project.repository_tree returns 404 for non-existent path (used to return an empty list)

[elboulangero]

Description of the problem, including code/CLI snippet

project.repository_tree(path="xxx") throws a 404 exception if the path xxx doesn't exist. Before it didn't and just returned an empty list. It seems that the behavior changed during last week.

Working example:

import gitlab
from pprint import pprint

# Configuration
SERVER_URL = "https://gitlab.com"
GROUP_ID = 5054009
PROJECT_NAME = "kali-docs"

# Initialization
GL = gitlab.Gitlab(SERVER_URL)
group = GL.groups.get(GROUP_ID)
projects = group.projects.list(all=True)

# Select a project to work with
gproj = [p for p in projects if p.name == PROJECT_NAME][0]
pprint(gproj.attributes)

# Get a "manageable project"
proj = GL.projects.get(gproj.id)

# Get repo tree for a non-existent path
tree = proj.repository_tree(path="non-existent")

It would be nice if someone can confirm this change of behavior.

Expected Behavior

proj.repository_tree(path="non-existent") used to return an empty list, there was no need to catch any exception.

Actual Behavior

>>> tree = proj.repository_tree(path="non-existent")
Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/gitlab/exceptions.py", line 337, in wrapped_f
    return f(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/gitlab/v4/objects/repositories.py", line 80, in repository_tree
    return self.manager.gitlab.http_list(gl_path, query_data=query_data, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/gitlab/client.py", line 944, in http_list
    gl_list = GitlabList(self, url, query_data, get_next=False, **kwargs)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/gitlab/client.py", line 1146, in __init__
    self._query(url, query_data, **self._kwargs)
  File "/usr/lib/python3/dist-packages/gitlab/client.py", line 1156, in _query
    result = self._gl.http_request("get", url, query_data=query_data, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/gitlab/client.py", line 800, in http_request
    raise gitlab.exceptions.GitlabHttpError(
gitlab.exceptions.GitlabHttpError: 404: 404 invalid revision or path Not Found

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python3/dist-packages/gitlab/cli.py", line 71, in wrapped_f
    return f(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/gitlab/exceptions.py", line 339, in wrapped_f
    raise error(e.error_message, e.response_code, e.response_body) from e
gitlab.exceptions.GitlabGetError: 404: 404 invalid revision or path Not Found

Specifications

• python-gitlab version: tested 2.5.0-1 (Debian 11 bullseye) and 3.12.0-1 (Debian unstable)
• API version you are using (v3/v4): v4
• Gitlab server version (or gitlab.com): gitlab.com

[JohnVillalovos]

This sounds like a GitLab issue. python-gitlab is just a thin wrapper around the API. So if before GitLab returned an empty list, but now it returns a 404, then that is what python-gitlab will do.

[nejch]

I agree with @JohnVillalovos here - @elboulangero, the GitLab API often has subtle breaking changes like this one even in minor releases and across different endpoints, and for us to try and handle individual cases like this, the library would become a whole mess and it gets a bit hard to track.

I suggest you handle the error in your user code if this is a recent change in GitLab behavior 🙇 but if there's anything else you're having issues with feel free to reopen!

[elboulangero]

It's easy to catch and handle the exception in user code, no problem. Thanks for the quick and informative answer!