fix(shell): quote variables for SC2086/SC2248/SC2046

[kojiromike]

Short description of what this resolves:

Fix shellcheck SC2086, SC2248, and SC2046 (variable-quoting) findings across the shell files checked in CI. This is the first of several focused PRs chipping away at the full shellcheck backlog — remaining codes (SC2044 find-loops, SC2154 unset vars, SC2310/SC2312 set -e behavior, etc.) are queued for follow-up PRs.

After this PR, shellcheck --include=SC2086,SC2248,SC2046 is clean on every file the CI workflow scans.

Changes proposed in this pull request:

• Quote variables in docker/openemr/{7.0.4,8.0.0}/{openemr,ssl}.sh, the fsupgrade-*.sh scripts, packages/standard/{ami,scripts}/*.sh, and the three utilities/ installer/migrator/monitor scripts.
• Where word-splitting is intentional, add targeted per-line # shellcheck disable=SC2086 with a reason comment:
  • openemr.sh: CONFIGURATION intentionally splits into multiple -f flags (see fix(docker): auto_configure.php inconsistency between flex and other versions #539).
  • ssl.sh: EMAIL intentionally splits to pass -m address as separate args.
• In fsupgrade-*.sh scripts, remove redundant if [ ! -d X ]; then mkdir -p X; fi guards (mkdir -p is already idempotent) and collapse the 13 sequential mkdir -p calls into a single multi-arg mkdir -p.
• Re-sync docker/openemr/8.0.0/upgrade/fsupgrade-{1..7}.sh from the updated 7.0.4/ copies so the CI content-hash dedup still collapses byte-identical files.

CI's shellcheck workflow remains failing — those failures are all from SC codes queued for follow-up PRs, not from any fix in this change.

AI disclosure:

• Yes, I used AI to help with this PR

[jesdynf]

@bradymiller I don't think this is exceptional and I've spot-checked it (and tested things on the commandline until I got what it was trying to get across) but it touches every script we own. Is there anything you want to peek at it before we let it through?

[bradymiller]

Will this require this change (both for 7.0.3 and 7.0.4) if not already there:

openemr-devops/docker/openemr/flex-edge/auto_configure.php

Lines 27 to 48 in 8f3eca7

	// Collect parameters(if exist) for installation configuration settings
	for ($i=1; $i < count($argv); $i++) {
	if ($argv[$i] == '-f' && isset($argv[$i+1])) {
	// Handle case where a single string contains all parameters
	$configString = $argv[$i+1];
	$configPairs = preg_split('/\s+/', trim($configString));
	foreach ($configPairs as $pair) {
	if (strpos($pair, '=') !== false) {
	list($index, $value) = explode('=', $pair, 2);
	$installSettings[$index] = $value;
	}
	}
	// Skip the next argument since we already processed it
	$i++;
	} else {
	// Handle standard key=value parameters
	$indexandvalue = explode("=", $argv[$i]);
	$index = $indexandvalue[0];
	$value = $indexandvalue[1] ?? '';
	$installSettings[$index] = $value;
	}
	}

[kojiromike]

bah, there was a bug in my git ls-files and this didn't actually catch all the shellcheck issues. There are still a ton in the xbackup scripts.

[jesdynf]

Those might not be worth fighting. I need to migrate away from the current MySQL image with Percona's XtraBackup stuff to something with MariaDB.

[jesdynf]

@kojiromike I've migrated off the xbackup stuff, which was blocking you, but this pull is aging -- should we close or can we move on it?

[kojiromike]

I'll move on it soon.

[jesdynf]

Looks okay, but a lot of the constructs aren't ones I use regularly. Which, I suppose, is the point of the commit.

[bradymiller]

note there is a plan for 7.0.5:
#534
(ie. don't remove it. I'll move it over to 8.0.1 soon)

[kojiromike]

note there is a plan for 7.0.5: #534 (ie. don't remove it. I'll move it over to 8.0.1 soon)

ack, I've moved this back to draft until I can manage 7.0.5.

I'm a bit stuck on this change because:

1. I don't want to do one version at a time and have divergence between versions.
2. If I try to do all versions at once then the PR is really big, and differences between the versions makes it harder.

[bradymiller]

7.0.5 is Jake's new docker updates (needs more testing, which why becoming 8.0.1 instead of 8.0.0). I think all his work passed shellcheck so hopefully won't require much (all new production dockers from 8.0.1 on will be based on it; ie. will be the future production dockers). Ideal times for this PR would be:

• now (to provide time to test prior to 8.0.0 release in 2.5-3 weeks)
  OR
• after 8.0.0 release in 2.5-3 weeks