[WIP] [CI] new command implementation for tests

[apostasie]

Note this is a large PR, though the bulk of the changes are under a couple of files in mod/tigron/internal.

This PR is a tigron changeset addressing a few important issues:

• CI: icmd is not usable to test when binary expect a console #3883
• [CI]: fix commands being stuck problem #3996
• unlogged issue: dangling processes left behind

In a shell: icmd does not provide value for us. Most of its features we do not use, and the simple wrapping around golang command exec is causing more issues than it solves - icmd does replace the process standard streams with its own buffers, which gets in the way of using ptys, does not provide proper timeout logic for stuck processes, and does not seem to kill process groups properly.

This PR does away with it, and replace it with an internal com.Command that does offer:

• proper timeout
• pty handling
• proper process group termination

... along with absorbing a large part of the logic that was previously baked into test/command.go (env filtering, cloning, command wrapping).

While this should technically be (mostly) transparent for tests, some of the logic changes with timeout and pty / stdin handling required adaptation of some functions signatures.

Specifically:

• Background(timeout) is now simply Background(), as the new WithTimeout helper now applies to any command, and not just the background-ed ones
• WithPTY no longer supports the extra "writers" argument. To write to stdin, the same methods can now be used whether or not we are working with a PTY
• the previous WithStdin method was ill-named/ill-advised, as it is not setting stdin, but writing to stdin, can be used repeatedly, and lacked flexibility. It has been replaced by Feed(io.Reader) for the simpler cases, and WithFeeder(func()io.Reader) that allows for more complex logic before writing to stdin (like Pty(writers) used to)

Finally, this PR also provides stuff that came up as necessary while working on the new command:
- minimal improvements over debugging information output readability
- (cosmetic) changes to prepare for version 2 of golangci
- the testing of internal Command itself did motivate the addition of leak detection and an internal assert library

While on the surface this PR is minimally impactful, this is actually a profound change, and I assume a few rounds will be needed before taking it out of "draft". Specifically, the current default timeout for commands execution is set at 3 minutes, and we might want to adjust that (eg: shorten it generally, and set it explicitly for tests that require longer)

Blocked by:

• [CI] Tigron breakout: golangci v2 and cleanups #4041
• [CI] Tigron breakout 2: remove gotest.tools assert dependency internally #4042
• [CI] Tigron breakout 3: leak detection #4044
• [CI] Tigron breakout 4: adding a few tests  #4045
• [CI] Tigron breakout 5: enable lint and tests on the CI #4047
• [fix] Avoid console SetRaw #4054

Preferable to merge after:

• [CI] Update golangci to v2 #4048
• [CI] golangci v2: review, fix, document, rules #4051

TODO:

• green CI
• mute the logging from pipes by default

[apostasie]

@AkihiroSuda thanks for the early review on this.

I am breaking it down in smaller chunks to simplify the discussion, so, let see what remains here once we are done with the more trivial stuff in the breakout PRs.

[apostasie]

Note: in some cases, go exec may return -1 exit code, which would be colliding with these in case someone wanted to test that specific condition.

[apostasie]

Geeeeeeeez github.

@AkihiroSuda moving to #4056

Sorry about that.

Done:

• yanked out the logger
• CI is (should be) green
• cleaned-up the commits