Hi folks, Newbie question here I am looking into ...
# general
r
Hi folks, Newbie question here I am looking into Pants to help control jenkinsfile sprawl across hundreds of repos and trying to see how to use pants to encapsulate the build logic. Question for the pants community • we have many repos and at the moment there is zero appetite for a monorepo. lets say I have 10 repos for 10 java/gradle applications. Should I copy/paste the pants.tom across these 10 repos or is there a better way to share pants toml configs across repos • now lets say I have many repos for various applications across different frameworks (python, ios, java, etc). Should we have language/frameworks specific pants.toml or is there some way to have all of them group together somehow
c
What I would try in a situation like this is to crate a repo with the pants basics and add the other repos as submodules. Then it will look like a monorepo from that point of view, but developers can keep ignoring it 🙂 The problem then will become keeping the top repo updated when there are changes, and git submodules are far from ideal... But it's an idea.
It's also possible to go with the google "repo" tool that also gathers multiple repos into one, but it's not simple either.
r
hmmm... i have had some bad experiences with submodules so not sure i want to do this 🙂
so we have gitlab for scm but use jenkins for CI. I was thinking the following • sprinkle/configure pants across all the repos • use gitlab CI to do the build process. ◦ this will allow us to create a parallel build system (gitlab) while dev keeps working on jenkins ci ◦ this allows us to show that gitlab ci process is a lot faster thanks to pants ◦ this also allow us to have developer use pants locally and still benefit from distributed caching • once we show that the bits being produced from jenkins ci and gitlab ci w/pants are basically the same, we can cut off jenkins
I am thinking that is the step 1 . control CI sprawl
step 2 : look into monorepo
thoughts ?