![]() ![]() So there you have it: simple GitHub Actions commit linting in most of Viget's favorite languages (try as I might, I could not figure out how to do this in Elixir, at least not in a concise way). Message 3' | docker run -rm -i python:3.11.3 python -c ' Print("Commits must begin with # or (%s)" % msg.strip(), file=sys.stderr) ![]() Python # git log -format=format:%s HEAD ^origin/main | python -c ' Warn %(Commits must begin with or (#įwrite(STDERR, "Commits must begin with # or (". Git log -format=format:%s HEAD ^origin/main | ruby -e ' If you just want something you can drop into a GitHub Actions YAML file to lint your commits, here it is (but stick around and I'll break it down and then show how to do it in a few other languages): steps: After struggling with it for a few hours, I decided to just DIY it with a simple inline script. I looked into commitlint, but I found it a lot more opinionated than I am - I really just want to make sure commits begin with either (an issue number) or - and rather difficult to reconfigure. I figured some sort of commit linting would be a good candidate for continuous integration - when a team member pushes a branch up to GitHub, check the commits and make sure they include a reference to a ticket. In a recent project retrospective, the team identified that we weren't being as consistent with this as we'd like, and decided to take action. And, conversely, nothing sparks nerd rage like fix bug or PR feedback or, heaven forbid, oops. As someone who regularly commits code to apps created in the Obama era, nothing warms my heart like running :Git blame on some confusing code and seeing a reference to a GitHub Issue where I can get the necessary context. But one hill upon which I will die is this: referencing tickets in commit messages pays enormous dividends over the long haul and you should always do it. Looks good.I don't believe there is a right way to do software I think teams can be effective (or ineffective!) in a lot of different ways using all sorts of methodologies and technologies. ![]() Update: Jonathan Wight pointed me to Inkling as an example of a similar platform. I would not at all be surprised if a product appears out of this, and I wouldn’t be surprised if it makes a ton of money. I’ve spent years thinking about it, dreaming about it, fantasizing about it - which is probably only enough to have a small notion of the difficulties involved and no real concept of what it really takes to create a platform and production system like this. They had to create an entire platform for this.Īnd that’s no small achievement. How do you specify the different content and interactions and layouts? You can’t just take snapshots from InDesign and push them out: you need something almost like a Hypercard stack (Hyperpad?) that says what should happen when you tap here and rotate and swipe and so on. What makes me sweat - I can’t think about it too much, because I’ve already spent the last couple years trying not to think about it too much - is what the whole thing looks like under-the-hood. Well, the future will be faster and better, sure, but this is one of the first big sights to see. It very much rewards exploration - I was delighted when I found the 360 panorama of Venice, for instance. ![]() There are plenty of apps out there that do less than a single page of the The Daily does. Maybe I’m speaking more as a fellow developer than a potential The Daily subscriber, and maybe I’m aware that some of my friends worked on it (and I probably can’t say who) - but I am, nonetheless, impressed. I’m not terrifically surprised that a 1.0 app has room for improvement. There are easy criticisms to make: the carousel is a bit slow, transitions between pages sometimes seem laggy, etc. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |