ahlec / alexandria Goto Github PK

Conformity tests: A special subset of our unit tests that test to make sure that the websites are in the format that we expect and that our code is still parsing the data correctly (we're making sure that we still conform to AO3/FF.net/etc)

"Searching" is where you're configuring the query you want to make, and then you query that search and everything after that point (going to the website and then processing the data and such) is part of "Querying"

All other delineations/subgroups of unit tests are no longer necessary.

just wanted to say hi since our projects are somewhat related :)
https://github.com/Zeks/ffnet_search

Ok @ahlec, I have 2 questions:

1. Is Alexandria functional yet? (Like is it ready to be used in a project)

   and

2. I've only ever imported libs from NuGet and MyGet before, how would I install this to my project?

Any help is much appreciated,
-Max

This will wind up flooding the project with warnings, but it will also make figuring out what isn't documented SO much easier.

(which will mostly be exposing tag-related things in a more public/specific way)

The general differentiation is fine, but how do we solve overlapping things? ie, do we have two tests for every feature -- one that's a conformity test to ensure that our API still works with the website, and one that just makes sure it deserializes fine?

We need concrete documentation for the difference between the two and when to write one as one thing and not the other.

We'll need to wrap the RequestHandleBase.Request function in a bunch of code that handles cases for where we have HTTP errors or a timeout or other network irregularity.

It's not a 1:1 relationship, it's a 1:many relationship.

Fix all violations of StyleCop rule SA1201 across the entire solution. This honestly is just a bunch of reordering, so it sounds worse than it actually is.

Under the current architecture, every time that AO3 adds/removes a language, it breaks the entire assembly. Even if we were to detect that change within, say, 1 hour and deploy a fix for it immediately, the requires that every user of Alexandria ALSO update immediately, because all of their applications will break. This is not a sustainable system, and this is a must-fix before we can launch.

We don't need to document the unit tests like we're documenting the rest of the project, so we should disable that analyzer rule for Alexandria.Tests