 1998-12-16 15:37               ` Chad Gatesman
  1999-01-22  9:26                 ` Swen Thuemmler
  1998-12-17  8:21               ` Chad Gatesman
  1 sibling, 1 reply; 15+ messages in thread
From: Chad Gatesman @ 1998-12-16 15:37 UTC (permalink / raw)
  To: H.J. Lu; +Cc: leei, oliva, egcs, egcs-bugs

"H.J. Lu" wrote:

> > Yes.  I understood you, and I am saying it still does not work for me if I put -frepo on the link line.  Here are the versions
> > I am using.  Could it possibly be a problem there, or maybe a platform specific bug?  I am using Linux 2.1.129 on an Intel
> > Pentium using libc.so.5.4.46.
> >
> >
> > chad:/tmp> g++ -v
> > Reading specs from /usr/local/lib/gcc-lib/i586-pc-linux-gnulibc1/pgcc-2.91.60/specs
> > gcc version pgcc-2.91.60 19981201 (egcs-1.1.1 release)
> > chad:/tmp> ld -v
> > GNU ld version 2.9.1 (with BFD 2.9.1)
>
> If you use Linux, you should use my binutils. The
> current one is 2.9.1.0.19a.

This solved my problem.
Thanks for all the help  :)

--
Chad Gatesman
Software Engineer
Bristol Technology
http://www.bristol.com/


^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1998-12-16  9:55   ` Chad Gatesman
  1998-12-16 11:07     ` Lee Iverson
@ 1998-12-17  8:06     ` Horst von Brand
  1 sibling, 0 replies; 15+ messages in thread
From: Horst von Brand @ 1998-12-17  8:06 UTC (permalink / raw)
  To: Chad Gatesman; +Cc: Alexandre Oliva, egcs, egcs-bugs

Chad Gatesman <chadg@redrose.net> said:
> This can be reproduced very easily.  I will show you an example using
> egcs-1.1.1:
> 
> _________________________________
> 
> // repo-fail.cc
> #include <vector>
> 
> void main()
> {
>     vector<int> v;
> 
>     v.push_back(1);
> }
> 
> _________________________________
> 
> 
> chad:/tmp> g++ -c -frepo repo-fail.cc
> chad:/tmp> g++ -o repo-fail repo-fail.o
> repo-fail.o: In function `vector<int, __default_alloc_template<false, 0>
> >::push_back(int const &)':
> repo-fail.o(.vector<int, __default_alloc_template<0, 0>
> >::gnu.linkonce.t.push_back(int const &)+0x35): undefined reference to
> `vector<int, __default_alloc_template<false, 0> >::insert_aux(int *, int
> const &)'
> collect2: ld returned 1 exit status

egcs-19981213 (sparc-sun-solaris2.5.1) gives:

vonbrand@pincoya$ c++ -c -frepo repo.cc 
vonbrand@pincoya$ c++ -o repo repo.o  
collect: recompiling repo.cc
collect: relinking
-- 
Dr. Horst H. von Brand                       mailto:vonbrand@inf.utfsm.cl
Departamento de Informatica                     Fono: +56 32 654431
Universidad Tecnica Federico Santa Maria              +56 32 654239
Casilla 110-V, Valparaiso, Chile                Fax:  +56 32 797513

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1998-12-16 12:25             ` H.J. Lu
  1998-12-16 15:37               ` Chad Gatesman
@ 1998-12-17  8:21               ` Chad Gatesman
  1 sibling, 0 replies; 15+ messages in thread
From: Chad Gatesman @ 1998-12-17  8:21 UTC (permalink / raw)
  To: H.J. Lu; +Cc: leei, oliva, egcs, egcs-bugs

"H.J. Lu" wrote:

> > Yes.  I understood you, and I am saying it still does not work for me if I put -frepo on the link line.  Here are the versions
> > I am using.  Could it possibly be a problem there, or maybe a platform specific bug?  I am using Linux 2.1.129 on an Intel
> > Pentium using libc.so.5.4.46.
> >
> >
> > chad:/tmp> g++ -v
> > Reading specs from /usr/local/lib/gcc-lib/i586-pc-linux-gnulibc1/pgcc-2.91.60/specs
> > gcc version pgcc-2.91.60 19981201 (egcs-1.1.1 release)
> > chad:/tmp> ld -v
> > GNU ld version 2.9.1 (with BFD 2.9.1)
>
> If you use Linux, you should use my binutils. The
> current one is 2.9.1.0.19a.

Apparently, I spoke too soon.  This did fix the problem for simple samples, but after I tried my large project (it is very template
and STL intensive) out I get undefined symbols from some of the templates where I didn't with egcs-1.0.3.  Begin such a large
project it is hard to determine what may be causing this.

I decided to add some explicit instantiations of templates that looked to be problematic.  After doing so collect recompiled two of
the modules over and over in what appeared to be an infinite loop (I let it go overnight.).

I don't know if I can isolate this problem in a simple sample, but I may find the time to try.  Has anyone seen such strange
behavior with large STL or template projects using -frepo?  Are there any known outstanding bugs/issues with -frepo?

Thanks again.

--
 Chad Gatesman
 Software Engineer
 Bristol Technology
 http://www.bristol.com/



^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1998-12-16 15:37               ` Chad Gatesman
@ 1999-01-22  9:26                 ` Swen Thuemmler
  1999-01-22 10:54                   ` Chad Gatesman
  1999-01-31 23:58                   ` Swen Thuemmler
  0 siblings, 2 replies; 15+ messages in thread
From: Swen Thuemmler @ 1999-01-22  9:26 UTC (permalink / raw)
  To: Chad Gatesman; +Cc: egcs, egcs-bugs

On Wed, 16 Dec 1998, Chad Gatesman wrote:

> > > chad:/tmp> g++ -v
> > > Reading specs from /usr/local/lib/gcc-lib/i586-pc-linux-gnulibc1/pgcc-2.91.60/specs
> > > gcc version pgcc-2.91.60 19981201 (egcs-1.1.1 release)
> > > chad:/tmp> ld -v
> > > GNU ld version 2.9.1 (with BFD 2.9.1)
> >
> > If you use Linux, you should use my binutils. The
> > current one is 2.9.1.0.19a.
> 
> This solved my problem.
> Thanks for all the help  :)

Lucky man. I have the same problems on AIX 4.2

$ g++ -v  
Reading specs from /pronet/local/gnu/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/egcs-2.91.60/specs
gcc version egcs-2.91.60 19981201 (egcs-1.1.1 release)
$ g++ -frepo -c repo-fail.cc
$ g++ -frepo -o repo-fail repo-fail.o
ld: 0711-317 ERROR: Undefined symbol: .vector<int, __default_alloc_template<0, 0> >::insert_aux(int *, int const &)
ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more information.
collect2: ld returned 8 exit status
$ 

So there seems to be a bug lurking somewhere...

--Swen



^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1999-01-22  9:26                 ` Swen Thuemmler
@ 1999-01-22 10:54                   ` Chad Gatesman
  1999-01-31 23:58                     ` Chad Gatesman
  1999-01-31 23:58                   ` Swen Thuemmler
  1 sibling, 1 reply; 15+ messages in thread
From: Chad Gatesman @ 1999-01-22 10:54 UTC (permalink / raw)
  To: Swen.Thuemmler; +Cc: egcs, egcs-bugs

Swen Thuemmler wrote:

> On Wed, 16 Dec 1998, Chad Gatesman wrote:
>
> > > > chad:/tmp> g++ -v
> > > > Reading specs from /usr/local/lib/gcc-lib/i586-pc-linux-gnulibc1/pgcc-2.91.60/specs
> > > > gcc version pgcc-2.91.60 19981201 (egcs-1.1.1 release)
> > > > chad:/tmp> ld -v
> > > > GNU ld version 2.9.1 (with BFD 2.9.1)
> > >
> > > If you use Linux, you should use my binutils. The
> > > current one is 2.9.1.0.19a.
> >
> > This solved my problem.
> > Thanks for all the help  :)
>
> Lucky man. I have the same problems on AIX 4.2
>
> $ g++ -v
> Reading specs from /pronet/local/gnu/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/egcs-2.91.60/specs
> gcc version egcs-2.91.60 19981201 (egcs-1.1.1 release)
> $ g++ -frepo -c repo-fail.cc
> $ g++ -frepo -o repo-fail repo-fail.o
> ld: 0711-317 ERROR: Undefined symbol: .vector<int, __default_alloc_template<0, 0> >::insert_aux(int *, int const &)
> ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more information.
> collect2: ld returned 8 exit status
> $
>
> So there seems to be a bug lurking somewhere...

Actually...not so lucky.  After further use, I found that this still doesn't work with my much larger STL projects.  So I think I have to agree with you.  There is some bug in there somewhere.

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1999-01-22 10:54                   ` Chad Gatesman
@ 1999-01-31 23:58                     ` Chad Gatesman
  0 siblings, 0 replies; 15+ messages in thread
From: Chad Gatesman @ 1999-01-31 23:58 UTC (permalink / raw)
  To: Swen.Thuemmler; +Cc: egcs, egcs-bugs

Swen Thuemmler wrote:

> On Wed, 16 Dec 1998, Chad Gatesman wrote:
>
> > > > chad:/tmp> g++ -v
> > > > Reading specs from /usr/local/lib/gcc-lib/i586-pc-linux-gnulibc1/pgcc-2.91.60/specs
> > > > gcc version pgcc-2.91.60 19981201 (egcs-1.1.1 release)
> > > > chad:/tmp> ld -v
> > > > GNU ld version 2.9.1 (with BFD 2.9.1)
> > >
> > > If you use Linux, you should use my binutils. The
> > > current one is 2.9.1.0.19a.
> >
> > This solved my problem.
> > Thanks for all the help  :)
>
> Lucky man. I have the same problems on AIX 4.2
>
> $ g++ -v
> Reading specs from /pronet/local/gnu/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/egcs-2.91.60/specs
> gcc version egcs-2.91.60 19981201 (egcs-1.1.1 release)
> $ g++ -frepo -c repo-fail.cc
> $ g++ -frepo -o repo-fail repo-fail.o
> ld: 0711-317 ERROR: Undefined symbol: .vector<int, __default_alloc_template<0, 0> >::insert_aux(int *, int const &)
> ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more information.
> collect2: ld returned 8 exit status
> $
>
> So there seems to be a bug lurking somewhere...

Actually...not so lucky.  After further use, I found that this still doesn't work with my much larger STL projects.  So I think I have to agree with you.  There is some bug in there somewhere.

^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: -frepo not working?
  1999-01-22  9:26                 ` Swen Thuemmler
  1999-01-22 10:54                   ` Chad Gatesman
@ 1999-01-31 23:58                   ` Swen Thuemmler
  1 sibling, 0 replies; 15+ messages in thread
From: Swen Thuemmler @ 1999-01-31 23:58 UTC (permalink / raw)
  To: Chad Gatesman; +Cc: egcs, egcs-bugs

On Wed, 16 Dec 1998, Chad Gatesman wrote:

> > > chad:/tmp> g++ -v
> > > Reading specs from /usr/local/lib/gcc-lib/i586-pc-linux-gnulibc1/pgcc-2.91.60/specs
> > > gcc version pgcc-2.91.60 19981201 (egcs-1.1.1 release)
> > > chad:/tmp> ld -v
> > > GNU ld version 2.9.1 (with BFD 2.9.1)
> >
> > If you use Linux, you should use my binutils. The
> > current one is 2.9.1.0.19a.
> 
> This solved my problem.
> Thanks for all the help  :)

Lucky man. I have the same problems on AIX 4.2

$ g++ -v  
Reading specs from /pronet/local/gnu/lib/gcc-lib/powerpc-ibm-aix4.2.1.0/egcs-2.91.60/specs
gcc version egcs-2.91.60 19981201 (egcs-1.1.1 release)
$ g++ -frepo -c repo-fail.cc
$ g++ -frepo -o repo-fail repo-fail.o
ld: 0711-317 ERROR: Undefined symbol: .vector<int, __default_alloc_template<0, 0> >::insert_aux(int *, int const &)
ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more information.
collect2: ld returned 8 exit status
$ 

So there seems to be a bug lurking somewhere...

--Swen


^ permalink raw reply	[flat|nested] 15+ messages in thread

end of thread, other threads:[~1999-01-31 23:58 UTC | newest]

Thread overview: 15+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1998-12-14 11:13 -frepo not working? Chad Gatesman
1998-12-16  0:41 ` Alexandre Oliva
1998-12-16  9:55   ` Chad Gatesman
1998-12-16 11:07     ` Lee Iverson
1998-12-16 11:20       ` Chad Gatesman
1998-12-16 11:33         ` Lee Iverson
1998-12-16 12:04           ` Chad Gatesman
1998-12-16 12:25             ` H.J. Lu
1998-12-16 15:37               ` Chad Gatesman
1999-01-22  9:26                 ` Swen Thuemmler
1999-01-22 10:54                   ` Chad Gatesman
1999-01-31 23:58                     ` Chad Gatesman
1999-01-31 23:58                   ` Swen Thuemmler
1998-12-17  8:21               ` Chad Gatesman
1998-12-17  8:06     ` Horst von Brand

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).