Out of date: This is not the most recent version of this page. Please see the most recent version

Ignoring files from mbed build

The .mbedignore file tells the mbed build command which files and directories to ignore (not process).


You can place the .mbedignore file in any directory where mbed build is going to search for source files. The most convenient place is the root directory of the library or application. However, this is not a requirement.

Avoid defining rules that would cross the library boundaries; these can lead to side effects or build problems that are hard to find.


Each line in the .mbedignore file is a file pattern used for matching files. Each matched file or directory is ignored while building.

The following wildcards are accepted:

Pattern Meaning
* Matches everything.
? Matches any single character.
[seq] Matches any character in seq.
[!seq] Matches any character not in seq.

The file is parsed with Python’s fnmatch functionality, so the syntax follows basic shell patterns with the following exceptions:

  1. Each line is internally prefixed with the path of the .mbedignore file.
  2. A line cannot start with . or / (because of rule 1).

The globbing functionality is not used, so you cannot recursively match a specific file pattern. Instead, you need to define a rule per directory.

You can use relative paths, so you can match files deeper in the build tree. However, avoid crossing library boundaries.


A file located in source/obsolete/.mbedignore with the following content:


After applying rule 1, the actual patterns used internally for matching the source files are: