This reverts commit 72ef52d12e (#723).
The test failure on Windows will need to be debugged before this can
be re-merged.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
The username doesn't include the Matrix homeserver domain name, while
the mxid does. Since we change the usage of the old "username" to
include the domain, it is incorrect to remain saying it as "username."
Without universal_newlines=True or text=True, subprocess.check_output
returns bytes, not str, so it makes no sense to compare its return to
"true". But upstream Git’s behavior only depends on the filename, not
whether the repository is bare; emulate this more closely.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
We uses `pyupgrade --py3-plus` to automatically replace all occurence
of `Text`. But manual fix is required to remove the unused imports. Note
that with this configuration pyupgrade also convert string literals to
.format(...) style, which is manually not included in the commit as well.
The mirror has some chance of running on Python 3 now, once the
python-zephyr patch is rebased on 0.2.1, though it’s untested.
Signed-off-by: Anders Kaseorg <anders@zulip.com>
- Call `super()` in the init method.
- Change return hashlib.md5 syntax to encode properly for Python 3
Signed-off-by: Manu LN <manu+github@lacavernedemanu.fr>
Additionally, pin the irc library version to 18.0 because the newer
version has an error in the AioReactor:
```
File "/home/rht/code/venv/lib/python3.8/site-packages/irc/bot.py", line 108, in run
self.bot.reactor.scheduler.execute_after(intvl, self.check)
AttributeError: 'AioReactor' object has no attribute 'scheduler'
```
This reduces the number of retries that might spam APIs.
There is some complexity here which is left un-managed -- for
instance, maybe_restart_mirroring_script does a number of restart
attempts, and then fails, but will be retried every 15s by the
surrounding `process_loop`. Previously, it would merely have looped
forever inside maybe_restart_mirroring_script.
Three loops are intentionally left as infinite `while True` loops,
that merely cap their backoff at the default 90s. Their callers do
not expect, or have any way to handle more gracefully, a failure of
the expected-infinite-loop in `process_loop` or `zulip_to_zephyr`.
They maintain their previous behavior of retrying forever, albeit more
slowly.
