Use alternate screen buffer unconditionally

[heaths]

Fixes regression from #6356. Previous assumption before #5681 was that VT cursor movement and screen clearing was supported, so this effectively restores that assumption while remaining separate from color support and virtual terminal processing support.

[heaths]

@mislav @samcoe this effectively restore the old assumption before alternate screen buffers that VT100 cursor movements and screen clearing was supported, while remaining separate from color support: https://github.com/cli/cli/pull/5681/files#diff-4ca7c158ba1609a9968f3f5bd80a5138fd242ed3b07882e12774ff0ac3188fcfL215-L223

At the very least, it's no worse than before.

A longer-term solution might be to, on Windows, check if virtual terminal processing is enabled, and on other platforms:

1. For 256- or truecolor detection, you can write a sequence like '\e[48:2:1:2:3m\eP$qm\e\\' (for truecolor) to the console and read it back to see if the color matches: https://github.com/termstandard/colors#querying-the-terminal
2. For cursor movement, alternate screen buffers, etc., if you wanted to lump them together (should be fairly safe) you could try moving the cursor and querying for the current position, then reset it.

All this would, of course, only be done if stdout is TTY.

[mislav]

@heaths Good catch and thank you! But, I was already starting to address this in https://github.com/cli/cli/pull/6421/files#diff-f27e472205191c606d87c322264c60b4e47e333d45866f76f5a822e35b22a3baR8