refs: https://github.com/TryGhost/Team/issues/558
- we can't use our logging lib to log a config error, because logging requires config to work
- we have to use console, as there's nothing else available at this point
- we can't even use notify to tell CLI a bad thing happened here, because that also requires config and logging
- there is no precedent or ref material for the code 999
- it's just meant to be a clearly different code to -1/1/2/3 we've used elsewhere
- it signifies that things went so wrong we weren't even able to handle the error gracefully with logging
refs: https://github.com/TryGhost/Team/issues/510
- added and wired up the new limit service, which is a lazy-loaded service
- this handles the case that there are host limits set in config, and wraps all the logic needed for detecting exceeded limits & throwing limit errors
- expects limits to be set in config under `host_settings.limits`
- supported limits are managed in the limit service, outside of core
refs https://github.com/TryGhost/Team/issues/513
- The bump does not effect scheduled post as URL with authentication token is generated during runtime for DefaultScheduler.
- This needs to be checked for SchedulerAdapter implementations which use persistant storage
- Continuing cleanup and clarification of the boot process
- By locating requires and inits more closely together its easier to see what happens where and to get timings
- Improved consistency of the debug begin/end statements
- Added "Step" comments to the main process
- Inlined the mountGhost function, for readability and consistency with the other steps
- There is now one true way to start Ghost - you create a server, and then call start with an express app
- We may well expand this again to improve testing pathways in future, but it will be done with a bit more clarity about expectations
- moved logging variable outside of the try-catch block, so it doesn't need a 2nd require
- I was considering having config,logging etc shared components required in a function together, but it's less readable
- Having a try-catch for loading shared components before main boot is also confusing
- This is simplest and most readable IMO
refs dd715b33dc
- this is the last event that is used to trigger part of the standard boot process
- events make the code harder to read/reason about
- the urlservice is one of the most core and critical components in Ghost, possibly the biggest consumer of time and memory
- we want to have the work it is doing front and center so that we can improve it
refs b1a98b0b67
- note: I already replaced server.start with themes.ready in the above commit
- events make the code harder to read/reason about
- long term it would be nice to have a concept of hooks for services, but for now explicit is clearer
- In the old boot the server wasn't started til we were ready
- In new boot, we start the server immediately and send the old started event
- Then, when we are ready to accept some traffic, we send a ready event
- At the moment, ready isn't quite sent at the right time:
- It _should_ be when we're ready to serve real traffic, not just send 503s
- This is after the URL generation has finished
- But this requires more refactoring work :(
- So for now we send when everything else is ready
- This really needs some tests
- Notify is a more familiar name e.g. systemd has the sd_notify system which this is similar to
- We're actually announcing the server started, it's not actually ready for traffic (will serve 503s)
- This commit removes the old boot process and any files that are no longer needed as a result
- Remove the duplicate event for triggering inactive themes to load
- Tidied up a few other bits
- moved the DB init back into the boot file so everything is colocated - want it to be EASY to reason about whats happening
- added some more debug, and made the method signatures all the same
- expanded on comments to make it clear what the intentions are here, as there's a lot more work to do
- With the new boot finalised it's clearer to see where this needs to go
- Update comments to add more clarity around what's happening and why
- Cleaned up the version require, as this was prep for maybe using version via config, which won't work in MigratorConfig
- We currently have 2 boot systems in the codebase, until now the tests were still running against the old one
- Regression tests using the forceStart + subdir flag no longer work. These need reworking, but this _should_ be easier later
- Note those tests work fine if they were the first tests run, it is only the case of trying to restart the server with a subdirectory after starting without that doesn't work
- Currently the new boot process breaks the loading of inactive themes because the theme loader event isn't wired until after the server.start event is fired
- This is a WIP fix, there needs to be some bigger refactoring of the boot process once the old process is gone
refs 0b79abf5b2
- in the referenced commit, we refactored the boot code to simplify it
- one of the required files was not updated with the change, so the path
was wrong
- this wouldn't have been easily visible from local testing because it's
based upon a config flag that isn't on by default
- this commit updates the path to the file
Background:
- Ghosts existing boot process is split across multiple files, has affordances for outdated ways of running Ghost and is generally non-linear making it nigh-impossible to follow
- The web of dependencies that are loaded on boot are also impossible to unpick, which makes it really hard to decouple Ghost
- With 4.0 we want to introduce a new, linear, simpler, clearer way to boot up Ghost to unlock decoupling Ghost into much smaller pieces
This commit:
- adds a new ghost.js file which switches between boot mode with `node index` or `node index old` so that if we find bugs we can work around them this week
- Note: the old boot process will go away very soon, but ghost.js will remain as the interface between the command to start Ghost and the application code
- reworks the database migration process into a standalone utility, so that the DB is handled as one simple step of the boot process, decoupled from everything else
- is missing tests for this new db utility
- leaves a lot of work to do around loading core code, services, express apps in a sensible order, as work to fix this would start to break the old boot process
- doesn't use the new maintenance app because we aren't restarting the server here, instead we have the concept of a "core app" that starts in maintenance mode - need to think about how apps will be decoupled in the near future
