dev/core#339 Install CiviCRM tables using utf8mb4

[seamuslee001]

Overview

This changes the standard civicrm mysql tables to be in utf8mb4 and associated charset

Before

Tables installed using utf8

After

Tables installed using utf8mb4

ping @eileenmcnaughton @totten @demeritcowboy

[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).

[demeritcowboy]

Cool. FYI these 4 tests should fail: #18828 (comment)

The last one I think I was mistaken and it's also just a test update needed.

[eileenmcnaughton]

Ok - so to articulate this makes it so that new installs install in utf8mb4 & the arguments are

1. converting is already supported so the fact more sites will be on utf8mb4 is not a material change
2. there is some risk that some extensions will install in utf8 AND the relevant fields will be used in a join - in which case they are incompatible but the majorty of extensions are probably fine
3. new installs should get the benefit of the new approach & not have to later update
4. this is in effect upping the presure on our existing path
5. we no longer support a level of mysql that does not support uft8mb4 - it's likely some sites still use one but unlikely for new installs

[seamuslee001]

@eileenmcnaughton fixed the tests

[demeritcowboy]

If you install sample data the custom field tables are still utf8 - from civicrm_generated.mysql. Otherwise looks good to me.

[seamuslee001]

Thanks @demeritcowboy updated that now @eileenmcnaughton you good to MOP this?

[seamuslee001]

Jenkins retest this please

[demeritcowboy]

@seamuslee001 that was the 4th test that failed for me earlier. Me thinks it's a real fail and it doesn't like COMPRESSED. I was surprised it didn't fail the first time here.

[seamuslee001]

@demeritcowboy I think I have found it, I had updated install query to specify the ROW_FORMAT was to be DYNAMIC, the problem being is that that was being carried through to the log table even tho the test was specifying COMPRESSED which would have been ok other than the fact that the test was putting in a KEY_BLOCK_SIZE setting for the table which is incompatible with DYNAMIC row format as per https://dev.mysql.com/doc/refman/5.6/en/innodb-compression-syntax-warnings.html

[eileenmcnaughton]

This was notified in the dev-digest. Tests now passing & we have discussed this - merging