David Anthoff
David Anthoff
More Roslyn links: - [Persistence, Facades and Roslyn's Red-Green Trees](https://docs.microsoft.com/en-us/archive/blogs/ericlippert/persistence-facades-and-roslyns-red-green-trees) - [Red-Green Trees](https://blog.yaakov.online/red-green-trees/) - [INSIDE THE .NET COMPILER PLATFORM – PERFORMANCE CONSIDERATIONS DURING SYNTAX ANALYSIS (#SPEAKROSLYN)](https://robinsedlaczek.com/2015/04/29/inside-the-net-compiler-platform-performance-considerations-during-syntax-analysis-speakroslyn/) - [Roslyn Immutable Trees](https://github.com/KirillOsenkov/Bliki/wiki/Roslyn-Immutable-Trees)...
You can kill test runner processes on the Julia Workspace, they show up there. All the things you describe sound like bugs to me, we should fix those!
Hm, this must somehow depend on the `show` method for types, right? Is https://github.com/JuliaStrings/InlineStrings.jl/blob/9b4ff8ff67dc0441b58383becba970370a30c1d1/src/InlineStrings.jl#L179 the line where the different output originates, right? Do you know under what circumstance that `show`...
The relevant pieces of code are https://github.com/julia-vscode/TestItemRunner.jl/blob/main/src/TestItemRunner.jl#L62 and https://github.com/julia-vscode/julia-vscode/blob/main/scripts/packages/VSCodeTestServer/src/VSCodeTestServer.jl#L56. From just staring at it, I can't really see a difference that would explain this... In both cases we create a...
> (Also there seems to be an off-by-one error in the reporting of the failed test line number, see screenshot below) Hm, weird, that doesn't reproduce on my system: 
Actually, I know what is happening here :) https://github.com/JuliaStrings/InlineStrings.jl/pull/62/files#diff-3b9314a6f9f2d7eec1d0ef69fa76cfabafdbe6d0df923768f9ec32f27a249c63R2 is the reason there is a difference. That line loads the `InlineStrings` package into the `Main` module. When that is the...
I think this is kind of a hard corner case, because the behavior of the `show` method from base simply is so different depending on the context from which one...
Another idea would be to add something like a `@testfolder` macro. Things might look like this: ```julia @testfolder "Foo" begin @testfolder "Bar" begin @testitem "Something begin end end end ```...
New idea: we conceptually change things so that each test item has a name/id and a label. The name/id can be hierarchical. If no label is specified, the name automatically...
Ah, I should clarify: my idea is that the file/folder type nesting we have already stays in place, this would just be an optional way to add more nesting below...