I have a collection of static HTML pages that I want to publish to GitLab Pages. These HTML documents were rendered as part of a computationally intensive snakemake workflow running on a HPC. The workflow "knits" HTML reports from Rmarkdown as part of its final steps.
For simplicity, let's assume that I have a single HTML file called index.html and that it contains only "Hello world!!!". Publishing this to GitHub pages is very easy. See this repo which gets published here.
GitLab, on the other hand, seems to insist that I set up CI/CD, even when the static HTML files are already available. I can see the advantages of using CI/CD but in this particular case this will involve setting up a runner and then getting the runner to run the snakemake workflow on the HPC, which could take a couple of days to complete. So I would prefer to solve the simple case first.
Is there a way to bypass the CI/CD? Or can I set up a kind of "null" CI/CD that simply takes the existing static HTML files and places them in public/, for example? What might such a "null" .gitlab-ci.yml file look like?
I tried to create a simple CI/CD pipeline that just copied index.html into public/. The GitLab repo is here. The pipeline fails for reasons that are completely opaque to me. How can I check error logs for the pipeline?
