search

gui-cs / TerminalGuiDesigner

Forms Designer for Terminal.Gui (aka gui.cs)
MIT License
423 stars 28 forks source link

Round 5 of test conversions #283

Closed dodexahedron closed 8 months ago

dodexahedron commented 8 months ago

Opening this one now as a unit, as it contains a few fairly lengthy tests, as well as a couple of TODO notes for potential issues and future work items to address.

A couple of tests did get some pretty significant upgrades, as well as some additional coverage from things like adding additional types/cases to certain tests (more can still be done, but I am more interested in completing the conversion, right now).

This branch is branched from convert-unit-tests-to-constraint-model-4, which is branched from v2 as it was a couple of days ago, after you merged some of your other work and updated the TG reference.

This is of course cumulative with convert-unit-tests-to-constraint-model-4, and they can either be individually merged in order or 4 can be closed and this one merged all at once.

This will stay as-is unless edits are needed. Continued work will be in a new branch, in the same style as the rest (convert-unit-tests-to-constraint-model-6), branched from here.

Ready for merge at any time.

dodexahedron commented 8 months ago

Just a note:

Round 6, which is branched from this, is in progress right now, and has had the base v2 branch merged into it, as of a few minutes ago. There weren't even any conflicts, no build errors, and all tests pass. 🥳

So, as of right now, the convert-unit-tests-to-constraint-model-6 branch is equal to v2 here plus everything I've done since the last merged PR.

tznind commented 8 months ago

I have started reviewing this, looks great but have been sidetracked as I test stuff a bit as I have found several new bugs.

I will focus on this one as 6 has a bug in resizing (see my comment on that PR).

dodexahedron commented 8 months ago

Yeah I'm thinking the bug in 6 may actually just be that a different pre-existing bug is now being hit, with the fix no longer hiding it.

But, I am not certain of that yet - maybe 70% confident, since the tests show hit testing working properly for identifying IsLowerRight when it should and not when it shouldn't. That makes me think something else is calculating it on its own, but I'll have to check it out. I'll work on that before I continue with more conversions.