Support groupmark

[LilithHafner]

Closes #626. These features don't implement themselves, even when seven separate people comment asking for them.

See also https://github.com/JuliaData/Parsers.jl/issues/168. This PR fixes none of that.

This is my first time working with this codebase and its style and I may have missed something. I'd appreciate a careful review.

[codecov[bot]]

Codecov Report

Patch coverage: 75.00% and no project coverage change.

Comparison is base (ae05b87) 90.40% compared to head (0e1d923) 90.40%.

:exclamation: Current head 0e1d923 differs from pull request most recent head fdc5253. Consider uploading reports for the commit fdc5253 to get more accurate results

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #1093 +/- ## ======================================= Coverage 90.40% 90.40% ======================================= Files 9 9 Lines 2293 2294 +1 ======================================= + Hits 2073 2074 +1 Misses 220 220 ``` | [Impacted Files](https://app.codecov.io/gh/JuliaData/CSV.jl/pull/1093?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=JuliaData) | Coverage Δ | | |---|---|---| | [src/context.jl](https://app.codecov.io/gh/JuliaData/CSV.jl/pull/1093?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=JuliaData#diff-c3JjL2NvbnRleHQuamw=) | `88.99% <60.00%> (ø)` | | | [src/chunks.jl](https://app.codecov.io/gh/JuliaData/CSV.jl/pull/1093?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=JuliaData#diff-c3JjL2NodW5rcy5qbA==) | `91.30% <100.00%> (ø)` | | | [src/file.jl](https://app.codecov.io/gh/JuliaData/CSV.jl/pull/1093?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=JuliaData#diff-c3JjL2ZpbGUuamw=) | `94.42% <100.00%> (ø)` | | | [src/rows.jl](https://app.codecov.io/gh/JuliaData/CSV.jl/pull/1093?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=JuliaData#diff-c3JjL3Jvd3Muamw=) | `84.81% <100.00%> (ø)` | | ... and [1 file with indirect coverage changes](https://app.codecov.io/gh/JuliaData/CSV.jl/pull/1093/indirect-changes?src=pr&el=tree-more&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=JuliaData)

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.

[LilithHafner]

That's the kind of request that leads to well-documented packages. Thank you.

[JackDunnNZ]

After this PR, I am seeing a compile error in conjunction with Parsers 2.4.2, so I think Parsers compat needs to be bumped to 2.5.

pkg> add CSV@0.10.10 Parsers@2.4

julia> using CSV
[ Info: Precompiling CSV [336ed68f-0bac-5ca0-87d4-7b16caf5d00b]
ERROR: LoadError: MethodError: Cannot `convert` an object of type Missing to an object of type Vector{String}

Closest candidates are:
  convert(::Type{T}, ::LinearAlgebra.Factorization) where T<:AbstractArray
   @ LinearAlgebra ~/.julia/juliaup/julia-1.9.1+0.x64.apple.darwin14/share/julia/stdlib/v1.9/LinearAlgebra/src/factorization.jl:59
  convert(::Type{Array{S, N}}, ::PooledArrays.PooledArray{T, R, N}) where {S, T, R, N}
   @ PooledArrays ~/.julia/packages/PooledArrays/DXlaI/src/PooledArrays.jl:499
  convert(::Type{T}, ::AbstractArray) where T<:Array
   @ Base array.jl:613
  ...

Stacktrace:
  [1] Parsers.Options(refs::Missing, sentinel::UInt8, ignorerepeated::UInt8, ignoreemptylines::UInt8, wh1::UInt8, wh2::UInt8, quoted::UInt8, oq::UInt8, cq::Vector{String}, e::Vector{String}, delim::Nothing, decimal::Bool, trues::Bool, falses::Nothing, dateformat::Bool, cmt::Bool, stripwhitespace::Bool, stripquoted::Bool, groupmark::Nothing)
    @ Parsers ~/.julia/packages/Parsers/aQvnF/src/Parsers.jl:61

FWIW, it seems like the PR exposed a bug where the parsingdebug and stripwhitespace flags had actually been getting passed as the values to stripwhitespace and stripquoted in older versions of Parsers, as the debug flag was only added in Parsers 2.5

[LilithHafner]

That is unfortunate—I imagine the best fix is the stop mixing up those arguments rather than restricting compat. Unfortunately, I do not have the bandwidth to debug that rn. Perhaps @quinnj is willing to look into it?

[quinnj]

Good catch @JackDunnNZ; mind making a PR to bump Parsers compat up to 2.5?