Kristoffer Carlsson
Kristoffer Carlsson
> I think it is the "escapechar", so usually you'd pass opts::Options.e Hm, I have no `Options` struct. I just parsed as: ``` julia> str = "\"foo\" # comment" "\"foo\"...
> Also, is "foo" the result you expected here? That xparse call is using quoted=true. If you wanted "\"foo\" # comment" you may want to use quoted=false Actually, that is...
I think this is a bit difficult to do because the all the deprecations fixes are run and if two deprecations want to fix overlapping source code, one of them...
Perhaps, I'll experiment a bit.
Could also put CSTParser.jl and Deprecations.jl in sysimg with generated precompile files as well?
> how do we dynamically add these jobs into CI runs though? I wrote > If active is set to false the prerelease version CI scripts should immediately return success....
> I think the main problem that needs to be solved for this would be updating versions.json. The current setup (or any GHA-based setup, I think) wouldn't allow for frequent...
Aha, okay. Maybe that info should be stored somewhere else then.
Need a while loop over `edit_text` so that it will run until convergence, see: https://github.com/JuliaComputing/FemtoCleaner.jl/blob/f3f6a1f53ab8c74960696fdef71a94f9be9b2900/src/FemtoCleaner.jl#L80-L89
Is the memory consumption here including the compilation of the first call to `@interpret`? I realize we still want that to be in the measurement for the total time but...