Some feedback to migration guide.

[ScarletKuro]

I hope it’s okay to share some feedback.

It looks like the migration guide at
https://github.com/bUnit-dev/bUnit/blob/main/MIGRATION.md may be missing a few items or could use some clarification:

1. The ToMarkup and ToHtml extension methods (for IEnumerable<INode>) appear to be removed. For example, MudBlazor was relying on these.

2. .Refresh() has been removed. I understand that the removal of IRefreshableElementCollection is mentioned, but when using var xyz = Find(...), this change is not very obvious.

3. Please consider changing the wording:

   - DisposeComponents is now asynchronous and is called DisposeComponentsAsync.
   + DisposeComponents is now asynchronous and has been renamed to DisposeComponentsAsync.

   It took me a bit of time to realize that the only required change was charging from DisposeComponents to DisposeComponentsAsync.

4. ComponentParameterFactory has been removed -> parameters should be set using Render method.

5. ComponentParameter also appears to be gone (internal), but it was used by MudBlazor as well.

6. The guide at https://bunit.dev/docs/migrations/1to2.html seems to be out of sync with the migration guide in the repository. The bunit.dev was the first thing I checked, so the inconsistency was a bit confusing.

[linkdotnet]

Hey - good input.FYI: The MIGRATION.md was our working document and should have been deleted :D.
An oversight on my side.

The rest I will update for sure! Thanks for that!

[linkdotnet]

@ScarletKuro Making many of the methods internal was intentional. Is there any concrete use case where you need ToHtml or ToMarkup (given that this is not bUnit specific but "AngleSharp" specific)?

More general: Is there anything you can't migrate?

[ScarletKuro]

Also, idk if it's worth to mention, but it seems it was changed how RenderCount works. what used to be 1 is now 12, but I just removed these, only six tests used them, and the check itself was somewhat arbitrary and random, like verifying the render count wasn’t really the point of those tests

[linkdotnet]

It would be interesting (probably better in a different Discussion on what AutocompleteConverterStrictTest looks like).

Also, idk if it's worth to mention, but it seems it was changed how RenderCount works. what used to be 1 is now 12, but I just removed these, only six tests used them, and the check itself was somewhat arbitrary and random, like verifying the render count wasn’t really the point of those tests

Hmm that is interesting. While we did change some stuff, it is interesting to see that the number got that high.

@egil and I did some changes regarding traversing the DOM. It might be that this trigger the RenderCount increment quite often.

[linkdotnet]

Also feel free to "copy&paste": NodePrintExtensions.cs into your codebase for the time being.

[ScarletKuro]

Hmm that is interesting. While we did change some stuff, it is interesting to see that the number got that high.

If you are interested to look, here is the test code + razor
https://pastebin.com/HN2RLEHj
With bUnit 2.x that throws Expected comp.RenderCount to be 1, but found 13 (difference of 12).
When a test fails, sometimes I can see in the bUnit exception a render count of 280+.
I honestly can’t say whether it’s accurate now or not, because MudBlazor uses antipatterns like custom logic in setters/getters in Blazor parameters or setting parameters directly, which messes with parent–child communication. I’m still in the process of rewriting that, so I’m not really paying attention to render counts for now.
But I know that this for example works correctly, where the RenderCount does matter: https://github.com/MudBlazor/MudBlazor/blob/dev/src/MudBlazor.UnitTests/Utilities/EventUtilTests.cs as it tests this.

Here is the autocomplete test that was before the migration and started to fail with bUnit 2.x:
https://pastebin.com/dnzk0acA

If you want to discuss that, feel free to start a new discussion and ping me.

[linkdotnet]

I do think there might be a few things at play here. For sure something @egil and I can have a look at. Thanks for providing the example.