Fix mutation coverage #485
Conversation
veewee
force-pushed
the
mutation-coverage
branch
4 times, most recently
from
8cb2355 to
701f013
Compare
Pull Request Test Coverage Report for Build 10633692680Details
💛 - Coveralls |
veewee
force-pushed
the
mutation-coverage
branch
from
701f013 to
e526235
Compare
veewee
force-pushed
the
mutation-coverage
branch
from
e526235 to
59cddb9
Compare
| @@ -32,7 +32,8 @@ | |||
| }, | |||
| "DecrementInteger": { | |||
| "ignore": [ | |||
| "Psl\\DataStructure\\PriorityQueue::peek" | |||
| "Psl\\DataStructure\\PriorityQueue::peek", | |||
| "Psl\\DateTime\\TemporalConvenienceMethodsTrait::since" | |||
There was a problem hiding this comment.
-> Altering since() causes timeouts in the async component.
| @@ -58,6 +60,7 @@ | |||
| }, | |||
| "Throw_": { | |||
| "ignore": [ | |||
| "Psl\\DateTime\\DateTime::__construct", | |||
There was a problem hiding this comment.
-> It's impossible to test the InvalidArgumentExceptions that are being thrown. I'm wondering if it would make more sense to use the same checks from fromParts in the constructor directly.
The constructor does a range check, the fromParts does a calendar (intl) check.
Both result in different type of exceptions with different messages.
Would it make sense to always go for the fromParts() checks instead?
| * @psalm-suppress ImpureMethodCall | ||
| */ | ||
| return Internal\create_intl_date_formatter($date_style, $time_style, null, $timezone ?? $this->getTimezone(), $locale) | ||
| ->format($timestamp->getSeconds() + ($timestamp->getNanoseconds() / NANOSECONDS_PER_SECOND)); | ||
| ->format($timestamp->getSeconds()); |
There was a problem hiding this comment.
the default pattern does not represent nanoseconds.
it is not possible (to my knowledge) to get a timestamp to contain nanoseconds > NANOSECONDS_PER_SECOND - 1.
I did not find a case to cover this calculation
|
|
||
| // $end is the sizeof($values), use static value for better performance. | ||
| $end = 3; | ||
| while ($end > 0 && $values[$end - 1][0] === '0') { |
There was a problem hiding this comment.
mutating the while's results in timeouts. I decided to go for a couple of simple if statements instead that results in the same behaviour.
This PR fixes the infection errors in CI (related to the DateTime component)