You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The current tests as they stand just take too long. In part, this is because some datasets get hit repeatedly (accessing annual datasets to retrieve individual storm data). But the slowest by far are the forecast/advisory tests. These have been trimmed down but they still take a bit of time.
The forecast/advisory tests should be expanded. A concern is modifiying a regex pattern to fit an issue for a specific advisory but consequentially breaking it for another advisory; there are many inconsistencies in these products. At this point of time there is no real way to know that patching one bug will not open up another. That's an issue.
I'm opening this issue to start noting ideas and methods to redevelop the tests. The time issue isn't going to get better; it just is what it is. But the tests needs to be redeveloped to accomodate the issues noted above and streamline the tests as much as possible.
The text was updated successfully, but these errors were encountered:
Many tests were removed during development of hotfix/v0.2.0-6. Some are ok as they tested functions that no longer exist.
However, tests removed in commit dea1060] need to be re-added. Unsure at the moment why they were failling. When ran tests check they failed. But if ran from script or console they passed.
The current tests as they stand just take too long. In part, this is because some datasets get hit repeatedly (accessing annual datasets to retrieve individual storm data). But the slowest by far are the forecast/advisory tests. These have been trimmed down but they still take a bit of time.
The forecast/advisory tests should be expanded. A concern is modifiying a regex pattern to fit an issue for a specific advisory but consequentially breaking it for another advisory; there are many inconsistencies in these products. At this point of time there is no real way to know that patching one bug will not open up another. That's an issue.
I'm opening this issue to start noting ideas and methods to redevelop the tests. The time issue isn't going to get better; it just is what it is. But the tests needs to be redeveloped to accomodate the issues noted above and streamline the tests as much as possible.
The text was updated successfully, but these errors were encountered: