wrbl

Wouldn’t it be nice to deploy code from anywhere? Imagine being in a gym, and after an agent finished the feature, you test it and the feature still needs a slight manual improvement. You do it in-between your exercises, and deploy — all within your phone?

Couldn’t stomach some poorly generated images on socials, so I made a gamified quiz to show some proper examples.

I'm building an app for coding on the go. It's heavily reliant on a SSH connection for runtime, but file browsing, and most importantly, the code edition is completely re-imagined from what e.g. Termius or other terminals offer.

https://codeusse.wrbl.xyz

Coding on the go can be convenient. Select lines of codes by dragging through the line indicators you want to change, and then describe the desired changes in pain English.

tldr; I've made a PR for Mozilla addons-linter and want it to have more visibility for the Mozilla team, so I ask you to make an interaction with the original issue.

--- A while back I've made what-to-click (https://addons.mozilla.org/firefox/addon/what-to-click/) -- an extension for Firefox that let's you create step-by-step documentation from recording your page clicks. Each step is added to a guide with a basic pre-generated message and a screenshot.

It's working pretty well, but one glaring issue is that it doesn't have any way to edit the screenshots, which may contain sensitive information. This prevents usage in some cases, especially when trying to generate a document for general public where each identifiable information may turn problematic to the creator.

Since this is my side-project, and also I want to keep it as simple (yet precise) as possible, so rather than creating a small Paint clone, I tackled the anonymisation with help of tesseract.js. In short, for each screenshot, OCR finds words and boxes those fit into. For each box I create a button that toggles scrubbing. I can't imagine anything simpler, yet precise for the end-user to edit the screenshot. You can see it in action here: https://www.youtube.com/watch?v=2RIWDiZ8X04 .

It's good enough, sometimes it misses words written in less contrast colours, but that's a trade-off between OCR analysis speed and precision I'm willing to take. The problem is that, since I want to keep the extension fully offline, I have to have tesseract's implementation included into my extension. Since tesseract is using WASM, the resulting files are quite big, since those contain binaries that will run heavy loads efficiently. Those big files are making Firefox addon portal linter throw an error when I try to submit the new version. Linter allows max. 4MB files, but the largest one in the project is 4.8MB. With this error I can't proceed with the submission.

The suggestion is to split the file into smaller ones, however I don't see a way of doing so for the WASM module and the way tesseract is loading this file. If you see any way of splitting the file, please let me know, the code can be found here: https://github.com/wrbl606/what-to-click .

Since all my attempts of splitting the file failed (can't use import/export to join files in the runtime, can't access the blob in the window element), I've created an issue following with fixing PR for the addons-linter from Mozilla, that will bump max. single file size to 5MB instead of currently used 4MB. The 4MB limit was set in 2016 and is getting a bit outdated since WASM modules must contain the binary blob afaik, and blocking extensions with WASM modules in them will make us miss many exicting projects in the future (or maybe even in the past, don't really know how many devs already abandoned projects for Firefox because of this limitation).

So, in short, since I don't have any response from Mozilla on my PR, I wanted to ask you to leave a comment or an emote on the original issue for better visibility for the Mozilla team. The issue can be found here: https://github.com/mozilla/addons-linter/issues/4748 .

Thanks in advance <3

Many companies are unable to create builds for Android/Java projects while JitPack team remains completely silent.

Still early in development but already usable. Feedback will be very much appreciated!

Ah, you’re right, here’s the right one: https://addons.mozilla.org/en-US/firefox/addon/what-to-click...

I’ve made a fully offline, open-source alternative to Scribe. It’s in early stages of development but can be already usable.

I’ve made quick and messy SurrealDB usage example for authentication and user data storage in a Flutter app: https://github.com/wrbl606/flutter_surrealdb_client_example

I really like my floating package for Flutter https://pub.dev/packages/floating . It’s very niche as it works only for Android and there’re not so many apps that can use it in a meaningful way yet I like the effect so much that I have to force myself to NOT include it in every app that I write.

Quick and messy SurrealDB usage example for authentication and user data storage in a Flutter app.