When a product is saved its validate_props method is invoked,
and this recalculates the stock_status property based on whether
the product manages stock or not, the stock quantity, and the
value of the woocommerce_notify_no_stock_amount option.
In the case of variable products, and when stock is managed, the stock
was set to "instock" when the current stock was enough, but only
if the "stock_quantity" property was in the list of changed properties
for the object (the method in the base product class doen't check
for changed properties). This is a problem because the
wc_update_product_stock function updates stock_quantity but via direct
database modification, and thus stock_quantity isn't considered
modified. Therefore stock modifications via wc_update_product_stock
don't update stock_status on the product (e.g. when going from 0 to 1
after a refund the stock status will remain as "outofstock").
The fix consists of removing the check for changed properties since
it's not done anyway in the other cases (when stock is below the
woocommerce_notify_no_stock_amount threshold) nor in the base class.
Also, validate_props is refactored for readabiliyy, and an useless
set_stock_status() call placed right before save()
in wc_update_product_stock is removed.
The mapping of the "Automattic\WooCommerce\Testing\Tools\" namespace
to the "tests/tools" directory is moved from manual registration
inside the tests bootstrap constructor to a declaration inside the
autload-dev section in composer.json.
The code hacker as originally designed, as a mechanism that allowed
to enable hacks at the individual test level, is flawed because it
assumes that code files are loaded before each test, but actually
the PHP engine loads code files only once.
Therefore this commit redesigns it so that the two existing main hacks,
the functions mocker and the static methods hacker, are applied
to all the relevant functions and classes at bootstrap time, and
mocks for each individual function/method can be registered at the
beginning of each test. See README for the full details.
Right now, when a product having a parent (e.g. a variation having a
parent variable product) is saved, wc_deferred_product_sync is
executed so that product sync is performed at the end of the request.
This commit implements the same when the product is deleted.
The testing tools (only the code hacker at this time) have been moved
from 'src' to 'tests/Tools', since many opcode cache plugins
load the whole src folder in production.
Also, an extra autoloader is set in the tests bootstrap so that
the 'tests/Tools' directory corresponds, using PSR4, to the
'Automattic\WooCommerce\Testing\Tools' namespace.
- Add methods to temporarily disable and reenable the code hacker.
The code hacker is causing issues in some tests that perform
write operations to the local filesystem. Since this happens only
in a few cases, the easiest fix is to temporarily disable the
code hacker when that happens. This commit adds two new methods
for that in `WC_Unit_Test_Case`: `disable_code_hacker` and
`reenable_code_hacker`.
These methods use a disabling requests count so that the hacker
isn't enabled before it should. E.g. you call `disable`, then
a helper method that does `disable` and `enable`, then `enable` -
then only the last `enable` will have effect.
- `CodeHacker::add_hack` has now a boolean `persistent` parameter.
Persistent hacks won't be cleared by `clear_hacks`.
- `CodeHackerTestHook::executeAfterTest` will now disable the hacker
only if no persistent hacks are registered.
- The existing `file_copy` method is made static for consistency.
- `CodeHacker::restore` method renamed to `disable` for clarity.
The unit testing bootstrap loads and initializes WooCommerce, this
loads a bunch of code files that can't then be hacked in the test hooks.
A workaround is provided in this commit for the case of hacking
static methods. A new StaticWrapper class is created that allows
defining mock methods after the code file has been loaded.
This is applied to all classes from a fixed list in the bootstrap,
before WooCommerce is initialized. The list should be kept up to date
with the list of classes that require such workaround.
- Fix how CodeHackerTestHook::executeBeforeTest parses the test name,
to account for warnings and tests with data sets.
- CodeHackerTestHook now includes a executeAfterTest hook that
disables the code hacker (needed to prevent it from inadvertently
altering further tests). Also, clear_hacks is executed in
executeBeforeTest for the same reason.
- CodeHacker gets restore, clear_hacks and is_enabled methods
to support the changes in CodeHackerTestHook.
- FunctionsMockerHack fixed so that it doesn't modify strings
that are class method definitions.
- Added the WC_Unit_Test_Case::file_copy method, it must be used
instead of the PHP built-in "copy" in tests, otherwise tests
that run with the code hacker active will fail.
This is something to investigate.
Now @hack class and method annotations can be used to register
code hacks as an alternative to using before_ methods.
The syntax is /* @hack HackClassName param1 param2 */
where parameters will be passed to the class constructor.
If the class name ends with "Hack", then that suffix can be
omitted (e.g. "Foo" can be specified instead of "FooHack").
The "code hacker" is a class that hooks on filesystem events
(using stream_wrapper_unregister) in order to allow for dynamically
modifying the content of PHP code files while they are loaded.
The code hacker class allows registering hacks, which are
functions that take source code as input and return the modified code.
A hack can be a standalone function or a class with a "hack" method.
A few hacks are provided off the shelf. One allows mocking standalone
PHP functions (WP, WOO or not), another one allows mocking static
methods, and there's the one that removes the "final" qualifier
from a class definition. This helps unit testing stuff that would
otherwise be quite hard to test.
Since we were converting the field to lowercase we ended up inserting meta in all lowercase, regardless of what it was in the CSV file. We should only be using the normalized field name when looking at the default columns, and should instead rely on a case-insensitive regex for the special columns.
One thing to note is that we're still defaulting the $headers array to the normalized field, as we don't want to change what is being passed to the filter for unmapped columns.
Nestor Soriano
Ron Rennick
Julia Amosova
Peter Fabian
Vedanshu Jain
Claudio Sanches
Christopher Allford
Boro Sitnikovski