Eric Huber

I see one of the coderabbit nitpick comments wants us to combine COPY and RUN into a single command, but as far as I know, COPY doesn't even work like...

I'm not sure if it's actually related, but I recalled a much older issue that was similar: https://github.com/PrairieLearn/PrairieLearn/issues/1505 Linking it here to avoid deja vu later.

I have a proof of concept JupyterLab here with an example question. Open the terminal in Jupyter and run `start-psql.sh` (it should be found in the path already). This is...

I should note that as it is currently, this workspace will only work in the default way of local testing where PL runs workspaces as 0:0. If you do force...

> I don't know much to comment on working with the file size of the database and how to improve that. Would it be possible to include the world.sql (or...

Can you add demo question(s) to the example course to replicate how you're trying to start your workspace? I had messed with the jupyter version previously but not the other...

It looks like the `/browser/[...]` endpoints are baked in to pgadmin and it does try to redirect from `/` to `/browser`, so the rewriteUrl feature of PL won't give us...

I fought with Caddy for hours trying to do a simpler reverse proxy and it's looking like that won't be as clean as I'd hoped. The way that pgadmin and...

I made some progress on a Caddy reverse proxy setup but there are still some issues to debug with how it behaves in production. Locally it seems to work okay....

It's starting up pgadmin (the interface only) pretty well now. It takes about 8 seconds if the image is already on the runner. This is well within the allowed timeouts...