Photos and text or diagrams in the same image file, at a small size, without losing cripsness
With essentially all browsers being some variation of evergreen now, we can start tossing out old assumptions, like how stark the tradeoff is between filesize, image quality, and features (like transparency) in image compression, especially in images that mix photos with other graphics (like lineart or text).
Extending the HTMLInputElement and using custom events for easy hooks
I’m in the process of moving the (non-wasm parts) of my web front-end for cega over to vanilla Jasvacript WebComponents. I’ve liked how much it cleans up and encapsulates the logic without even needing a framework. The hooks for parsing the file into a byte array on the client-side (for passing to the wasm) were an obvious extraction point, so I pulled it out to a gist micro-project: file-byte-reader.
Dropping Yew and just calling the wasm_bindgen exported function directly
Earlier I figured out how to get HTML file input selectors being called from my Rust code using yew but eventually the complexity of doing everything in essentially js translated to Rust grated on my sensibilities. I can handroll the basic HTML and Javascript easier (and maybe wrap stuff neatly in a WebComponent) and then call out to my Rust lib when I need it.
If you are already using Mastodon to share your content, why not use it as the hub of discussion for your content
Since I’m running a statically generated site via Jekyll, adding comments takes some addition thought. The last time I was looking into this Disqus was starting to take off, but using a third party service for something like this never quite sat right (and some privacy issues they went through later proved that out).
Making a Scrabble word-finder by adding subtraction and overloading the subset operator on #.tally supplied frequency hashes
I choked a bit on a tech screen today (at least by my own standards), by rushing to a solution too quickly, instead of working out a proper tested class. So after dinner I thought I’d revisit it and see what a cleaner solution might look like.
Idiomatic breadth-first versus depth-first tree-traversals in Ruby
Anyone who has gotten very far in a basic “implement this data structure” tech-screen, using Ruby, hopefully knows about just making your class an Enumerable and implementing #each to get all of the methods of that Mixin, which people are used to using on collections in any Ruby project. Less obvious is how you tackle a more complex case, like trees (another interview classic), which have different possible traversal methods of even the same instance.
Pull requests and example usage of yew, gloo, and web-sys
I submitted some pull requests for additional examples to the top level library I’m using for Rust Wasm, yew. Some of the underlying tech and libraries have advanced since most of the examples were created, and there wasn’t a simple one for a very common use case: using a form with multiple fields, including a file-input selector.
Basically just a wget call
ETA: this, of course, I then rolled into a bash script with way too nice of usage()* for a few-liner. [*bash convention for the help function]
Parsing content with Rust regexes in Zola
As you probably know, Google Docs “Publish to Web” settings doesn’t provide the cleanest html, not to mention the giant header it tacks on.