elijahwright_

there's a good reason why this isn't done this way. Git is very open ended but there are .gitignore templates out there. ideally anyone submitting a pull request should be looking at what they're adding. but the real issue here is this:

> Luckily, you realize that you can turn the blacklist of files (the gitignore) into a whitelist, by just ignoring everything and manually un-ignoring desired files.

this will not work and tbh it's fitting a square peg in a round hole. it technically might fit if you shave off the square enough but large repos don't do this, they use a round peg: https://github.com/github/gitignore

I wouldn't know but at least this code has a good foundation to add onto

it works with multiple W-2s I'm pretty sure. the goal is to make it work for everyone which is pretty difficult but I'm motivated to at least work on getting it to work with IRAs because I have one and I want to use this for next year

well, for one, it's more more than 0.1 em dashes per page. the SHARE IT Act has 10 on each page[0]. I don't know how many the 2017 tax cut bill had but it's more than 1,000 and that was over 185 pages[1], and obviously that was before LLMs like ChatGPT. so I don't really know why this is the measure of AI or not, especially because bills have always had a lot of em dashes to start. if you're not analyzing the text of the bill then it's just not going to be accurate

[0] https://www.congress.gov/118/plaws/publ187/PLAW-118publ187.p...

[1] https://www.congress.gov/115/statute/STATUTE-131/STATUTE-131...

this article doesn't make any sense. the bill has a lot of em dashes because that's how bills are expressed and it's a large bill. bills in Congress aren't written with em dashes because it can be confusing with the bill syntax and there's not a reason to do it that way

sorry I changed the name, here's the correct URL - https://docs.openfile.tax/en/latest/reference.html

the code they released was everything that is necessary to run Direct File in development, but they removed code relating to MeF (the IRS's online submission API) and SADI (the IRS's auth system which is integrated with ID.me). most of the code is the backend, Fact Graph (which is a very complex rules engine), the client app (both the form and the screener when you go to directfile.irs.gov), and the state tax API

I'll fix the docs today, thanks