Kristoffer Carlsson
Kristoffer Carlsson
Also got bitten by this.
I guess that is breaking so perhaps it is not worth doing. Disregarding that, I do think it is ok to have to "double jump" like that though.
It feels a bit tricky to support this. For the REPL, we can look in the history file but for a potentially unsaved notebook file I don't see where we...
Yeah, I don't know how often they are autosaved but parsing that would indeed be a possibility.
I guess `code_string` should return `nothing` if `definition` returns `nothing` though?
Exact same problem here as poster above. Using the fixtures, the data shows up in "Comparison" but "Timeline" and "Changes" are empty.
@mwatts15 Yes! Setting the `DEF_BRANCH` made everything show up. Thank you!
Note that using the same CPU target that Julia itself uses is likely a good idea: https://github.com/KristofferC/PackageCompilerX.jl/blob/1aedf81497fe195cca28d7e6287a189e1fb26518/src/PackageCompilerX.jl#L13
The depot path is set relative to the absolute path of the executable (which is looked up at runtime). I'll have a proof of concept of it in PackageCompilerX in...
As an example https://github.com/KristofferC/PackageCompilerX.jl/pull/13. Running `PackageCompilerX.create_app("examples/MyApp/")` in the repo creates the folder `MyApp` inside that folder with the file structure ``` ❯ tree . . ├── artifacts │ └── 43563e7631a7eafae1f9f8d9d332e3de44ad7239...