work around problem with finding stdlib files

[KristofferC]

Commited on top of #145

Previously we couldnt see the source for stdlib files:

julia> @enter norm([1,2])
In norm(itr, p) at C:\cygwin\home\Administrator\buildbot\worker\package_win64\build\usr\share\julia\stdlib\v1.1\LinearAlgebra\src\generic.jl:448
1   1 ── %1  = (isempty)(itr)
2   └───       goto #4 if not %1
3   2 ── %3  = (eltype)(itr)

About to run: (isempty)([1, 2])
1|debug> 

Now we can:

julia> @enter norm([1,2])
In norm(itr, p) at C:\cygwin\home\Administrator\buildbot\worker\package_win64\build\usr\share\julia\stdlib\v1.1\LinearAlgebra\src\generic.jl:448
448       isempty(itr) && return float(norm(zero(eltype(itr))))
449       if p == 2
450           return norm2(itr)

The method printing is still wrong though.

[KristofferC]

I pushed a commit to also fixup the printing of the frame.

If this is too "hand-wavy" I'm happy just moving this to the debugger instead.

[pfitzseb]

I still kinda feel like CodeTracking's whereis should handle this.

[KristofferC]

Yeah, I'm inclined to agree. Would be simple to move there. Let's wait and see what Tim thinks.

[KristofferC]

I guess we would still need it here for the frame printing and getfile but we could call the equivalent of maybe_find_stdlib_file from CodeTracking then.

[timholy]

#144 (comment).

But I'd be happy to have a PR in CodeTracking with this replacement operation! It should be performed here.

[KristofferC]

timholy/CodeTracking.jl#16

[KristofferC]

Updated the PR based on timholy/CodeTracking.jl#16. Note that the Manifest here for CodeTracking now refers to that PR branch (to run CI) on it but should not be merged as is.