feature request: allow use of .mailmap #764
Labels
Comments
|
As this library is in |
|
Thanks for the feedback @Byron. The easiest workaround seems to be to read in the |
HelgeCPH
added a commit
to HelgeCPH/pydriller
that referenced
this issue
Oct 11, 2024
PyDriller does not support [Git `.mailmap` files](https://git-scm.com/docs/gitmailmap). That is, authors or committers, which are represented as `Developer` objects always get assigned the name and email values that are stated in respective commits. I would like that PyDriller supported use of `.mailmap` files directly. So far, I have worked around lacking mailmap support by letting my external scripts map users with multiple email addresses or slightly different names to canonical data. I believe there might be more users than me who could be interested in this feature. Currently, PyDriller does not support `.mailmap` files since it relies on GitPython to receive names and emails for `Developer`s. GitPython does not support `.mailmap` files either, see [here](gitpython-developers/GitPython#764). Since they state that GitPython is in maintenance mode, I contribute support of `.mailmap` files to PyDriller and not to the underlying dependency. I believe that this functionality is also better suited to PyDriller, a library with which users want to run analysis of Git repositories, where analysis allow for "deduplication" of physical authors into logical authors. Since parsing of `.mailmap` files is not straight forward and since I do not want to introduce an algorithm that might produce results dissimilar to Git, I decided to wrap the `git check-mailmap` [CLI command](https://git-scm.com/docs/git-check-mailmap/2.31.0). Most of this is done in the new file [pydriller/utils/mailmap.py](utils/mailmap.py). To make the feature work, I had to extend the [`Developer`](pydriller/domain/developer.py) class slightly.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
A
.mailmapfile is used to map commit author info where committers use multiple names or email addresses to a single unique name/email. It is usually located in the root of a repo. See e.g. https://stackoverflow.com/questions/13777171/configuring-git-log-to-use-mailmap-by-defaultI cannot find any mention of the feature in this repo;
GitPyhonseems to ignore the file. It would be quite useful to allow using that file. Maybe replacing the author name and email inraw, or adding a secondrow-mailmappedcolumn for example.The text was updated successfully, but these errors were encountered: