Make TeleIRC bridges respect default_branch override #100
Conversation
There was a problem hiding this comment.
I'm wondering if this is the best approach for the role? It makes the Ansible playbook run very slowly since the binary has to be compiled multiple times for each individual bot. I am not sure it is a major advantage to have multiple bridges running on different versions. Maybe I am not seeing the use case?
| @@ -9,22 +9,25 @@ | |||
| - name: git clone/pull RITlug/teleirc | |||
| git: | |||
| repo: "https://github.com/RITlug/teleirc.git" | |||
| dest: "/tmp/teleirc" | |||
| version: "{{ default_version }}" | |||
| dest: "/tmp/teleirc/{{ item.value.cn }}" | |||
There was a problem hiding this comment.
Does /tmp/teleirc/ need to be created first?
There was a problem hiding this comment.
I believe it creates /tmp/teleirc/ if it doesn't exist yet (pretty sure I blew away that directory during testing).
I simply needed a way to be able to test branch changes over several different bridges. I agree I still feel a little wonky doing things this way as I did not merge this PR in yet even though this playbook is currently running in prod. I'm not sure what the best path forward is for testing without doing things in this manner. I thought of only making two binaries, one running off whatever the default is, and one that is the branch to be tested, but i'm not sure how feasible that is to do with ansible. |
|
@Tjzabel I guess the longer arc to this is RITlug/teleirc#343, so we could pluck the pre-built binaries we want from a tagged release. We could keep this Pull Request and git branch open as a working solution, but I think the long arc here is getting the release automation in place so there is less work on building a binary. What do you think? |
TeleIRC bridges running the Go deployment previously only built a single Go binary running the
{{ default_branch }}.This PR updates that to build each repository individually and starts each project with their respective built binary.
closes #99