Contents
This package provides an integration of the Python debugger pdb into the
VIM editor.
vimpdb has been used successfully under Linux, Mac OSX and Windows.
It is compatible with Python 2.7, 2.6, 2.5 and 2.4. It is not compatible with Python 3.1 (it should be the same for 3.0).
vimpdb requires an installation of VIM that supports both python and
clientserver options.
Find out if it is the case by issuing the following command at the VIM prompt:
:version
If the options are supported, you will see +clientserver and +python in the
output. In the opposite case, you will see -clientserver or -python.
On Linux and Windows, the default VIM build should already be server-enabled.
On Windows, the python option compiled in VIM depends on a specific Python
version. Find out if that specific version is installed and works in VIM by
issuing the following command at the VIM prompt:
:python import sys; print sys.version
On Mac OSX, you'll want to use MacVIM. MacVIM also has the python option
compiled in by default.
Just install this package using easy_install
$ easy_install vimpdb
You can obviously also use pip.
If you look inside the package, you will see a VIM script file:vimpdb.vim. Do not move it to VIM configuration directory (like~/.vim/plugin). vimpdb knows how to make the script available to VIM.
vimpdb tries to avoid depending on any user configuration. If it cannot detect the right configuration by itself, it will ask a few questions which you should be able to answer easily.
When launched, vimpdb looks for its RC file : ~/.vimpdbrc. If it does
not find it, vimpdb creates that file for you from default values.
vimpdb tries a set of default values that should work. It checks if those default values are appropriate. If the default values do not work, vimpdb asks for other values interactively until it has checked that the values provided actually work.
The default values per OS are listed hereunder.
For Linux:
vim_client_script = vim vim_server_script = vim server_name = VIM port = 6666
For MacOSX:
vim_client_script = mvim vim_server_script = mvim server_name = VIM port = 6666
For Windows:
vim_client_script = vim.exe vim_server_script = gvim.exe server_name = VIM port = 6666
See below for details about each option.
You are obviously allowed to create and tune that RC file. Nevertheless, the RC file should hold values for all 4 options. If one of them is missing, vimpdb breaks and complains accordingly.
To communicate with the VIM instance where debugging happens, vimpdb needs to launch another VIM instance in client mode.
vim_client_script option holds the script used to launch that VIM instance
with clientserver support.
On Windows, it should hold vim.exe, not gvim.exe.
Furthermore, do not include quotes in the value to take care
of whitespace in the path.
In case no VIM instance is running, vimpdb launches a VIM instance in server mode.
vim_server_script option holds the script used to launch that VIM instance
with clientserver support. As debugging in the VIM instance is written with
python, that instance must have python support.
On MacOSX and Linux, vim_server_script and vim_client_script can hold
the same value.
On Windows, only the graphical VIM can be used as server, reason for the two separate default values as seen above.
The VIM instance in server mode has a name.
By default, vimpdb speaks to the server named VIM, which
is the default servername used by VIM.
If you want vimpdb to use another server name, modify the
server_name option. It should hold the name of the VIM
server you want to be used for debugging.
You may list the currently running VIM servers using:
$ vim --serverlist VIM
Or, on a Mac:
$ /Applications/MacVim.app/Contents/MacOS/Vim --serverlist VIM
When a VIM instance with clientserver support is running, you can find its
name by issuing the following command at the VIM prompt:
:echo v:servername
VIM communicates to vimpdb through a UDP socket. By default, the socket is opened on port 6666.
If that socket is not available in your system, you can specify an available
port number with the port option.
Using vimpdb is easy -- just insert a call to set_trace in your code
almost as usual:
import vimpdb; vimpdb.set_trace()
Then start your python application/script.
When the python interpreter hits that line, vimpdb will launch a VIM instance. VIM should get the focus; it loads the source file at the right line.
In VIM, you may now use the following commands:
| Key | Command |
|---|---|
| n | Next |
| s | Step |
| a | Args |
| u | Up |
| d | Down |
| r | Return |
| c | Continue |
| b | Sets a breakpoint at the line on which the cursor is sitting. |
| w | Displays the value of the word on which the cursor is sitting. |
| x | Switch to debugging in shell with standard pdb. |
| v | Switch back to vimpdb from plain pdb. |
If you find it hard to change habits and keep on typing
import pdb; pdb.set_trace()
you can add the following line to the .pdbrc file sitting in your home
folder:
import vimpdb; vimpdb.hookPdb()
This way, the command v(im) mentioned above is added to your standard
pdb and you can switch to vimpdb at any time.
- No convenient way to evaluate a Python expression.
Before version 0.4.1, vimpdb RC file (~/.vimpdbrc) had a single
script option. That option has been turned into the vim_client_script
option. The upgrade should be transparent.
Before version 0.4.0, vimpdb was configured through environment variables.
If you had a working configuration, upgrade should be transparent.
The values of VIMPDB_SERVERNAME and VIMPDB_VIMSCRIPT environment
variables are setup in the RC file (~/.vimpdbrc).
They are put respectively in server_name and script options.
See changelog