We needed to update our GitHub app's permissions today to empower some cool new stuff we're building for merge queues and PR preview apps. And it turns out you only get 240 characters(!) to explain why you need new permissions. GitHub, why?
We raised a $7M seed round led by @QuietCapital, and today we're launching Mint, the Continuous Integration and Continuous Deployment (CI/CD) platform with the best developer experience.
https://t.co/BqDU5g02BZ
@sebmarkbage How does checking if `document.createElement` is a function tell you that this is anything like the function you think it is? I think these questions are similar -- you have to make some assumptions about the environment either way.
📣 We just released ABQ 1.5!
🆕 Local test replay lets you replay a test run from CI on your local machine. Debugging test failures in CI has never been easier!
https://t.co/YYHlPV0sRi
@thdxr@flydotio We’re doing something along these lines with Mint, our new build tool. Ephemeral infrastructure has a lot of advantages, but it also results in tons of wasted compute running the same setup steps over and over and over. https://t.co/4LxS3cz6ba
@tekknolagi@rtfeldman We haven’t said much publicly yet, but there’s a little information at https://t.co/4LxS3cz6ba. We have another preview session 5/31 at 10 eastern as well!
@louispilfold@rtfeldman If you need more flexibility, you could wire some of it outside of GHA — e.g., have a GHA script echo “go to x url to enter a TOTP” and then have the script poll some endpoint at that server to grab the TOTP once it’s entered.
This does something similar: https://t.co/trlsWoDbXP
The content-based caching behavior from Mint, the new build tool we're developing at @rwx_research, supports hit/miss/hit sequences. We've been talking about this for a while, and @kylekthomp and @ayaz_hafiz_ made it happen 😀
The latest Code with Jason episode is out!
In this one I talk with @dan_manges, founder of RWX, about how to fix flaky tests. Enjoy!
https://t.co/Sz6f2O71Ai
I love the idea of unifying build and CI tools, but remote executors typically require some system setup which is usually presumed to already be done locally. Alternatively, you could run local builds in VMs/containers, but the overhead is too high.
@dbarrosop@github Ah! Well, if you're still having trouble getting ahold of someone at GitHub, we'd be happy to get you set up with our managed runners so you can keep running your actions workflows while you get it sorted out. DM me if you're interested!
ABQ is highly efficient with CPU utilization. Here's the queue monitoring from a company that went live today with dozens of engineers running their test suite in 16 parallel processes.
https://t.co/KkRmpo2IQA