Julia: Stacktrace counter sometimes gets stuck and doesn't start at one

[KristofferC]

Error During Test at /Users/kristoffer/julia/test/testdefs.jl:19
Stacktrace:
   [53] pkgerror(::String) at /Users/kristoffer/julia/usr/share/julia/stdlib/v1.0/Pkg/src/Types.jl:120
   [53] #generate#2(::Base.Iterators.Pairs{Union{},Union{},Tuple{},NamedTuple{(),Tuple{}}}, ::Function, ::Pkg.Types.Context, ::String) at /Users/kristoffer/julia/usr/share/julia/stdlib/v1.0/Pkg/src/generate.jl:8
   [53] mktempdir at ./file.jl:595 [inlined]
   [53] temp_pkg_dir(::getfield(Main.Test56Main_Pkg_pkg.OperationsTest.REPLTests, Symbol("##36#47"))) at /Users/kristoffer/julia/usr/share/julia/stdlib/v1.0/Pkg/test/utils.jl:18
   [53] top-level scope at none:0
   [53] include at ./boot.jl:317 [inlined]
   [53] include_relative(::Module, ::String) at ./loading.jl:1040
   [53] include at ./sysimg.jl:29 [inlined]
   [53] include(::String) at /Users/kristoffer/julia/usr/share/julia/stdlib/v1.0/Pkg/test/pkg.jl:3
   [53] top-level scope at none:0
   [53] include at ./boot.jl:317 [inlined]
   [53] include_relative(::Module, ::String) at ./loading.jl:1040
   [53] include at ./sysimg.jl:29 [inlined]
   [53] include(::String) at /Users/kristoffer/julia/test/testdefs.jl:13
   [53] macro expansion at /Users/kristoffer/julia/test/testdefs.jl:22 [inlined]

Perhaps https://github.com/JuliaLang/julia/pull/28016 @Liozou?

[ararslan]

Stack traces are now 53-indexed instead of 1-indexed, nbd. :trollface:

[Liozou]

That's odd. Do you have a MWE?

[Liozou]

I think I can safely git blame @timholy for https://github.com/JuliaLang/julia/commit/c33de5809b3fb6e2b77f4d43a5a0acd911c8152a#diff-a98b3bbc7110e1f8b626430f6e4975feR540: the loop does not increment frame_counter accordingly. I don't know the implication on the rest of the code so I'd rather not touch it.

[Liozou]

What bugs me here is that the stacktrace has less than 50 stackframes, so the faulty function show_reduced_backtrace should not even have been called...

[KristofferC]

I didn't paste the full output. It could have been quite long.

[Liozou]

OK, that makes sense

[Liozou]

28328 is actually breaking #28016 pretty badly, I'm looking into that.