Debugging Octave: Difference between revisions

From Octave
Jump to navigation Jump to search
(Created page with "= Basic debugging processes = == Error & trace the stack == = Tools for debugging = == GBD == === Most used commands === [http: link to gbd manual] == Emacs == [http: link t...")
 
(→‎Tools for debugging: __debug_octave__ is in released versions now)
 
(44 intermediate revisions by 13 users not shown)
Line 1: Line 1:
= Basic debugging processes =
= Preliminaries =
== Error & trace the stack ==
 
Since compilation of all the source from scratch can take long it is good to have a source folder where most of the source has been compiled. To do this, you can create a parallel build:
 
<syntaxhighlight lang="bash">
mkdir dbg-octave
cd dbg-octave
/path/to/octave/source/configure FFLAGS=-g CFLAGS=-g CXXFLAGS=-g --enable-address-sanitizer-flags --prefix=/opt/dbg-octave
make # or make -jN where N is the number of CPU cores
</syntaxhighlight>
 
This will create a new build of Octave in a different directory without optimizations (no -O2 gcc parameter) and with debug symbols compiled in. This build is useful for debugging Octave itself.


= Tools for debugging =
= Tools for debugging =
== GBD ==
=== Most used commands ===
[http: link to gbd manual]


== Emacs ==
There are different tools for debugging. This article concentrates on describing how to use <code>gdb</code>.
[http: link to emacs manual]
 
If you run Octave from a build tree, execute <code>./run-octave -g</code> to start a gdb session that is prepared to run Octave (with the necessary environment correctly set up).  Note that when Octave runs in GUI mode, it forks at startup on Linux and MacOS systems, so this method will only work if <code>gdb</code> correctly follows the process across the <code>fork</code> and <code>exec</code> system calls.
 
Alternatively, you can attach a debugger to a running Octave session.  Current versions of Octave include the command <code>__debug_octave__</code> to manage the details.  Executing this command at the Octave prompt should open a separate window for a debugger session attached to the current Octave process.  On Linux systems, the default terminal window is <code>gnome-terminal</code>.  On MacOS systems, the default debugger is <code>lldb</code>.
 
For some kinds of errors on some OS, the last approach might not be useful. The OS might kill the shell that runs gdb as soon as the spawning process (i.e. Octave) crashes. In that case, you can attach to Octave from an "independent" shell. Execute <code>getpid ()</code> in Octave and take note of the displayed *PID*. Open a shell and execute <code>gdb -p *PID*</code> (replace <code>*PID*</code> with the actual PID). On Windows, use the msys2 shell that can be started with the file <code>cmdshell.bat</code> in Octave's installation folder.
 
Independent of how <code>gdb</code> was started and Octave was attached to it, it is now possible to issue gdb commands on the <code>(gdb)</code> prompt. See e.g. the [https://sourceware.org/gdb/download/onlinedocs/gdb/index.html gdb documentation]. To return to Octave while gdb is still attached to it, execute <code>continue</code> (or <code>c</code>) at the <code>(gdb)</code> prompt.
 
*NOTE:  Ubuntu introduced a patch to disallow ptracing of non-child processes by non-root users - i.e. only a process which is a parent of another process can ptrace it for normal users - whilst root can still ptrace every process.
That's why gdb won't be able to link to the octave process if you start gdb from within an Octave session using the <code>__debug_octave__</code> command.
 
You can temporarily disable this restriction by doing:
<syntaxhighlight lang="bash">
echo 0 | sudo tee /proc/sys/kernel/yama/ptrace_scope
</syntaxhighlight>
and then reopen <code>gdb</code> using the command mentioned above from within an Octave session or if you have admin rights you can simply do:
<syntaxhighlight lang="bash">
__debug_octave__ ("gnome-terminal --command 'sudo gdb -p %d'")
</syntaxhighlight>
 
= Debugging oct-files =
 
To debug oct-files, avoid making any optimization during compilation. Use <code>export CXXFLAGS="-ggdb -Wall -O0"</code> for C++ code or <code>export CFLAGS="-ggdb -Wall -O0"</code> for C code to suppress optimization. Compile the oct-file with the debug flag <code>-g</code> which enables debug symbols
 
<syntaxhighlight lang="bash">
mkoctfile -g file.cpp
</syntaxhighlight>
 
In next step you will use GNU debugger or gdb. The symbols from your oct-file will only be available to gdb once the oct-file is loaded in Octave.  To do that without executing any functions from the oct-file, you can ask Octave to display the help for the oct-file:
 
<syntaxhighlight lang="bash">
octave> help file
</syntaxhighlight>
 
start now the GNU debugger with octave by following the instructions above.
 
Now you can set a breakpoint in the line of interest of your oct-file in gdb prompt:
 
<syntaxhighlight lang="bash">
(gdb) b file.cpp:40
</syntaxhighlight>
 
by typing c the execution of octave will continue and you can run your oct-file directly or via an m-script.
 
<syntaxhighlight lang="octave">
octave> x = file(y)
</syntaxhighlight>
 
the debugger will stop on the above defined line and you can start debugging according to the manual of GNU debugger.
 
== Producing a stack trace ==
 
Sometimes Octave might crash, meaning, it terminates abruptly and returns control to the operating system shell. In these cases, it is very helpful to produce a stack trace to diagnose the problem. For this, it can be useful to (re)compile Octave with debugging symbols. Otherwise, the stack trace can be harder to read and optimizations might make debugging more difficult. (But it is also possible to produce a stack trace with a "standard" build.)
 
Attach <code>gdb</code> to Octave as described before and return execution to Octave. Then, execute whatever commands are necessary to cause Octave to crash. At that point, you will be back in the gdb session. Type <code>where</code> or <code>bt</code> at the gdb prompt to obtain a stack trace.
 
When running in GUI mode or debugging threading issues, it is usually useful to get information about all the execution threads at the point of the crash.  To do that, use the gdb command <code>thread apply all bt</code>.
 
You could also get some help from your system tools. In most GNU/Linux systems whenever a crash happens in a software, a <i>core dump</i> will be generated. These core dumps are handled by a registered component of the system and finally might be stored somewhere in the directory tree. You should find them, view them and inspect them.
 
=== Where are core dumps stored? ===
 
It differs on each system. First you should see how core dumps are handled on your system. To do so, type this in a shell terminal:
<syntaxhighlight lang="bash">
$ cat /proc/sys/kernel/core_pattern
</syntaxhighlight>
This may print a file name pattern along with a path where all core dumps will be saved <b>ONLY</b> if it does not start by a pipe or '|'. If it does, <i>the kernel will treat the rest of the pattern as a command to run.  The core dump will be written to the standard input of that program instead of to a file</i>, and you need to consult that program's help or manual.
 
=== How to view a core dump? ===
 
To do this you should use gdb. Core dumps are saved under root user, so you may need to change owner of the core dump you are interested in if you are not logged in as root. After that type in the terminal:
<syntaxhighlight lang="bash">
gdb octave -c <Path to core dump>
</syntaxhighlight>
 
Always expect some warnings from gdb in a few first times of doing this. Most likely gdb will tell you that:
 
1. The core dump file is not in the correct format. It is the case if the core dump handler of your system compresses core dumps before storing them, and you need to decompress the core dump first.
 
2. Core dump was generated by a-path-to/octave-gui. Then quit gdb and start it again by:
<syntaxhighlight lang="bash">
gdb octave-gui -c <Path to core dump>
</syntaxhighlight>
 
3. Some debug info are missing. In this case gdb itself will tell you how to install them. Install them and start gdb again.
 
If everything worked fine, you can use <code>where</code> command in gdb prompt to see a full stack trace of the crash.
 
=== Helpful gdb commands ===
 
[http://www.gnu.org/software/gdb/documentation gdb documentation]
 
The following command shows the back trace of all threads belonging to the Octave process:
<syntaxhighlight lang="bash">
(gdb) thread apply all bt
</syntaxhighlight>
 
For debugging octave_value variables (e.g. <code>my_octave_value</code>) to find out what the variable actually is (instead of just it's base class):
<syntaxhighlight lang="bash">
(gdb) print *(my_octave_value.rep)
</syntaxhighlight>
 
The file <code>etc/gdbinit</code> in the Octave repository contains some macros that can be helpful:
* <code>display-dims DIM_VECTOR</code>: Display the contents of an Octave dimension vector.
* <code>display-dense-array ARRAY</code>: Display the contents of an ordinary, i.e., dense Octave array.
* <code>display-sparse-array SPARSE_ARRAY</code>: Display the contents of a sparse Octave array.
* <code>show-octave-dbstack</code>: Display the contents of the current Octave debugging stack.
 
== ddd ==
 
[http://www.gnu.org/software/ddd GUI for gdb]


[[Category:Development]]
[[Category:Development]]

Latest revision as of 15:25, 22 December 2022

Preliminaries[edit]

Since compilation of all the source from scratch can take long it is good to have a source folder where most of the source has been compiled. To do this, you can create a parallel build:

mkdir dbg-octave
cd dbg-octave
/path/to/octave/source/configure FFLAGS=-g CFLAGS=-g CXXFLAGS=-g --enable-address-sanitizer-flags --prefix=/opt/dbg-octave
make # or make -jN where N is the number of CPU cores

This will create a new build of Octave in a different directory without optimizations (no -O2 gcc parameter) and with debug symbols compiled in. This build is useful for debugging Octave itself.

Tools for debugging[edit]

There are different tools for debugging. This article concentrates on describing how to use gdb.

If you run Octave from a build tree, execute ./run-octave -g to start a gdb session that is prepared to run Octave (with the necessary environment correctly set up). Note that when Octave runs in GUI mode, it forks at startup on Linux and MacOS systems, so this method will only work if gdb correctly follows the process across the fork and exec system calls.

Alternatively, you can attach a debugger to a running Octave session. Current versions of Octave include the command __debug_octave__ to manage the details. Executing this command at the Octave prompt should open a separate window for a debugger session attached to the current Octave process. On Linux systems, the default terminal window is gnome-terminal. On MacOS systems, the default debugger is lldb.

For some kinds of errors on some OS, the last approach might not be useful. The OS might kill the shell that runs gdb as soon as the spawning process (i.e. Octave) crashes. In that case, you can attach to Octave from an "independent" shell. Execute getpid () in Octave and take note of the displayed *PID*. Open a shell and execute gdb -p *PID* (replace *PID* with the actual PID). On Windows, use the msys2 shell that can be started with the file cmdshell.bat in Octave's installation folder.

Independent of how gdb was started and Octave was attached to it, it is now possible to issue gdb commands on the (gdb) prompt. See e.g. the gdb documentation. To return to Octave while gdb is still attached to it, execute continue (or c) at the (gdb) prompt.

  • NOTE: Ubuntu introduced a patch to disallow ptracing of non-child processes by non-root users - i.e. only a process which is a parent of another process can ptrace it for normal users - whilst root can still ptrace every process.

That's why gdb won't be able to link to the octave process if you start gdb from within an Octave session using the __debug_octave__ command.

You can temporarily disable this restriction by doing:

echo 0 | sudo tee /proc/sys/kernel/yama/ptrace_scope

and then reopen gdb using the command mentioned above from within an Octave session or if you have admin rights you can simply do:

__debug_octave__ ("gnome-terminal --command 'sudo gdb -p %d'")

Debugging oct-files[edit]

To debug oct-files, avoid making any optimization during compilation. Use export CXXFLAGS="-ggdb -Wall -O0" for C++ code or export CFLAGS="-ggdb -Wall -O0" for C code to suppress optimization. Compile the oct-file with the debug flag -g which enables debug symbols

mkoctfile -g file.cpp

In next step you will use GNU debugger or gdb. The symbols from your oct-file will only be available to gdb once the oct-file is loaded in Octave. To do that without executing any functions from the oct-file, you can ask Octave to display the help for the oct-file:

octave> help file

start now the GNU debugger with octave by following the instructions above.

Now you can set a breakpoint in the line of interest of your oct-file in gdb prompt:

(gdb) b file.cpp:40

by typing c the execution of octave will continue and you can run your oct-file directly or via an m-script.

octave> x = file(y)

the debugger will stop on the above defined line and you can start debugging according to the manual of GNU debugger.

Producing a stack trace[edit]

Sometimes Octave might crash, meaning, it terminates abruptly and returns control to the operating system shell. In these cases, it is very helpful to produce a stack trace to diagnose the problem. For this, it can be useful to (re)compile Octave with debugging symbols. Otherwise, the stack trace can be harder to read and optimizations might make debugging more difficult. (But it is also possible to produce a stack trace with a "standard" build.)

Attach gdb to Octave as described before and return execution to Octave. Then, execute whatever commands are necessary to cause Octave to crash. At that point, you will be back in the gdb session. Type where or bt at the gdb prompt to obtain a stack trace.

When running in GUI mode or debugging threading issues, it is usually useful to get information about all the execution threads at the point of the crash. To do that, use the gdb command thread apply all bt.

You could also get some help from your system tools. In most GNU/Linux systems whenever a crash happens in a software, a core dump will be generated. These core dumps are handled by a registered component of the system and finally might be stored somewhere in the directory tree. You should find them, view them and inspect them.

Where are core dumps stored?[edit]

It differs on each system. First you should see how core dumps are handled on your system. To do so, type this in a shell terminal:

$ cat /proc/sys/kernel/core_pattern

This may print a file name pattern along with a path where all core dumps will be saved ONLY if it does not start by a pipe or '|'. If it does, the kernel will treat the rest of the pattern as a command to run. The core dump will be written to the standard input of that program instead of to a file, and you need to consult that program's help or manual.

How to view a core dump?[edit]

To do this you should use gdb. Core dumps are saved under root user, so you may need to change owner of the core dump you are interested in if you are not logged in as root. After that type in the terminal:

gdb octave -c <Path to core dump>

Always expect some warnings from gdb in a few first times of doing this. Most likely gdb will tell you that:

1. The core dump file is not in the correct format. It is the case if the core dump handler of your system compresses core dumps before storing them, and you need to decompress the core dump first.

2. Core dump was generated by a-path-to/octave-gui. Then quit gdb and start it again by:

gdb octave-gui -c <Path to core dump>

3. Some debug info are missing. In this case gdb itself will tell you how to install them. Install them and start gdb again.

If everything worked fine, you can use where command in gdb prompt to see a full stack trace of the crash.

Helpful gdb commands[edit]

gdb documentation

The following command shows the back trace of all threads belonging to the Octave process:

(gdb) thread apply all bt

For debugging octave_value variables (e.g. my_octave_value) to find out what the variable actually is (instead of just it's base class):

(gdb) print *(my_octave_value.rep)

The file etc/gdbinit in the Octave repository contains some macros that can be helpful:

  • display-dims DIM_VECTOR: Display the contents of an Octave dimension vector.
  • display-dense-array ARRAY: Display the contents of an ordinary, i.e., dense Octave array.
  • display-sparse-array SPARSE_ARRAY: Display the contents of a sparse Octave array.
  • show-octave-dbstack: Display the contents of the current Octave debugging stack.

ddd[edit]

GUI for gdb