Closed angelikatyborska closed 5 months ago
I'm not convinced we should use git submodules for the CI. On the Elm Analyzer repo, we have these lines to check out the analyzer, then the repo in some subdirectory (effectively, same result as a submodule but much cleaner) and then run a script for running the analyzer on the repo exercises. What do you think?
What's wrong with using a submodule?
I just copied the approach of the elixir-analyzer 🤷
Nothing wrong per se, it's just kind of a pain to keep up to date, the command is long, and even if we don't care about keeping it up to date it's still kind of noisy in git status
.
the command is long
This one: git submodule update --recursive --remote
? I never use that. I just cd elixir; git pull; cd ..; git add .
😁
I want a solution where I can run all the tests locally as well as in CI, and it should be reasonably fast locally.
If I see correctly, the elm analyzer accepts the path to the elm repo as an argument? I guess I can rewrite the Elixir tooling to do the same.
On second thought, I'm not crazy about the idea of having a clean state of my elixir repo locally to run analyzer/representer tests 🤔
Blocked by https://github.com/exercism/elixir-representer/pull/101
I extracted the fix to a separate PR so we can get it merged independently of the discussion about the right approach for this PR
That sounds like a nice idea, but maybe it would be possible to just trigger the CI on elixir-representer main branch whenever the CI on elixir main branch runs?
I haven’t done anything like that before, but sure if it’s possible it’s a good option.
WIP