"editor" command not working properly with spaces within an argument

[IcaroBritto]

I'm having a problem with the "editor" entry on the .jrnl_config file.
Basically what I am trying to do is start vim with the filetype defined to markdown.
The editor entry is like this:

{
...
   "editor": "nvim -c 'setlocal filetype=markdown'",
...
}

This command works fine on terminal, but this is not working on jrnl.
The problems are:

• The filetype is not set (it is empty);
• An error is shown when nvim opens like if the ' is being interpreted as a marker command.
• The jrnl temp file is not set in nvim.

I've tried double quotes ("editor": "nvim -c \"setlocal filetype=markdown\"") and escaping white spaces ("editor": "nvim -c setlocal\\ filetype=markdown") but they didn't work either.

Can you help me with that?
Sorry for my English.

Nice piece of software by the way.

[maebert]

Internally, the commands to launch an editor are split on whitespaces and stored in an array, so "nvim -c 'setlocal filetype=markdown'" becomes ["nvim", "-c", "'setlocal", "filetype=markdown'"], causing the ' to be misinterpreted. There's no super elegant way to fix that unfortunately, but you try to could store 'setlocal filetype=markdown' in an environment variable and change it to something like "nvim -c $SETMD"

[IcaroBritto]

Hi,
Thank you for the reply.
I did just like you said, but it didn't work :/
I think the environment variables aro not being passed when the editor command is executed.
I'm not a Python developer, but I was looking into the code and, correct if I'm wrong, looks like that the process is called at the line 134 of the util.py.
Maybe if the code was like this we could use environment variables:

...
env_copy = os.environ
subprocess.call(config['editor'].split() + [tmpfile], env=env_copy)
...

Again, I'm not a Python developer, so I'm totally not sure.
Thanks again!

[wren]

I've had similar issues. There's a PR that aims to address this bug, but I haven't gotten a chance to test it out.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[gregorybodnar]

Let's keep this open. I might be able to test the PR with some contrived invocations. I'll try for the weekend.

[gregorybodnar]

I built up a test with #435 cherry-picked onto 2.0-rc3 and using this in my config:

  editor_test:
    editor: vim -c 'r! date' -f
    journal: /home/greg/editortest.journal

The PR preserves the quotes that are being passed, so it looks like this issue would be resolved. I don't have a Windows system to try with, so I'm just going to have to trust the Python docs for that case.

[wren]

Yup, I agree that the PR looks pretty good. I want to give the original author a chance to update themselves. If they don't, though, we can look at our options. I'm not sure we can just cherry-pick their commit when we factor in licensing and stuff like that, but I'm sure we can make a PR with similar functionality (and more tests, anyway). Anyway, we can cross that bridge when we get to it.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[wren]

I think subprocess is the right way to go, but passing shell=True is just too dangerous of an approach when the config file isn't secured. We'd basically just be executing arbitrary commands on a user's machine. Which we sort of already are, but that feels much worse.

Other than that, though. The oringinal PR looks good.

[wren]

@gregorybodnar Looks like that PR staled out. Do you still have a chance to submit an updated PR?

[gregorybodnar]

I have a long weekend just about to start, so I'll get it going.

[wren]

So, here's a thing.

On a mac:

$ python -c 'import sys; print(sys.platform)'
darwin

And this is our editor code:
subprocess.call(shlex.split(config['editor'], posix="win" not in sys.platform) + [tmpfile])

posix="win" not in sys.platform

"win"

darwin

😂

[wren]

Fixing the above (change to win32) fixes all issues related to editor config values for me. I'm drafting a PR with this fix plus a few tests.