cryptopunksnotdead / punks.journal Goto Github PK

Hi! Reading through why-the-csv-stdlib-is-broken.md and reading the code examples, I think you'd get a lot of value out of a gem I maintain, Seeing is Believing. Here's a video showing how much easier the code samples could be than what you're probably doing now:

https://vimeo.com/287668754

Also, the published version of the gem has lots of print statements in it. Eg when I run the first example in the readme, it comes out like this:

If you want some suggestions on the code, LMK :) I don't want to come in and shit on things, b/c I also would like a better CSV lib, but kind of skimming through, there's a number of ways to make kind of annoying things a lot simpler, and it's not following gem etiquette on namespacing.

Maybe a last thought: you make a lot of good points, but a mandatory feature of a CSV lib, for me, is that I can use it in a streaming manner (eg process lines as they become available and emit them as they are processed). I can do that currently with the stdlib's CSV, csvreader will need this feature if it wants to compete.

require 'csv'

read, write = IO.pipe
csv = CSV.new read

write << "1,2,3\n4,5,6\n"
csv.readline # => ["1", "2", "3"]
csv.readline # => ["4", "5", "6"]

write << "7,8,9\n"
csv.readline # => ["7", "8", "9"]