Reuben

@edgarrmondragon What do you think about using `--` (double-dash) over `-` as the special "any" plugin type argument? `--` is used in `.lock` file names as a delimiter between the...

> @ReubenFrankel Just for posterity, I'll note that with *nix CLI applications `--` on its own commonly indicates that no more CLI options (of the form e.g. `-n`/`--name`) will be...

@edgarrmondragon Yeah, I would appreciate some help if you have any ideas. Might unblock https://github.com/meltano/edk/pull/75#issuecomment-1485849140 too if there is good pattern to follow.

> > @edgarrmondragon Yeah, I would appreciate some help if you have any ideas. Might unblock [meltano/edk#75 (comment)](https://github.com/meltano/edk/pull/75#issuecomment-1485849140) too if there is good pattern to follow. > > @ReubenFrankel So,...

@edgarrmondragon Not a hard blocker, no - I just haven't got an environment to manually test the fix with. I was thinking since CI is already setup with a Windows...

I think I finally made some progress on this last night - I'll push a test POC later today and we can decide whether it's worth having or not.

Hmm... Working locally no problem, but looks like there is some kind of race condition in reading the emulated tap process output for `test_validate_failure_subprocess_err` specifically. Failing tests on Windows indicate...

The [return code debug log](https://github.com/meltano/meltano/actions/runs/8903623347/job/24451671229#step:11:1670) seems to indicate that the process return code is `1` on Windows for both error and termination, which goes against what I thought was happening.

Decided to just flag if we encounter a `RECORD` message and use that to determine success, rather than rely on the process return code. I think this makes more sense...

Also loosely related to #7066