Asset compilation and build pipelines


  • I survived the hour long Uno hand

    We're working on a solution for Node.js, if that's useful.

    Here's the problem: we have two repositories in this particular dilemma. One is a web app. The other is a library of javascript components to be re-used across many apps.

    We want to minimize HTTP requests. Therefore, in our current build script, which is executed directly on the demo server, we grab the (already minified) components the app needs and concatenate them, so we can serve the result, which is cacheable.

    However, I'm building a build server. The web app repo won't build by itself, it needs the components in a specific spot on the server to build.

    Where should this logic go? What do other people do?



  • Our front-end people "solved" this by using Git Submodules. I do not recommend that solution.

    Does the build server not have the ability to check-out a repo? At my previous employer, we used a build server that had read-only access to the repo and just checked-out or updated stuff as needed.


  • I survived the hour long Uno hand

    It can. Is that the best option? Having duplicate checkouts of the component library on the build server for each application that uses it?

    I'm kind of leaning toward segregating the step that needs the extra library from the steps that do not and running the former during the deploy phase, on the server it's deployed to.



  • @Yamikuronue said:

    It can. Is that the best option? Having duplicate checkouts of the component library on the build server for each application that uses it?

    Set up the build server like you would a new developer. Somewhere on its local HD, you check-out all the repos so the build server can refer to them as needed. Write your source control updates in the script to happen when appropriate.


Log in to reply