https://github.com/henriqemalheiros/bug-report-nuxt-server-middleware-hot-reload
yarn dev;http://localhost:3000/api in your browser;~/api/index.js in your editor;return res.end('api test');http://localhost:3000/api in your browser;yarn dev again;http://localhost:3000/api in your browser;You should see "api test" appearing on the screen without needing to stop Nuxt.
Nuxt is properly watching the files in ~/api but it's not reloading them.
It seems that previously (#258 and #1819) we would use Nuxt alongside nodemon. But since v2.0.0 (#1509 and #3718), server middlewares have watchers support.
Thanks for the report 鈽猴笍
See #4301. Will be fixed in a future version (not v2.4 though)
@manniL Just to confirm the same problem here.
What do you mean exactly with not in this version? Will this still be in Nuxtjs v2? :)
Thanks.
@francoism90 I meant that the next minor release (v2.4) will likely not include a fix for this. I can't give an ETA or similar though.
Closing to unify the (almost same) issue in #4301
Hi, year and a half later - is there any ETA on this issue? This significantly spoils development experience :)
@ruzpuz If I understand the issue rightly, yes, this is no longer an issue. I can't verify it though as the repro repository is now gone. Feel free to open a new issue if not 馃檪
Um.... yeah nah, but... It very much is an issue... Looks like half (not literally but the part I need) of the project is outside webpack and does not get hot-reloaded properly. You can try webpack aliases inside of server-middleware, and it is not working as well.. Quite a deal breaker for me, although the idea is superb... :(
@ruzpuz Please do submit a code reproduction and a new issue! :smiley:
Would a fork of the repository be okay? I am getting prepared to port the project away from nuxt, but before I do that I can fork the repo here so you can play with it if you like
@ruzpuz That would be perfect :+1:
Most helpful comment
@francoism90 I meant that the next minor release (v2.4) will likely not include a fix for this. I can't give an ETA or similar though.
Closing to unify the (almost same) issue in #4301