Opinion from SG16 Unicode

---------- Forwarded message ---------
From: Steve Downey <sdowney_at_[hidden]>
Date: Mon, Mar 30, 2026 at 5:31 AM
Subject: The SG19 Priority List for C++29/32 -- feature to read csvfiles
parsing
To: Michael Wong <fraggamuffin_at_[hidden]>


CSV parsing isn't on SG16's list at the moment, but text parsing definitely
is. We will, I hope, start to deliver core library infrastructure for text
fairly soon, and basic parser infrastructure on top of that -- since it's
at this point a fairly well understood problem. Some sort of
rangie-sh combinator-ish things, depending on what we can get people to
propose. Constexpr so that we can interoperate with data from #embed if
necessary. But none of that is really what someone who wants modern data
formats, or even "legacy" like CSV, really wants.

I don't have a real idea what the useful output of a CSV parse would look
like for you. I've done a bit of work with various JSON parsers, and some
schema driven TSV and pipe delimited data parsers, but not ones that are
really suitable for standardization at the moment. Do you have something in
mind already, or is this an open wishlist? From an SG16 perspective, if we
can deliver CSV we should probably deliver JSON as well.

Right now I am cautiously optimistic about what we can deliver in C++29 and
am drawing up a prioritization and RFP paper for SG16 for Brno.

--SMD
Chair SG16