I’ve just bought myself a crimbo present of a decent computer rig for once. I’m thinking of putting it to work for Haiku / Paladin development and wanted some advice.
I want ideally to monitor the Paladin GitHub develop branch, fire up a new VM of Haiku for the following configurations, and then run a build and test suite for Paladin on each one:-
- 32 bit Haiku Beta 1 (Won’t currently work - installing devel:libpcre forces an upgrade to a later hrev)
- 32 bit nightly
- 64 bit Haiku Beta 1
- 64 bit Nightly
I also of course need a way to pull down the latest nightlies when a new build is needed.
If anyone has done this before please let me know. Also, if there’s any other automated build/test stuff the wider Haiku community would like testing but hasn’t got the compute to do it, just let me know.
Thanks in advance.
5 Likes
I’ve only done something sort of similar to this recently, when I started to find a way of automating the builds for larger ports from HaikuPorts like Rust, LLVM, V8 and WebKit to name a few by using several cloud providers to see if this is possible.
I used Jenkins for this and have a Haiku VM as a buildslave and this works ok with small VMs with some tweaking to get the memory usage down. The goal of this is to also automate the building and testing of recipes from GitHub PRs on HaikuPorts, possibly via a GitHub Webhook to monitor PRs coming in or something like that. Jenkins might also be suitable for creating an automated testsuite for Paladin with many Haiku VMs on a single host if the machine is powerful enough.
WebKit would also benefit more from having automated building and testing in my opinion.
Great! Any scripts/advice you have would be useful. Thanks.