@adamchainz NICE! Could be cool that a project is ready to run with harlequin when needed by the developer/admin (maybe outside the project env possibly?). It's useful in both development and production.
Can the manage.py harlequin provide help when harlequin and/or adapter doesn't exist?
> Harlequin was not detected. You should install it outside your project's env:
>
> pip -u install 'harlequin[postgres]'
@benjaoming I’ve just pushed a release that explicitly depends on Harlequin, so it should always be installed. I think Harlequin itself should have a good enough error when the adapter isn't installed?
@adamchainz Since Harlequin contains its own stack of dependencies, isn't it better to install in an environment that's external to the Django project's? I mean, if possible!
There's also the whole footprint of packages.. it's a bit huge :)
@benjaoming Yeah it has a lot of unnecessary upper version bounds... Maybe it would be better to not have it as a dependency, allowing pipx installs etc. I think I will roll that back, later.
@adamchainz super - it was just the input that I had from trying it out. I can make it work for my projects, and trying out Harlequin was really very very promising, so thank you for making it so easy 🫶
@adamchainz Didn't know about harlequin and it's great to have a shortcut from #Django, when tinkering with it. Only thing was that the django-harlequin package does not have harlequin itself as a dependency? In the end harlequin seems not to be my thing. The 90s norton commander vibe and especially the key bindings. And installing harlequin pulls in 29 packages with ~400MB by itself (10 x django), among them numpy, pyarrow. In my case a bit much for the occasional SQL queries I do.
@adamchainz For my own use I just modified the django-harlequin package to not use harlequin, but 🥁 #emacs 😆 . It works quite well for sqlite with emacs sqlite-mode and should work for postgresql with pgmacs as well.
Add comment