Skip to content

Modular visual interface for GDB in Python

Notifications You must be signed in to change notification settings

marcusobrien/gdb-dashboard

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 

Repository files navigation

GDB dashboard

Modular visual interface for GDB in Python.

This comes as a standalone single-file .gdbinit which, among the other things, enables a configurable dashboard showing the most relevant information during the program execution. Its main goal is to reduce the number of GDB commands issued to inspect the current program status allowing the programmer to focus on the control flow instead.

Screenshot

Installation

Just place .gdbinit in your home directory, for example:

wget -P ~ git.io/.gdbinit

Features

  • Single GDB init file.

  • Write the dashboard to the main GDB console or to an external file/TTY.

  • Interaction with GDB using the native Python API.

  • Several default modules are included to address the most basic needs: source code, assembly, registers, etc.).

  • User-defined modules can be easily developed by extending a Python class.

  • Additional configuration files (both GDB and Python) are read from ~/.gdbinit.d/.

  • Fully stylable user interface and dynamic command prompt.

  • Optional syntax highlighting using the Pygments Python library.

  • No GDB command has been redefined, instead all the features are available as subcommands of the main dashboard command.

Default modules

Follows the list of bundled default modules. Refer to the GDB help system for the full syntax.

  • assembly shows the disassembled code surrounding the program counter. The instructions constituting the current statement are marked, if available.

  • history lists the last entries of the GDB value history.

  • memory allows to inspect memory regions.

  • registers shows the CPU registers and their values.

  • source show the program source code, if available.

  • stack shows the current stack trace including the function name and the file location, if available. Optionally list the frame arguments and locals too.

  • threads lists the currently available threads.

  • expressions watches user expressions.

Dashboard output

By default the dashboard is displayed in the GDB terminal but the -output command of both the dashboard and modules can change this behavior. When the output of a module is not specified then the global output is used.

Display the whole dashboard in another terminal

Dashboard in another terminal

It may be useful to move the dashboard to another terminal so the main terminal can be used exclusively for GDB commands and target I/O.

  1. start GDB in one terminal;

  2. open another terminal (e.g. tmux pane) and get its TTY with the tty command (e.g. /dev/ttys001, the name may be different for a variety of reasons);

  3. issue the command dashboard -output /dev/ttys001 to redirect the dashboard output to the newly created terminal;

  4. debug as usual.

Display each module in a separate terminal

Modules distributed in multiple terminals

It is also possible to display the output of one or more modules to individual terminals. If two or more modules share the same output, they will be stacked as usual.

  1. start GDB in one terminal;

  2. open another terminal and get its TTY with the tty command;

  3. pick a module, say source, then issue the command dashboard source -output /dev/ttys001 to redirect its output to the newly created terminal;

  4. repeat for any other modules;

  5. debug as usual.

Display the whole dashboard in a web browser

Dashboard in a web browser

Pushing this even further, one could use a web browser as an auxiliary terminal using gotty. Of course, using the method described above, one can also display the output of individual modules in one or more web browser instances.

  1. start GDB in one terminal;

  2. open another terminal and execute gotty sh -c 'tty; cat';

  3. open a web browser, navigate to http://localhost:8080 and note the TTY;

  4. issue the command dashboard -output /dev/ttys001 to redirect the dashboard output to the web browser;

  5. debug as usual.

Commands

The GDB documentation is available at help dashboard. Just like any GDB command, abbreviations are possible, so da, dash, etc. all resolve to dashboard.

dashboard

This is the root command and it is used to manually redisplay the dashboard.

dashboard -configuration [<file>]

Display and optionally write to <file> the current configuration (layout, styles, outputs). This command allows to configure the dashboard live then make the changes permanent, for example:

dashboard -configuration ~/.gdbinit.d/auto

dashboard -output [<file>]

By default the dashboard is written to the GDB console but it is possible to redirect its output to a file or even to another terminal. If the target is a valid terminal TTY then its width is used to format the dashboard, otherwise fall back to the width of the main GDB console.

Without argument reset this setting to the default.

dashboard -enabled [on|off]

Enable or disable the automatic display of the dashboard whenever the target program stops. The dashboard is enabled by default and even when it is disabled, it can be manually displayed with dashboard.

Sometimes it may be convenient to redraw the dashboard even if the target program has not changed its execution status, for example when the programmer switches the currently selected frame with the up or down commands. It is possible to do so by setting up some GDB hooks in the user-defined init file, for example:

define hookpost-up
dashboard
end

define hookpost-down
dashboard
end

dashboard -layout [<directive>...]

By default, all the modules are enabled and placed within the dashboard in alphabetical order. As the number of modules grows, it is important to decide which modules will be part of the dashboard, and where.

Each directive is in the form [!]<module>, when the ! is present then the corresponding module is disabled by default. The order of directives denotes the display order within the dashboard. For example:

dashboard -layout source !assembly stack

Modules which do not appear in the list are disabled and placed after the last element in alphabetical order.

When executed without arguments, this command lists all the available modules in the form of a list of directives followed by the status of the output files of the modules.

dashboard -style [<name> [<value>]]

Access to the stylable attributes of the dashboard, see Stylable attributes. For example, to change the prompt to something more familiar:

dashboard -style prompt '(gdb)'

The argument is parsed as a Python literal and converted to the proper type.

When only the name is specified this command shows the current value, whereas without arguments prints all the attributes.

Modules subcommands

Every module adds its own subcommand dashboard <module> which is used to toggle the enable flag and to redisplay the dashboard.

Modules may also declare additional subcommands, see help dashboard <module> from GDB.

There are two additional predefined subcommands: -style and -output.

-style

If a module declares some stylable attributes then the command dashboard <module> -style will be available. Its functioning is equivalent to the dashboard -style command but it does apply to a module.

-output

Similarly, the dashboard <module> -output mimics the dashboard -style command but allows a finer grain of operation.

Configuration

Files in ~/.gdbinit.d/ are executed in alphabetical order, but the preference is given to Python files. If there are subdirectories, they are walked recursively. The idea is to keep separated the custom modules definition from the configuration itself.

By convention, the main configuration file should be placed in ~/.gdbinit.d/ (say ~/.gdbinit.d/init) and can be used to tune the dashboard styles and modules configuration but also the usual GDB parameters.

The alternative is to hard code changes in the provided .gdbinit, to do so just add new modules and GDB settings under # Default modules and # Better GDB defaults respectively.

Per-project configuration

GDB natively support the auto-loading of .gdbinit files, this can come in handy to set up a different dashboard style according to the current project type (e.g., C++ development, reverse engineering, etc.). This feature is disabled by default for security reasons. To enable the auto-loading everywhere in the file system add this line to the main configuration file:

set auto-load safe-path /

Stylable attributes

There is number of attributes that can be used to customize the aspect of the dashboard and of its modules. They are documented within the GDB help system. For what concerns the dashboard itself it can be reached with:

help dashboard -style

Whereas for modules:

help dashboard <module> -style

ANSI escape codes

Colors and text styles are specified using ANSI escape codes. For example setting a style to 1;31 will produce ^[[1;31m, which will result in displaying the text red (31) and bright (1). The ANSI output can be disabled by setting the ansi attribute to False (note that this will not affect the command prompt).

Syntax highlighting

When the ansi attribute is set to True the Pygments Python library may be used by modules to provide syntax highlighting of the source code.

The syntax_highlighting stylable attribute is a string which defines the Pygments style to use.

The list of all the available styles can be obtained with (from GDB itself):

python from pygments.styles import get_all_styles as styles
python for s in styles(): print(s)

Dividers

A divider is basically a terminal-wide horizontal line with an optional label. Primary dividers are those used to separate the modules, whereas secondary dividers may be used inside modules to logically separate different sections. When a section or module is empty then the styles used for the divider are those with the off qualifier.

Common styles

These are general purpose ANSI styles defined for convenience and used within the default modules.

  • style_selected_1
  • style_selected_2
  • style_low
  • style_high
  • style_error

Custom modules

The idea of custom modules is that they provide ways to access readonly information from the target program status; it is safe to assume that they will be queried during the program execution only.

Custom modules must inherit the Dashboard.Module class and define some methods:

  • label returns the module label which will appear in the divider.

  • lines return a list of strings which will form the module content. When a module is temporarily unable to produce its content, it should return an empty list; its divider will then use the styles with the off qualifier.

The name of a module is automatically obtained by the class name.

Modules are instantiated once at initialization time and kept during the whole the GDB session.

Optionally, a module may include a description which will appear in the GDB help system by specifying a Python docstring for the class.

Optionally, a module may define stylable attributes by defining the attributes method returning a dictionary in which the key is the attribute name and the value is another dictionary:

  1. default is the initial value for this attribute.

  2. doc is the documentation of this attribute which will appear in the GDB help system. This key can be omitted.

  3. name is the name of the attribute of the Python object, defaults to the key value.

  4. type is the type of this attribute, it is used to coerce the value passed as an argument to the proper type, or raise an exception. This key defaults to the str type.

  5. check is a control callback which accept the coerced value and returns True if the value satisfies the constraint and False otherwise. This key is optional, when omitted no check is performed.

Optionally, a module may declare subcommands by defining the commands method returning a dictionary in which the key is the command name and the value is another dictionary:

  1. action is the callback to be executed which accepts the raw input string from the GDB prompt. Callbacks may raise exceptions to notify erroneous situations which message will be shown automatically to the user.

  2. doc is the command documentation.

  3. completion is the completion policy, one of the gdb.COMPLETE_* constants defined in the reference manual. This key is optional and defaults to None which is equivalent to gdb.COMPLETE_NONE.

Common functions

A number of auxiliary common functions are defined in the global scope, they can be found in the provided .gdbinit and concern topics like ANSI output, divider formatting, conversion callbacks, etc. They should be more or less self-documented, some usage examples can be found within the bundled default modules.

Example

Default modules already provide a good example, but here is a simple module which may be used as a template for new custom modules, it allows the programmer to note down some snippets of text during the debugging session.

class Notes(Dashboard.Module):
    """Simple user-defined notes."""

    def __init__(self):
        self.notes = []

    def label(self):
        return 'Notes'

    def lines(self, term_width, style_changed):
        out = []
        for note in self.notes:
            out.append(note)
            if self.divider:
                out.append(divider())
        return out[:-1] if self.divider else out

    def add(self, arg):
        if arg:
            self.notes.append(arg)
        else:
            raise Exception('Cannot add an empty note')

    def clear(self, arg):
        self.notes = []

    def commands(self):
        return {
            'add': {
                'action': self.add,
                'doc': 'Add a note.'
            },
            'clear': {
                'action': self.clear,
                'doc': 'Remove all the notes.'
            }
        }

    def attributes(self):
        return {
            'divider': {
                'doc': 'Divider visibility flag.',
                'default': True,
                'type': bool
            }
        }

To use the above just save it in a Python file, say notes.py, inside ~/.gdbinit.d/, the following commands (together with the help) will be available:

dashboard notes
dashboard notes add
dashboard notes clear
dashboard notes -style

Minimal requirements

GDB dashboard requires at least GDB 7.7 compiled with Python 2.7 in order to work properly.

See #1 for more details/workarounds.

Additional GDB front ends

GDB dashboard is not meant to work seamlessly with additional front ends, e.g., TUI, Nemiver, QtCreator, etc.

There are basically two options to work around this:

  • if the main debugging tool is GDB dashboard then it is recommended to prevent the front end from loading the .gdbinit file, they usually have an option to do so;

  • otherwise it is possible to load GDB dashboard manually, that is, install as usual then:

     mv ~/.gdbinit ~/.gdb-dashboard
    

    finally load it when needed from the GDB shell:

     source ~/.gdb-dashboard
    

Resources

License

Copyright (c) 2015-2017 Andrea Cardaci [email protected]

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

About

Modular visual interface for GDB in Python

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 100.0%