spacemit: shellfmt again after quoting fixes

[rpardini]

spacemit: shellfmt again after quoting fixes

• spacemit: revert shellfmt that messed up
  • revert spacemit.conf changes from 2f63a9c
• spacemit: apply quoting to fix more glaring problems
  • add a TODO for the original author of a "if $(...)" statement which I can't understand; it's still wrong
• spacemit: shellfmt again after quoting fixes
  • problematic if $(...) remains; just indent changed

[rpardini]

@The-going identified the mess introduced by shellfmt. I've reverted the shellfmt, fixed a few quoting issues, and re-ran it. Thanks @The-going for finding and reporting it

[The-going]

spacemit: shellfmt again after quoting fixes

shellfmt script\program is most likely an assistant for a person who can find typos and suggest something,
but he is very often mistaken.
I interpret everything that shellfmt writes\prompts like this:
"Hey, man, I think there's something wrong with this place. Pay attention to this."
But no more.

I'll check the changes in more detail tomorrow.

[The-going]

I understand your desire to make the script better.
But! But this is a system file!
And if we make any changes, we are obliged to test these changes on various
possible operating systems.

I do not know for sure in which environment this file will be executed.
With which environment variables and how a completely specific
operating system shell will handle the disclosure of double surrounding quotes.
This file was written by two people and checked each other after making changes.
Personally, I tested it in debian and openSUSE.

[The-going]

@rpardini
What are we going to do about it?

1. Exclude the system file from the purposes of shelfmt analysis.
2. Fix it and spend a lot more time on tests in various shells.

[rpardini]

Dunno. Leave it broken, then.

[The-going]

Dunno. Leave it broken, then.

Did I misunderstand?

[rpardini]

1. I broke a file that was according to you carefully hand-crafted, by running shellfmt against it.
2. you pointed out the breakage
3. I sent a new PR (this one) un-doing my shellfmt, fixing the more glaring/blatant problems found in the original, and running it again.
4. you claim that is not good enough and that shellfmt is not good enough and that changes should be tested in different shells (?) and different OS's (?). Despite my best efforts, I've no time for such, thus I closed this PR. A shame, really, since I broke something and wanted to fix it, but it seems it's impossible to reason.

Thus the closure, as I simply don't have the resources (time) to keep on discussing this. Terribly sorry.

[The-going]

Terribly sorry.

No problem!
I'll do the cleaning.
At least tell me which version of shellfmt I need to use for testing in a production environment.

[rpardini]

We've GHA workflows that auto-update shellfmt and shellcheck.
Please use
bash lib/tools/shellcheck.sh
and
bash lib/tools/shellfmt.sh
those use the exact versions and configurations that everyone is expected to use, and are also used by the CI jobs ran.