Make textual apps pipeable by default

[yggdr]

It is currently undocumented how textual behaves when you pipe into a program echo test | my_textual_program. There are a few issues and discussions here on GitHub that will tell you that textual polls STDIN regularly, so if that is not connected to a controlling TTY, textual will crash.
Also in these issues and discussions one can find various workarounds, but the best one seems to be what is done in the Textualize/toolong project (https://github.com/Textualize/toolong/blob/main/src/toolong/cli.py#L36), where the app detects that sys.stdin is a pipe, reopens the pipelines controlling TTY, and runs itself as a subprocess to which it attaches the reopend TTY as its stdin. Then it reads from its own stdin and writes everything that gets piped in into a tempfile that the subprocess then reads.

IMO this should be incorporated into textual proper in the form of a e.g. a helper function to get at least simple apps going that allow data being piped in.
But in the short term this should be at least documented so a search on textual.textualize.io for terms like stdin, pipe, or pipeline can find the info that this is currently only supported via a workaround.

[github-actions[bot]]

We found the following entry in the FAQ which you may find helpful:

• Does Textual support images?

Feel free to close this issue if you found an answer in the FAQ. Otherwise, please give us a little time to review.

This is an automated reply, generated by FAQtory

[yggdr]

So what is it that you think needs further discussion that should here instead of an enhancement issue? As I mentioned, there have already been a few places here where this has been discussed, and workarounds proposed. Is the very fact whether this should be added as a built-in feature contentious, or is it the precise form such a feature should take?
At least the point about mentioning it in the documentation should be uncontroversial, right?