Remove the hard-coded list of allowed CORS origins, and build it
dynamically from the list of "own URLs", i.e. the URLs at which the
Manager expects to be available.
This list of "own URLs" is constructed from the available network
interfaces.
Firefox will still have an additional 1px vertical line at the bottom of
the page, even when `#app` should take 100% of the vertical space. This
causes a scrollbar to show up. Setting the height at `calc(100vh - 1px)`
solves this.
Setting the CSS `color` and `background-color` properties on the top-level
elements will make Firefox understand the dark theme, and colour the scroll
bar appropriately.
This ensures it uses virtual DOM, which is waaaaaaay faster.
CSS variables are used to compute the max height for the Jobs Table grid
cell, to make sure the Tabulator height can be set to 100% without
overflowing.
By having all info for the jobs table in the `JobUpdate` schema, it won't
have to query for the full job when a new job is added. This fetching of
the full job is now delayed until someone clicks on the table row.
Instead of just emitting the straight parsed-JSON object that the SocketIO
library gives us, feed it to the OpenAPI client type system to ensure it's
parsed in the same way as any regular API response.
This is mostly to get datetimes as JS `Date` object instead of as string.
Replace the Vue v2 webapp with a Vue v3 one, and embed the OpenAPI
client in the webapp itself (instead of being its own npm project).
- Vue v2.x -> v3.x
- Tabulator v4.x -> v5.1
- Moment JS -> replaced with Luxon JS
- Vue CLI/UI -> replaced with Vite
This helps to get things consistent on Windows and Linux. Otherwise a path
like `/some/path` is absolute on one platform but not on the other. This is
mostly for getting the unit tests in this package to work on Windows, but
using absolute paths also helps in clarity of error logging.
The upstream buffer takes care of two things: communication with Flamenco
Manager (first context) and buffering things in a SQLite database (second
context). This commit separates those two contexts, so that shutting down
the application isn't going to prevent buffering things in the database.
When shutting down, the main worker context closes. This causes the
subprocess to be killed, which in turn caused a task execution error. This
now no longer happens, as such errors are expected on shutdown and do not
indicate task failure.
