| Blog | https://runspired.com |
| GitHub | https://github.com/runspired |
| https://www.linkedin.com/in/runspired/ |
| Blog | https://runspired.com |
| GitHub | https://github.com/runspired |
| https://www.linkedin.com/in/runspired/ |
One thing I found interesting in the upset proposal was the callout that some JS engines optimize a specific pattern for upsert and its easy to not hit the pattern.
The pattern they showed was not the pattern I use. So I wrote a micro bench. As always I don't super trust micro benches but it does appear my intuition for which would be faster beats the one that JS engines supposedly optimize.
~30% faster on maps of 100 items, ~10% faster on maps of 10k items, ~20% faster on maps of 100k items.
Apparently Metatext didn’t upload the meme photo 🤣
So here is another attempt
Added in AbortController support and tests. You need not define your own, but if you do it gets entangled. Handlers can abort individual bits via the controller they define and own. Top level signal aborts all!
Things that did not age well: these slides I made last month for a new #javascript talk on how data IS THE framework (titled “loading, please wait…”) 🤣
Starting new features by writing READMEs and docs is awesome #emberdata 🎉
https://github.com/emberjs/data/blob/feat/request-manager/packages/request/README.md