The method context::insert_context_record() sets the context::$id to
an int, while when we create an instance, the $id is often read from
a DB $record object, and is a string. The ID should always be an int.
The sequence of questions that made up a quiz used to be stored as a
comma-separated list in quiz.questions. Now the same information is
stored in the rows in the quiz_slots table. This is not just 'better' in
a database design sense, but it allows for the future changes we will
need as we enhance the quiz in the MDL-40987 epic.
Having changed the database structure, all the rest of the code needs to
be changed to account for it, and that is done here.
Note that there are not many unit tests for the changed bit. That is
because as part of MDL-40987 we will be changing the code further, and
we will add unit tests then.
Only MySQL is allowed to run Moodle instances without
(with an empty) database prefix. The rationale is that
"user" is a reserved word in every RDBMS but MySQL.
Before this patch, installation was breaking badly, when
the "user" CREATE TABLE statement was issued. With this
patch the error is shown earlier (on connection) preventing
any advance/use of the database.
This check is present in all the drivers, just sqlsrv was missing it.
The previous behaviour delegates the responsability
of overwriting the 3 public methods to the child
classes, now the field type or moodleform field is
guessed earlier in the execution flow.
Other changes introduced here:
- Fix wrong moodleform detection when there is a form
in the page but the field we are dealing with is
not inside it.
- Updating the last week feature files to these new
step definitions.
- Some coding style changes.
Every single step that sets or gets a value from a field
has been updated to follow the same behaviour both when
using it through a single step or through generic steps
like "I fill the moodle form with:", to resume all the
changes:
- Created a behat_form_group to re-guess the field type
and act appropriately setting and getting it's value
- Normalize all getters and setters to use behat_form_field
children
- Complete behat_form_checkbox to trigger the appropiate JS
event needed to perform some JS scripts that are listening
- Refactor MDL-43713 multi-select changes and remove
the two new steps introduced there as that behaviour can
be managed from the generic getter
- Added a new step definition to check a capability permission
as we changed the way radio buttons gets it's value
This patch implements:
1) Normalization of options. Before the patch options
in a select were being returned as "op1 op2 op3" by selenium
and "op1 op2 op3" by goutte. With the patch, those lists
are always returned like "op1, op2, op3". If real commas are
needed when handling multiple selects they should
be escaped with backslash in feature files.
2) Support for selecting multiple options. Before the patch
only one option was selected and a new selection was cleaning the
previous one. With the patch it's possible to pass "op1, op2" in
these steps:
- I fill the moodle form with (table)
- I select "OPTION_STRING" from "SELECT_STRING"
3) Ability to match multiple options in this steps. Before the
patch matching of multiple was really random, now every every
passed option ("opt1, opt2") is individually verified. It applies
to these 2 steps:
- the "ELEMENT" select box should contain "OPTIONS"
- the "ELEMENT" select box should not contain "OPTIONS"
4) Two new steps able to verify if a form have some options selected or no:
- the "ELEMENT" select box should contain "OPTIONS" selected
- the "ELEMENT" select box should contain "OPTIONS" not selected
5) Change get_value from xpath search to Mink's getValue() that is immediate
(does not need form submission) and works for all browsers but Safari, that
fails because of the extra ->click() issued.
Note all the changes 1-4 only affect to multi-select fields. Single
selects should continue working 100% the same.
The change 5) causes Safari to fail. The problem has been traced down to
the extra ->click() present there. Anyway there are not test cases
requiring that "immediate" evaluation right now. Only the special feature
file attached verifies it.
The previous (slightly whacky) way that we were doing this in previous
versions is no longer supported (and was arguably only supported by being
buggy).
The correct way of doing this is to use a trigger.
