Spack currently cannot run as a background process uninterrupted because some of the logging functions used in the install method (especially to create the dynamic verbosity toggle with the v key) cause the OS to issue a SIGTTOU to Spack when it's backgrounded.

This PR puts the necessary gatekeeping in place so that Spack doesn't do anything that will cause a signal to stop the process when operating as a background process.

With the addition of #13100 (parallel builds), people are going to try to
run Spack in the background more often.  This makes Spack handle that
situation gracefully, the way a good POSIX program should.

Specifically:

1. When `spack install` is running, we disable echo and canonical input
   so that users can type `v` to toggle build output.  We do that in a
   safe way now, so that it does not generate `SIGTTOU` in the background
   (#14682 did this).

2. We properly disable keyboard input mode when Spack is placed in the
   background, and re-enable it when Spack is in the foreground.  This
   means that if you Ctrl-Z a spack install, your terminal won't be left
   in a weird state.

3. We'll continue writing verbose output when Spack is in the background.
   If you have `stty +tostop` on, it'll end up stopping the build when
   you try to run in the background, unless you redirect output. This is
   normal behavior and it lets you easily do thingslike `spack install &>
   log.txt &`

4. Spack works fine when stopped in the background or when running in the
   background.

(2) is handled mostly with signal handlers (the way things like `vi` and
`emacs` do it) -- see the code for how that's done -- it's a bit tricky
in Python, as Python did not support blocking signals until 3.8.  It
turns out we can still make it work.

With the addition of #13100 (parallel builds), people are going to try to
run Spack in the background more often.  This makes Spack handle that
situation gracefully, the way a good POSIX program should.

Specifically:

1. When `spack install` is running, we disable echo and canonical input
   so that users can type `v` to toggle build output.  We do that in a
   safe way now, so that it does not generate `SIGTTOU` in the background
   (#14682 did this).

2. We properly disable keyboard input mode when Spack is placed in the
   background, and re-enable it when Spack is in the foreground.  This
   means that if you Ctrl-Z a spack install, your terminal won't be left
   in a weird state.

3. We'll continue writing verbose output when Spack is in the background.
   If you have `stty +tostop` on, it'll end up stopping the build when
   you try to run in the background, unless you redirect output. This is
   normal behavior and it lets you easily do things like this:

       spack install -v &> log.txt &

4. Spack works fine when stopped in the background or when running in the
   background.

(2) is handled mostly with signal handlers (the way things like `vi` and
`emacs` do it) -- see the code for how that's done -- it's a bit tricky
in Python, as Python did not support blocking signals until 3.8.  It
turns out we can still make it work.

With the addition of #13100 (parallel builds), people are going to try to
run Spack in the background more often.  This makes Spack handle that
situation gracefully, the way a good POSIX program should.

Specifically:

1. When `spack install` is running, we disable echo and canonical input
   so that users can type `v` to toggle build output.  We do that in a
   safe way now, so that it does not generate `SIGTTOU` in the background
   (#14682 did this).

2. We properly disable keyboard input mode when Spack is placed in the
   background, and re-enable it when Spack is in the foreground.  This
   means that if you Ctrl-Z a spack install, your terminal won't be left
   in a weird state.

3. We'll continue writing verbose output when Spack is in the background.
   If you have `stty +tostop` on, it'll end up stopping the build when
   you try to run in the background, unless you redirect output. This is
   normal behavior and it lets you easily do things like this:

       spack install -v &> log.txt &

4. Spack works fine when stopped in the background or when running in the
   background.

(2) is handled mostly with signal handlers (the way things like `vi` and
`emacs` do it) -- see the code for how that's done -- it's a bit tricky
in Python, as Python did not support blocking signals until 3.8.  It
turns out we can still make it work.

With the addition of #13100 (parallel builds), people are going to try to
run Spack in the background more often.  This makes Spack handle that
situation gracefully, the way a good POSIX program should.

Specifically:

1. When `spack install` is running, we disable echo and canonical input
   so that users can type `v` to toggle build output.  We do that in a
   safe way now, so that it does not generate `SIGTTOU` in the background
   (#14682 did this).

2. We properly disable keyboard input mode when Spack is placed in the
   background, and re-enable it when Spack is in the foreground.  This
   means that if you Ctrl-Z a spack install, your terminal won't be left
   in a weird state.

3. We'll continue writing verbose output when Spack is in the background.
   If you have `stty +tostop` on, it'll end up stopping the build when
   you try to run in the background, unless you redirect output. This is
   normal behavior and it lets you easily do things like this:

       spack install -v &> log.txt &

4. Spack works fine when stopped in the background or when running in the
   background.

(2) is handled mostly with signal handlers (the way things like `vi` and
`emacs` do it) -- see the code for how that's done -- it's a bit tricky
in Python, as Python did not support blocking signals until 3.8.  It
turns out we can still make it work.

With the addition of #13100 (parallel builds), people are going to try to
run Spack in the background more often.  This makes Spack handle that
situation gracefully, the way a good POSIX program should.

Specifically:

1. When `spack install` is running, we disable echo and canonical input
   so that users can type `v` to toggle build output.  We do that in a
   safe way now, so that it does not generate `SIGTTOU` in the background
   (#14682 did this).

2. We properly disable keyboard input mode when Spack is placed in the
   background, and re-enable it when Spack is in the foreground.  This
   means that if you Ctrl-Z a spack install, your terminal won't be left
   in a weird state.

3. We'll continue writing verbose output when Spack is in the background.
   If you have `stty +tostop` on, it'll end up stopping the build when
   you try to run in the background, unless you redirect output. This is
   normal behavior and it lets you easily do things like this:

       spack install -v &> log.txt &

4. Spack works fine when stopped in the background or when running in the
   background.

(2) is handled mostly with signal handlers (the way things like `vi` and
`emacs` do it) -- see the code for how that's done -- it's a bit tricky
in Python, as Python did not support blocking signals until 3.8.  It
turns out we can still make it work.

…upstream_master

* commit 'e2b1737a42c9c0c796671f9dd0c39f623e4c91c0': (1343 commits)
  update CHANGELOG.md for 0.14.1
  version bump: 0.14.1
  multiprocessing: allow Spack to run uninterrupted in background (spack#14682)
  Cray bugfix: TERM missing while reading default target (spack#15381)
  Upstreams: don't write metadata directory to upstream DB (spack#15526)
  Creating versions from urls doesn't modify class attributes (spack#15452)
  bugfix: fix install_missing_compilers option bug from v0.14.0 (spack#15416)
  bugfix: installer.py shouldn't be executable (spack#15386)
  Add function replace_prefix_nullterm for use on mach-o rpaths. (spack#15347)
  ArchSpec: fix semantics of satisfies when not concrete and strict is true (spack#15319)
  suite-sparse: fix installation for v5.X (spack#15326)
  testing:  increase installer coverage (spack#15237)
  bugfix: resolve undefined source_pkg_dir failure (spack#15339)
  Bugfix: resolve StopIteration message attribute failure (spack#15341)
  Recover coverage from subprocesses during unit tests (spack#15354)
  Correct pytest.raises matches to match (spack#15346)
  bugfix:  Add dependents when initializing spec from yaml (spack#15220)
  Uniquify suffixes added to module names (spack#14920)
  bugfix: ensure proper dependency handling for package-only installs (spack#15197)
  Fix for being able to 'spack load' packages that have been renamed. (spack#14348)
  ...

# Conflicts:
#	.travis.yml
#	lib/spack/spack/modules/common.py
#	var/spack/repos/builtin/packages/mofem-cephas/package.py
#	var/spack/repos/builtin/packages/mofem-fracture-module/package.py
#	var/spack/repos/builtin/packages/mofem-users-modules/package.py
#	var/spack/repos/builtin/packages/python/package.py