REPL doc lookup assumed ASCII for the given string

[Seelengrab]

Fix #41589

Original code assumed ASCII, but that's not true. There may be other places in Base or an Stdlib where this happens, I'll grep and possibly open another PR if I find more.

Also, is there a better way to check whether something just doesn't throw? Checking the type of the returned value certainly seems like a way, but is of limited use for other code paths and kind of feels wrong...

[jakobnissen]

I'm pretty sure the indices 1:end-1 works for any string that ends in a =, so I'm not convinced the old behaviour is incorrect.
Edit: You are right and I was wrong.

[louie-github]

A quick test shows this not to be the case:

julia> "÷="[end-1]
ERROR: StringIndexError: invalid index [2], valid nearby indices [1]=>'÷', [3]=>'='
Stacktrace:
 [1] string_index_err(s::String, i::Int64)
   @ Base .\strings\string.jl:12
 [2] getindex_continued(s::String, i::Int64, u::UInt32)
   @ Base .\strings\string.jl:233
 [3] getindex(s::String, i::Int64)
   @ Base .\strings\string.jl:226
 [4] top-level scope
   @ REPL[1]:1

[Seelengrab]

I'm pretty sure the indices 1:end-1 works for any string that ends in a =, so I'm not convinced the old behaviour is incorrect.

The byte index before a valid Char index may not be a valid Char index itself (except for single-byte encodings like ASCII). In this case, the character in front of = is a multi-byte character of length 2 bytes, that's why end - 1 is not valid here.

---

Thanks for the quick review, did not know about chop! Should there be a comment next to the test, saying that the isa Markdwon.MD check is just used for confirming that it doesn't throw, instead of requiring that return type?

[KristofferC]

Should there be a comment next to the test, saying that the isa Markdwon.MD check is just used for confirming that it doesn't throw, instead of requiring that return type?

I don't think it really matters. Someone can always follow the issue -> PR trail from the testset name if they are really interested.