id,node_id,number,title,user,state,locked,assignee,milestone,comments,created_at,updated_at,closed_at,author_association,pull_request,body,repo,type,active_lock_reason,performed_via_github_app,reactions,draft,state_reason
794554881,MDU6SXNzdWU3OTQ1NTQ4ODE=,1208,A lot of open(file) functions are used without a context manager thus producing ResourceWarning: unclosed file <_io.TextIOWrapper,4488943,closed,0,,,2,2021-01-26T20:56:28Z,2021-03-11T16:15:49Z,2021-03-11T16:15:49Z,CONTRIBUTOR,,"Your code is full of open files that are never closed, especially when you deal with reading/writing json/yaml files.

If you run python with warnings enabled this problem becomes evident.
This probably contributes to some memory leaks in long running datasettes if the GC will not 'collect' those resources properly.

This is easily fixed by using a context manager instead of just using open:
```python
with open('some_file', 'w') as opened_file:
    opened_file.write('string')
```

In some newer parts of the code you use Path objects 'read_text' and 'write_text' functions which close the file properly and are prefered in some cases.


If you want I can create a PR for all places i found this pattern in.


Bellow is a fraction of places where i found a ResourceWarning:
```python

update-docs-help.py:
  20          actual = actual.replace(""Usage: cli "", ""Usage: datasette "")
  21:         open(docs_path / filename, ""w"").write(actual)
  22  

datasette\app.py:
  210          ):
  211:             inspect_data = json.load((config_dir / ""inspect-data.json"").open())
  212              if immutables is None:

  266          if config_dir and (config_dir / ""settings.json"").exists() and not config:
  267:             config = json.load((config_dir / ""settings.json"").open())
  268          self._settings = dict(DEFAULT_SETTINGS, **(config or {}))

  445              self._app_css_hash = hashlib.sha1(
  446:                 open(os.path.join(str(app_root), ""datasette/static/app.css""))
  447                  .read()

datasette\cli.py:
  130      else:
  131:         out = open(inspect_file, ""w"")
  132      loop = asyncio.get_event_loop()

  459      if inspect_file:
  460:         inspect_data = json.load(open(inspect_file))
  461  

```

",107914493,issue,,,"{""url"": ""https://api.github.com/repos/simonw/datasette/issues/1208/reactions"", ""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,completed
793881756,MDU6SXNzdWU3OTM4ODE3NTY=,1207,"Document the Datasette(..., pdb=True) testing pattern",9599,closed,0,,,1,2021-01-26T02:48:10Z,2021-01-29T02:37:19Z,2021-01-29T02:12:34Z,OWNER,,"If you're writing tests for a Datasette plugin and you get a 500 error from inside Datasette, you can cause Datasette to open a PDB session within the application server code by doing this:

```python
ds = Datasette([db_path], pdb=True)
response = await ds.client.get(""/"")
```

You'll need to run `pytest -s` to interact with the debugger, otherwise you'll get an error.",107914493,issue,,,"{""url"": ""https://api.github.com/repos/simonw/datasette/issues/1207/reactions"", ""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,completed