Improve output triggers when logged to file

[eileenmcnaughton]

Overview

When logging is enabled or disabled with the setting logging_no_trigger_permission set to TRUE the trigger sql is output to a file instead of being run. This is used by wmf & probably no-one else.

The issue we are having is the output sql has become a bit messy & inconsistent in terms of

1. extraneous db names in it Fix Schema calculation of usePrefix to cope with rpow: #20471
2. the sorting of the tables has somehow changed making a diff tricky
3. duplication - the calculation of triggers to drop happens more than once - resulting in them all being written out once and then being interspersed into the create statements
4. verbosity - the separators are unset & reset after each line.

This cleans it up by

1. see Fix Schema calculation of usePrefix to cope with rpow: #20471
2. adding an asort
3. putting all the triggers into an array keyed by the statements (for deduping) and only outputting them on class destruct
4. with them all being written at once we can open and close the delimiters just once

Before

drush cvapi Setting.create logging_no_trigger_permission=1
drush cvapi Setting.create logging=1

The output is to a file in config and log - it is overly verbose

After

Per above but logging is less verbose

Technical Details

Comments

[civibot]

(Standard links)

• If this is your first pull-request for CiviCRM, please browse CONTRIBUTING.md for information about the development and testing processes.
• If you are reviewing this pull-request, you may wish to consult the test sites and the Review Standards (long template, short template).

[eileenmcnaughton]

test this please

[seamuslee001]

O_o odd test failure lets see it if repeats

CRM_Utils_versionCheckTest::testCronFallback
Failed asserting that two arrays are equal.
--- Expected
+++ Actual
@@ @@
 Array (
     '4.3' => Array (...)
+    '4.2' => Array (...)
+    '4.4' => Array (...)
+    '4.5' => Array (...)
+    '4.6' => Array (...)
 )

[seamuslee001]

Jenkins re test this please

[demeritcowboy]

Is this the one you were looking for review? I can take a look.

[eileenmcnaughton]

Thanks @demeritcowboy - there is another one too which can be tested in conjuntion - #20471 - note we discussed #24071 & you just need to test for no breakage - the rpow stuff itself doesn't need testing as it's obscure / only used by us & the change should be a clean up on it's own

[demeritcowboy]

jenkins retest this please

[demeritcowboy]

I reviewed both PRs at the same time. Looks good.

I wasn't aware this feature existed. The only differences in the final result (when imported) I see is that the Definer is different but that's a good thing.

[eileenmcnaughton]

thanks @demeritcowboy !