The Candid Startup

One of my reasons for choosing TypeScript as a language stack is that it’s built on an asynchronous IO, event-driven programming model, from the ground up.

So far, my scalable spreadsheet implementation has concentrated entirely on the frontend.

Last time, we integrated NeverThrow into InfiniSheet to add Result<T,E> types for Rust style error handling.

Last time, we looked at general options for error handling in TypeScript.

All real world code needs error handling.

The JavaScript npm ecosystem is built on the foundation of semantic versioning.

There are three big outstanding updates that have been hanging over me for a while.

Last time, I created a reference implementation of the SpreadsheetData interface, including support for editing.

We’re on a roll, steadily filling out the Infinisheet architecture diagram with actual implementations.

Last time, I left you with a laundry list of things to do next.

I finally have a storybook that I’m happy to publish.

Last time, I put together a storybook for my InfiniSheet project.

Now that I have Storybook bootstrapped, I’m going through all my InfiniSheet components creating stories.

It’s been a couple of months since I last worked on features for my scalable React spreadsheet component.

I have an account on Bluesky now.

Something weird happened the other day.

Last time, we had a look at Vitest browser mode and Playwright component testing.

Last time we added Playwright to our toolbox for end to end testing.

As ever I’m spending the week after re:Invent gorging on YouTube videos at 1.5x speed.

I’m ready to add browser based automated testing to my InfiniSheet project.

Last time, I decided that I needed some additional tools for browser based automated testing.

Last time, I was tearing my hair out because I’d broken the core paged virtual scrolling functionality in the big 0.6.0 update.

I rebuilt my react-virtual-scroll package from the ground up.

As we discovered last time, traditional virtual scrolling implementations don’t work properly.

How can this be happening?

It’s time for more enhancements to react-virtual-scroll driven by requirements from react-spreadsheet.

Last time we got basic selection with a focus cell working.

The 0.4.0 release of react-virtual-scroll focused on customization.

I’m about to add support for editing data in my React spreadsheet component.

Last time, we defined a minimal data interface and hooked it up to VirtualSpreadsheet.

Last time, I got fed up with the hardcoded placeholder content in my react-spreadsheet component.

Last time, I left you feeling rather constrained by my starting spreadsheet size of 100 rows and 26 columns.

Last time, I ported some spreadsheet-ish sample code into my new stub react-spreadsheet package, and called it a spreadsheet.

I seem to have spent the past few weeks repeatedly finding reasons not to get my react-spreadsheet package off the ground.

This is my hundredth post to the Candid Startup.

I had just started turning my stub react-spreadsheet package into a real one, when I ran into a dilemma.

I was about to get started turning my stub react-spreadsheet package into a real one, when I noticed a new Vitest related warning from VS Code.

Last time we got good results when we tried out TypeDoc for API reference documentation generation.

Last time I gave up on using API Extractor to generate API Reference Documentation.

I’ve been working on my spreadsheet project for many months now.

I’m using TSDoc comments to provide IntelliSense based documentation for my open source project.

I’ve added some TSDoc based comments to my open source project.

Last time I left you with the realization that documentation is a good thing and that I should probably write some.

My VirtualList and VirtualGrid components use the same approach as React-Window.

I’m putting my money where my mouth is.

Many of us first became aware of the risks of a supply-chain attack with the SolarWinds hack of 2020.

I need a small test app with paged scrolling enabled to use for functional testing.

My monorepo is up and running.

Exciting news!

I figured out how to build usable library packages.