Nick Mills-Barrett

Also hit this issue, sometimes 3 copies of each VM. Resolved it by exporting them to `.utm` files and re-importing, so utm no longer manages the location of the VMs.

Finally got round to making this better: https://github.com/pyinfra-dev/pyinfra/commit/0b4572294a30c0ac72ad0c3710a38dd8fc5c3d18

@wowi42 do you mind giving me push access to the branch, I'd like to fix up the styling to get this merged!

> Much appreciated! It's my first PR for pyinfra and the lining has cost me some time, but eventually figured out how to do it properly Would love any comments/ideas/etc...

> The problematic part is in the fixup commit and it boils down to changing the GitConfig fact to enable the collection of the system level config. The issue is...

So the `local.include` function is already calling `host.deploy` under the hood ([link](https://github.com/pyinfra-dev/pyinfra/blob/aad6c3b84afc56f40c36b9e53dd277ce3c51b916/pyinfra/local.py#L36)) which already supports passing data, so this should be a pretty quick change. Once done the data is...

> Perhaps the error needs to include all candidates so it's a bit more obvious what's going on? Happy to do a PR if it's welcome! Definitely this! PR would...

Switched to using the user input value only (since we only add one more option ever): https://github.com/pyinfra-dev/pyinfra/commit/1ef3741359adf79d990d06adf7e7f31d25243eb5

Hi @JakkuSakura - I assume you mean the `FindFiles` or similar facts? Extending the arguments taken should be pretty simple and allow specifying any `find` flags, source: https://github.com/pyinfra-dev/pyinfra/blob/aad6c3b84afc56f40c36b9e53dd277ce3c51b916/pyinfra/facts/files.py#L325-L330

This is unfortunate. A lot of work went into the `MaskString` and various string command classes, unfortunately fact arguments are bypassing that entirely.