pullrequests: introduce limit to stop displaying additional changes
We've noticed some scalability issues when many descendants exist for the
changesets in a pull request.
To resolve this issue, we instead do not display any additional changes at all
if the amount of additional changes is larger than a configured limit.
(This occurred because we were merging a lot of heads in the repository.)
Throwing away some of the changesets would also keep the total amount more
manageable, but this would result in an (for the user) unpredictable subset of
changesets being shown.
This could be extended further with "too long to be shown - click here to
show", but that was quite a bit of additional work and did not cover our use
case of not allowing the display at all in case of too many additional changes.
repo group: fix logging of rename/move
After renaming a group, it would iterate over all the contained groups and
repos and update their full path while logging the update from the/old/path to
the the/new/path. Doing that, it would also visit the already renamed top level
group, but since the full path of that one already had been updated, it would
log it as renaming from the/new/path to the/new/path.
Fixed by logging when renaming in the first place, and skipping the top level
repo group while iterating.
To avoid redundant logging, only log (and rename) if the name or parent
actually change.
Based on a patch by toras9000.
setup: use old importlib_metadata version to fix kombu failing on python < 3.8
Many libraries use the importlib_metadata library as fallback when running on
Python versions older than 3.8 . For example setuptools when easy_install is
used for install the Kallithea console_scripts entrypoints in the bin folder.
The dependencies on importlib_metadata were indirect and without constrains on
version number.
The problem is that Celery uses Kombu, which (on Python < 3.8) uses
importlib_metadata in a way that is incompatible with importlib_metadata > 5.
Most obvious, building docs failed as:
Running Sphinx v5.1.1
Configuration error:
There is a programmable error in your configuration file:
Traceback (most recent call last):
File ".../kallithea/venv/lib64/python3.7/site-packages/sphinx/config.py", line 347, in eval_config_file
exec(code, namespace)
File ".../kallithea/docs/conf.py", line 17, in <module>
import kallithea
File ".../kallithea/kallithea/__init__.py", line 45, in <module>
CELERY_APP = celery.Celery() # needed at import time but is lazy and can be configured later
File ".../kallithea/venv/lib64/python3.7/site-packages/celery/local.py", line 492, in __getattr__
[name])
File ".../kallithea/venv/lib64/python3.7/site-packages/celery/app/__init__.py", line 2, in <module>
from celery import _state
File ".../kallithea/venv/lib64/python3.7/site-packages/celery/_state.py", line 15, in <module>
from celery.utils.threads import LocalStack
File ".../kallithea/venv/lib64/python3.7/site-packages/celery/utils/__init__.py", line 16, in <module>
from .nodenames import nodename, nodesplit, worker_direct
File ".../kallithea/venv/lib64/python3.7/site-packages/celery/utils/nodenames.py", line 6, in <module>
from kombu.entity import Exchange, Queue
File ".../kallithea/venv/lib64/python3.7/site-packages/kombu/entity.py", line 7, in <module>
from .serialization import prepare_accept_content
File ".../kallithea/venv/lib64/python3.7/site-packages/kombu/serialization.py", line 440, in <module>
for ep, args in entrypoints('kombu.serializers'): # pragma: no cover
File ".../kallithea/venv/lib64/python3.7/site-packages/kombu/utils/compat.py", line 82, in entrypoints
for ep in importlib_metadata.entry_points().get(namespace, [])
AttributeError: 'EntryPoints' object has no attribute 'get'
That made readthedocs builds fail, when it in the default web configuration
used Python 3.7 .
Fixed by introducing an explicit dependency on importlib_metadata < 5.