* FIX PHP 8.1 support in MySQLiConnector::query errors
The default error reporting mode in PHP 8.1 has changed from using
errors reported on the connection handle to throwing
mysqli_sql_exception. query() makes no allowance for this, and
functions up the call stack expect to catch
Silverstripe\ORM\Connect\DatabaseException instead - resulting in the
MySQLi exception going all the way up to halt the system.
We can use a try, catch, and finally to retain backwards compatibility,
no matter which setting (e.g. PHP version default) someone has enabled.
* Move MySQLConnector test skip call into setUp()
As review feedback; marking the test as skipped in a private function
obfuscated where the call was happening and made it harder to skimread
the tests. Moving this into a setUp function makes it obvious the check
is run before each test case, and skipped if necessary.
* ENH Added filter to remove empty/null from Path::join parts
This is to ensure null values are removed and trim() deprecated warnings are not thrown
Fixes#10802
* FIX revert public signature
Co-authored-by: Steve Boyd <emteknetnz@gmail.com>
---------
Co-authored-by: Steve Boyd <emteknetnz@gmail.com>
Since 4.12 the use of typehints and return types has caused issues with
values fetched directly from config without validation. This has lead to
upgrade woes in a minor version (#10721) with no immediate recourse
other than manual system intervention.
To use types, we should ensure types, leaving a stable API that won't
error on a bad value - or should give a thoughtful and directive error
message if so.
Issue #10721 summary:
SessionMiddleware runs before FlushMiddleware
SessionMiddleware causes a PHP fatal error passing `null` to a `string`
parameter.
`null` comes from config, because default string value doesn't exist. We
need flush for this - but system execution never makes it that far.
This is to support the mechanism working on all operating systems where Windows may produce a mix of forward and backward slashes in some paths.
For working with the files it may not be a problem, but for exact string comparison the path delimiters need to be unified.
