Getting Started with coala¶
Welcome to this little tutorial. It is meant to be a gentle introduction to the usage of coala.
In order to complete this tutorial you will need coala installed. Installation instructions can be found here.
Here’s a list of our supported languages.
Get Some Code¶
In order to perform a static code analysis on your code you will need some code to check. If you do not have your own code you want to check, you can retrieve our tutorial samples:
git clone https://github.com/coala/coala-tutorial
Please note that the commands given in this tutorial are intended for use with this sample code and may need minor adjustments.
There are two options how to let coala know what kind of analysis it should perform on which code.
Command Line Interface¶
In order to specify the files to analyze, you can use the
argument of coala like demonstrated below. For all file paths, you can
specify (recursive) globs.
Because analysis routines can do many various things we named them bears. A bear can check your code for potential problems, calculate metrics and even provide corrections for your code.
You can specify the bears that you want coala to run using the
cd coala-tutorial coala --files=src/\*.c --bears=SpaceConsistencyBear --save
You can use comma separated values to specify more than one item in
arguments! Do not use spaces as that would start a new argument.
coala will now ask you for missing values that are needed to perform the
analysis, which in this case is only the
use_spaces setting. We
recommend setting it to
coala will now check the code and, in case you use the tutorial code,
yield one result. SpaceConsistencyBear will detect a trailing whitespace at
the end of the line, after
#include <stdio.h> in the
main.c file. coala
will then ask you to remove the trailing space, by applying the suggested
patch (option 2).
You can also run coala in non interactive mode (given that all the settings
required by the bears you are using are provided in the
In this case there won’t be any interaction, the patch will be shown directly.
Feel free to experiment a bit. You’ve successfully analysed some code!
But don’t stop reading - you don’t have to enter all those values again!
We have given coala the
--save argument, which means that it will
automatically generate a
.coafile into the current directory. Read on!
Configuration Files - coafiles¶
coala supports a very simple configuration file. If you’ve executed the instructions from the CLI section above, coala will already have such a file readily prepared for you. Go, take a look at it:
If you are using Windows, you should use
type .coafile instead!
This should yield something like this:
[Default] bears = SpaceConsistencyBear files = src/*.c use_spaces = yeah
If you now invoke
coala it will parse this
.coafile from your
current directory. This makes it easy to specify once for your project
what is checked with which bears and make it available to all
Feel free to play around with this file. You can either edit it manually
or add/edit settings via
coala --save ... invocations. If you want
coala to save settings every time, you can add
save = True manually
Thats all nice and well but we also have a Makefile for our project we want to check. So let us introduce another feature of our configuration syntax: sections.
[Default] specifies that everything below will belong to
the Default section. If nothing is specified, a setting will implicitly
belong to this section.
Let’s check the line lengths of our Makefile:
coala -S Makefiles.bears=LineLengthBear Makefiles.files=Makefile --save
As you can see, the
--settings) option allows to specify
arbitrary settings. Settings can be directly stored into a section with
By default, the
LineLengthBear checks whether each line contains
79 chars or less in a line. To change this value, use the
max_line_length inside the
coala will now yield any result you didn’t correct last time, plus a new
one for the Makefile. This time coala (or better, the
LineLengthBear) doesn’t know how to fix the issue but still tries to
provide as much helpful information as possible and provides you the
option to directly open the file in an editor of your choice.
If your editor is already open this may not work, because the other process will shortly communicate with the existent process and return immediately. coala handles this for some editors automatically, if yours does not work yet - please file an issue so we can include it!
If you changed one file in multiple results, coala will merge the changes if this is possible.
coala should have appended something like this to your
[Makefiles] bears = LineLengthBear files = Makefile
As you see, sections provide a way to have different configurations for possibly different languages in one file. They are executed sequentially.
For a list of configuration options for the bears, take a look at our bear-docs documentation.
Often you don’t want to look at trivial results like spacing issues. For
that purpose coala includes a special setting called
that allows you to set the action for a bear that shall be automatically
applied on run. We have a command line alias
--apply-patches to make it
easier to use.
Let’s automatically fix Python code. Take a look at our sample Python code:
$ cat src/add.py """ This is a simple library that provides a function that can add numbers. Cheers! """ def add(a,b): return a+b; import sys
That looks horrible, doesn’t it? Let’s fix it!
$ coala -S python.bears=PEP8Bear python.files=\*\*/\*.py \ --apply-patches --save # other output ... Executing section python... [INFO][11:03:37] Applied 'ApplyPatchAction' for 'PEP8Bear'. [INFO][11:03:37] Applied 'ApplyPatchAction' for 'PEP8Bear'.
coala would now fix all spacing issues and without bothering you again.
All settings in the default section are implicitly inherited to all other sections (if they do not override their values). We can use that to save a few lines!
Lets add the following section to our
[TODOS] bears = KeywordBear
coala with the
-s argument which is the same as
--save. I recommend setting case insensitive keywords to
TODO, FIXME and case sensitive keywords empty.
After the results we’ve already seen, we’ll see a new informational one which informs us that we have a TODO in our code.
Did you note that we didn’t specify which files to check this time? This
is because all settings, including
files = src/*.c, from the Default
section are already available in every other section implicitly. Thus
the default section is a good point to set things like logging and
output settings or specifying a default set of files to check.
There are several ways to ignore certain issues, so you aren’t lost if any routines yield false positives.
coala lets you ignore whole files through the
ignore setting. In
addition to normal globs, coala offers
** to match all directories and
files = **/*.h ignore = **/resources.h
This configuration would include all header (
.h) files but leaves
out resource headers.
Ignoring Code Inside Files¶
Sometimes you need finer-graded ignores. Imagine you have a
LineLengthBear that shall not run on some code segments, because you
can’t wrap them:
code = "that's checked normally" # Ignore LineLengthBear unwrappable_string = "some string that is long and would exceed the limit"
You can also skip an area:
# Start ignoring LineLengthBear unwrappable_string_2 = unwrappable_string + "yeah it goes even further..." another_unwrappable_string = unwrappable_string + unwrappable_string_2 # Stop ignoring
You can also conditionally combine ignore rules! Bear names will be
split by comma and spaces, invalid bear names like
and will be
Also note that in the bear names delimited by commas and spaces, you may specify glob wildcards that match several bears:
# Start ignoring Line*, Py* unwrappable_string_2 = unwrappable_string + "yeah it goes even further..." another_unwrappable_string = unwrappable_string + unwrappable_string_2 # Stop ignoring
In the above example all bears matching the glob Line* and Py* will be ignored. You may also specify more complex globs here such as # Start ignoring (Line*|P[yx]*) which will ignore all bears’ names which start with Line, Py, and Px.
# Ignore LineLengthBear and SpaceConsistencyBear variable = "Why the heck are spaces used instead of tabs..." + "so_long"
If you put an
all instead of the bear names directly after the
ignoring keyword, the results of all bears affecting
those lines will be ignored.
If you’ve used another linter in the past, you don’t have to change your
pre-existing code with the
noqa keywords to
ignore as the examples
below work as well. If no bears are specified,
noqa will be applicable to
work for all bears.
# noqa long_line = "This is a long line ... "
If you wish to specify which bear to use with
noqa, as is done
ignore, you would have to proceed as follows:
# noqa LineLengthBear long_line = "This is a long line ... "
Now that we have sections we need some way to control, which sections are executed. coala provides two ways to do that:
If you add the line
TODOS.enabled=False to some arbitrary place to
.coafile or just
enabled=False into the
coala will not show the TODOs on every run.
Especially for those bears yielding informational messages which you might want to see from time to time this is a good way to silence them.
If you provide positional arguments, like
coala Makefiles, coala
will execute exclusively those sections that are specified. This will
not get stored in your
.coafile and will take precedence over all
enabled settings. You can specify several targets separated by a space.
What was that TODO again?
Continuing the Journey¶
If you want to know about more options, take a look at our coala settings documentation or with
coala -h. If you liked or disliked this tutorial, feel free to drop
us a note at our bug tracker or mailing list.
If you need more flexibility, know that coala is extensible in many ways due to its modular design:
- If you want to write your own bears, take a look at our tutorial.
- If you want to add custom actions for results, take a look at the code in coalib/results/results_actions.
- If you want to have some custom outputs (e.g. HTML pages, a GUI or voice interaction) take a look at modules lying in coalib/output.