insight for Solaris

insight for Solaris

[Yakov Nekritch]

Hello,


I tried to build Insight (under SPARC
Solaris 5.8 ) and got the following errors :

.........
rm -f gdb
gcc -g -O2        -o gdb gdb.o libgdb.a     ../bfd/libbfd.a 
../readline/libreadline.a ../opcodes/libopcodes.a 
../libiberty/libiberty.a     ../libgui/src/libgui.a 
-L/opt/scratch/yasha/insight-6.1/itcl/itcl -litcl3.2 
-L/opt/scratch/yasha/insight-6.1/itcl/itk -litk3.2 
-L/opt/scratch/yasha/insight-6.1/tk/unix -ltk8.4 
-Wl,-R,/opt/scratch/yasha/insight-6.1/local/lib 
-L/opt/scratch/yasha/insight-6.1/tcl/unix -ltcl8.4 
-L/usr/local/X11R6/lib -lX11 -ldl  -lsocket -lnsl -lm -ldl -lsocket 
-lnsl -lm -lncurses   -liconv  ../libiberty/libiberty.a
Undefined                       first referenced
  symbol                             in file
libintl_dgettext                    ../bfd/libbfd.a(bfd.o)
libintl_gettext                     libgdb.a(main.o)
libintl_textdomain                  libgdb.a(main.o)
libintl_bindtextdomain              libgdb.a(main.o)
ld: fatal: Symbol referencing errors. No output written to gdb
collect2: ld returned 1 exit status
*** Error code 1

Stop.
make: stopped in /opt/scratch/yasha/insight-6.1/gdb


Can someone help me with this?
Is it probably an error of glibc? But I could not find
those symbols (libintl_dgettext,  etc. ).

Thank you,
Yakov

==========================================================================
Dr. Yakov Nekrich		Phone:  +49-(0)228-734209
Computer Science Department	Fax:    +49-(0)228-734440
University of Bonn		e-mail: yasha@cs.uni-bonn.de
Römerstrasse 164,		http://theory.cs.uni-bonn.de/~yasha
53117 Bonn
Germany
==========================================================================

Re: insight for Solaris

[Keith Seitz]

On Fri, 2005-01-21 at 03:01, Yakov Nekritch wrote:
> rm -f gdb
> gcc -g -O2        -o gdb gdb.o libgdb.a     ../bfd/libbfd.a 
> ../readline/libreadline.a ../opcodes/libopcodes.a 
> ../libiberty/libiberty.a     ../libgui/src/libgui.a 
> -L/opt/scratch/yasha/insight-6.1/itcl/itcl -litcl3.2 
> -L/opt/scratch/yasha/insight-6.1/itcl/itk -litk3.2 
> -L/opt/scratch/yasha/insight-6.1/tk/unix -ltk8.4 
> -Wl,-R,/opt/scratch/yasha/insight-6.1/local/lib 
> -L/opt/scratch/yasha/insight-6.1/tcl/unix -ltcl8.4 
> -L/usr/local/X11R6/lib -lX11 -ldl  -lsocket -lnsl -lm -ldl -lsocket 
> -lnsl -lm -lncurses   -liconv  ../libiberty/libiberty.a
> Undefined                       first referenced
>   symbol                             in file
> libintl_dgettext                    ../bfd/libbfd.a(bfd.o)
> libintl_gettext                     libgdb.a(main.o)
> libintl_textdomain                  libgdb.a(main.o)
> libintl_bindtextdomain              libgdb.a(main.o)
> ld: fatal: Symbol referencing errors. No output written to gdb
> collect2: ld returned 1 exit status
> *** Error code 1

This is an error compiling gdb, not insight. Nonetheless, it looks like
your system does not have GNU gettext available. Try disabling native
language support in the build (pass "--disable-nls" to configure). If
that fails, try finding/building gettext for solaris.

If that still fails, ask on the gdb list for more help.

Keith

Re: insight for Solaris

[Keith Seitz]

[Back on list]

On Fri, 2005-01-21 at 06:54, Yakov Nekritch wrote:
> Hello,
> 
> thank you very much, this worked. Of course, with many warnings, like
> 
> gdb_locale.h:40: warning: `textdomain' redefined
> ../intl/libintl.h:171: warning: this is the location of the previous 
> definition
> gdb_locale.h:41: warning: `bindtextdomain' redefined
> ../intl/libintl.h:172: warning: this is the location of the previous 
> definition
> 
> But now I am getting strange errors in plugins/rhabout/Makefile

I've not heard of the above warnings. You might try asking on the gdb
list about them. [It could very well be that "--disable-nls" is not the
right way to go.]

> make: 
> "/opt/scratch/yasha/insight-6.1/gdb/gdbtk/plugins/rhabout/Makefile" line 
> 58: Missing dependency operator
> make: 
> "/opt/scratch/yasha/insight-6.1/gdb/gdbtk/plugins/rhabout/Makefile" line 
> 60: Need an operator
> make: 
> "/opt/scratch/yasha/insight-6.1/gdb/gdbtk/plugins/rhabout/Makefile" line 
> 62: Need an operator
> make: 
> "/opt/scratch/yasha/insight-6.1/gdb/gdbtk/plugins/rhabout/Makefile" line 
> 69: Missing dependency operator
> 
> 
> and much more. Do I need a special version of make (like GNU make )?

Ah, something that I know about! Yes, as a matter of fact, to "build"
the plug-ins, GNU make is required.

However, you should be able to delete/rename the plugins folder --
they're not necessary. You'll need to use "make -k" (or whatever your
make program's equivalent) to force it to complete, since a missing
plugin directory will cause a build error.

Keith

Re: insight for Solaris

[Yakov Nekritch]

Hello,

I was able to finish build  using gmake.
Thank you very much for your help.

However I think, the debugger still does not work properly.
(The program with which I tried it, stops when I press "Next" button ).

Some additional testing is still needed, but I think this is a gdb
problem.

Is it possible to "connect" the Insight with an older version of
GDB which is already installed on my computer and works properly?
(Instead of the version which is a part of Insight package ).


Thank you,
Yakov




Keith Seitz wrote:
> [Back on list]
> 
> On Fri, 2005-01-21 at 06:54, Yakov Nekritch wrote:
> 
>>Hello,
>>
>>thank you very much, this worked. Of course, with many warnings, like
>>
>>gdb_locale.h:40: warning: `textdomain' redefined
>>../intl/libintl.h:171: warning: this is the location of the previous 
>>definition
>>gdb_locale.h:41: warning: `bindtextdomain' redefined
>>../intl/libintl.h:172: warning: this is the location of the previous 
>>definition
>>
>>But now I am getting strange errors in plugins/rhabout/Makefile
> 
> 
> I've not heard of the above warnings. You might try asking on the gdb
> list about them. [It could very well be that "--disable-nls" is not the
> right way to go.]
> 
> 
>>make: 
>>"/opt/scratch/yasha/insight-6.1/gdb/gdbtk/plugins/rhabout/Makefile" line 
>>58: Missing dependency operator
>>make: 
>>"/opt/scratch/yasha/insight-6.1/gdb/gdbtk/plugins/rhabout/Makefile" line 
>>60: Need an operator
>>make: 
>>"/opt/scratch/yasha/insight-6.1/gdb/gdbtk/plugins/rhabout/Makefile" line 
>>62: Need an operator
>>make: 
>>"/opt/scratch/yasha/insight-6.1/gdb/gdbtk/plugins/rhabout/Makefile" line 
>>69: Missing dependency operator
>>
>>
>>and much more. Do I need a special version of make (like GNU make )?
> 
> 
> Ah, something that I know about! Yes, as a matter of fact, to "build"
> the plug-ins, GNU make is required.
> 
> However, you should be able to delete/rename the plugins folder --
> they're not necessary. You'll need to use "make -k" (or whatever your
> make program's equivalent) to force it to complete, since a missing
> plugin directory will cause a build error.
> 
> Keith
> 

Re: insight for Solaris

[Keith Seitz]

On Thu, 2005-01-27 at 04:00, Yakov Nekritch wrote:

> Is it possible to "connect" the Insight with an older version of
> GDB which is already installed on my computer and works properly?
> (Instead of the version which is a part of Insight package ).

Unfortunately not. Right now, Insight *is* gdb: they are the same
executable. As far as I know, Insight is the only gdb frontend that does
this.

I've always had plans to augment this by allowing insight to attach to
any MI-enabled gdb running anywhere, but, alas, that does not appear to
be in the cards.

Keith

RE: insight for Solaris

[Keith Seitz]

On Fri, 2005-01-21 at 09:21, KiranKumar B Shivananda wrote:
> Yes the command set
> *(int*)ADDRESS=INTEGER" is working
> 
> Can you give more info about
> target triple, host, target

What's the output of "show version"? That will give the host and target
"triples", i.e., i686-pc-linux, arm-unknown-elf (or simply arm-elf), and
so on.

The other bit is what target you are debugging: native? a board (which
one, what's your Target Settings dialog say the connection is)?
simulator?

Keith

RE: insight for Solaris

[Keith Seitz]

On Fri, 2005-01-21 at 08:29, KiranKumar B Shivananda wrote:
> But I didnt find gdb command to set memory.

Ah, oh. My bad. I forgot about that...

Well, here's what you can try just to see if gdb is working: "set
*(int*)ADDRESS=INTEGER". Change "int" to "char", "long", or whatnot
depending on the size of the memory chunk being modified.

If this works, I will probably need more info: target triple, host,
target. Maybe more.

Keith

RE: insight for Solaris

[Keith Seitz]

On Fri, 2005-01-21 at 06:00, KiranKumar B Shivananda wrote:
> I have a problem during writing to memory in insight.
> 
> I tried to write 8 bytes to memory. It gave the error "Integer value
> too large to represent".
> 
> Is it correct or a bug.

This is one of those things that sounds like a gdb problem. Can you
reproduce the problem using command-line gdb? If so, I highly recommend
sending your question to the gdb list. Be sure to include the text from
your session which demonstrates the problem.

There are far more people there who are all far more knowledgeable about
gdb than I am now. They will be able to help much faster and more
authoritatively than I will be able to.

Keith