New APIs to release code on Heroku | Hacker News Reader

New APIs to release code on Heroku

blog.heroku.com

86 pointsfriism10 comments

This is really awesome. I've been using Heroku for a while now, and the service just keeps getting better.

I think their platform API is actually one of the most underutilized tools out there for building really cool stuff. Last year a friend and I built postgression (http://postgression.com), to showcase how powerful it can be (for fun, of course).

I've got about a million other ideas for using the platform API to do awesome stuff.

Good job, guys (and gals!).

That's a neat tool to have.

Nice. Deploying a Scala / Play Framework project on Heroku is painful because the compile times out about half the time. Being able to compile locally and push up a zip is going to make like much nicer.

Hey there, I work at Heroku - we'd love for you to try out these new APIs, but also feel free to [log a ticket][1] and we'll see if there's a way to help you get faster builds.

[1]: https://help.heroku.com/

Meh I just finish writing a script that deals with deployment. This may actually solve my problem. http://stackoverflow.com/questions/20331106/application-stru...

Yes. Pretty much the only thing I really dislike about Scala is the speed of the compiler, and Heroku magnifies this about 70 times by requiring a complete rebuild of the slug on every push. It would be nice if it cached the build artifacts and only recompiled source files that changed - that would be a huge time saver and would mean I could deploy several times a day instead of only the one time I allow myself. Perhaps a workaround exists but I just haven't seen a way to accomplish this

To be honest I love using Git push to deploy. I just wish it was faster. This doesn't really solve that problem for me - building the slug on my own machine or as part of a CI setup seems like a step back.

That and Heroku's MySQL-compatible service, ClearDB, is very inconsistent, with query times for simple queries ranging from normal to a couple of seconds inexplicably.

You may well be interested in the build cache[1], which allows you to store build artifacts between pushes - I think it was designed for asset compilation, but there's no reason it shouldn't also work for compiled code.

[1] https://devcenter.heroku.com/articles/buildpack-api#caching

I hadn't seen that - sounds like it will help. Thanks a bunch.

Ever since the Rap Genius gaffe(http://news.rapgenius.com/James-somers-herokus-ugly-secret-l...), I view Heroku with a totally different angle.

I don't recall there is anything concrete offered to affected customers, except a generic apology, and this keeps me from doing business with Heroku. Right now I am building a web stack from scratch on a US-based hosting vendor. It is frustrating and we kind of are reinventing the wheel, but the reassurance that all things are transparent and within our control is priceless. Your opinion could be different.

My experience with Heroku has been drama free. I have never pushed a high-traffic system to their servers. For a reason: for the price, I'm better off managing my own boxes and not have to worry about data security and performance. Heroku for me is for projects that quite don't require much in order to run (as in not many users, data is not very important, etc).

Reply on news.ycombinator.com