TomasHubelbauer

Hey HN, I was frustrated with downloading GitHub invoices for my org by hand so I figured I'd use the GitHub API to download them instead.

…not so fast! The GitHub API doesn't allow one to do that. So I decided to stretch my Puppeteer muscle as I've been enjoying it more and more since I first looked into it and scrape the licenses from my own account.

The only thing I have to do now is fire up this script and enter the 2FA code. The rest is taken care of. Please note that this probably won't work headless as headless downloading still seems to be a to-do for Puppeteer: https://github.com/GoogleChrome/puppeteer/issues/299

I may have to rename this repository, because I have a feeling name-dropping GitHub in the title will get me asked to change it, but that said, I hope you find this useful if you also like to keep the GitHub invoices locally!

Hey everyone, I made this tool which allows one to use it as a Git hook which will then extract modern Office files (ZIP archives of XML files) to directories named by them and take advantage of GitHub/GitLab/Whatever's online diff when inspecting PRs and commits to a repository.

Diffing Office files was not an unsolved problem before, I cite some prior art, but I've decided to take a different route by tracking both the compressed binary files and the uncompressed plain text files in a folder in the repository. This way the diffs can be viewed in online Git hosting services.

I document features and drawbacks of this approach in my repo. It's certainly not efficient, but it solves a different need (the ability to get to the diff in minimal amount of steps) than other solutions and that's why I found it interesting to work on.

Diffs of both the actual XML files (all information, but too much cruft to get a mental picture of the files text content) and TXT files that contain only concatenated text node values of the files in order to get a view of content changes only are available.

In the future I'll work on making this tool faster by skipping unchanged files instead of regenerating them each file, making this suitable for Git repos that track a large number of Office files. I won't worry about the space inefficiency, because I think Git does a good job of compressing the plain text files anyway and there's not much that can be done in this regard on top of that because of the way the tool works.

Hi! I made this GitHub repo to try and show WebRTC data channels and the connection establishment flow in a different way from what I was able to find online.

My aim with this repository is to make WebRTC data channels accessible to any web developer by showcasing offerer and answerer separately for total novices (makes it much clearer what's going on on each end of the connection) as well as combined approach where a peer can be both offerer and answerer.

I use postMessage instead of a real signalling channel, because I got fed up with people bringing whatever 3rd party signalling solutions to their WebRTC tutorials and I want for this to be as simple to set up and run as possible. That's why it returns off file protocol - no server needed, just open and play!

One more goal with this tutorial is to keep it up to date. Once a supported browser (Chrome, Firefox) starts complaining about a deprecated API or the WebRTC spec evolves, I try to rewrite the demo to use the latest and greatest. You can see that in me using async/await instead of promises which also tremendously helps readability.

In the future, I will extend this to support mesh connection topology and make it possible to hold group conversations or multiple conversations at the same time.

If you've been waiting to see what all the fuss with WebRTC is about, I hope this encourages you to start exploring. Any feedback is welcome!