Ref: https://github.com/amberframework/recipes/issues/32
amber new app -r defaultExpected behavior: Create an app based on default recipe (or any recipe)
Actual behavior: Get an error output:
➜ ~ amber new blog -r default
12:11:40 Generate | (INFO) Rendering App blog in ./blog from default
12:11:41 Generate | (ERROR) Could not find the recipe default : 302 Found
Environment file not found for ./config/environments/production
Reproduces how often: 100%
Amber v0.8.0
Looks like this is issue is due crystal HTTP::Client doesn't support redirects yet, see: https://github.com/crystal-lang/crystal/issues/2721
/cc @damianham
I think we need to move to shards based install asap.
I think we need to move to shards based install asap.
@damianham Yeah, please! amberframework/recipes is already a nightmare to maintain :sweat: Should we archive it?
I'm working in a Vue.js recipe, see: https://github.com/faustinoaq/vue-amber-recipe (not published yet)
I think is better to use recipes like this:
amber new myapp --recipe https://codeload.github.com/faustinoaq/vue-amber-recipe/zip/0.0.1
Just like heroku buildpacks, WDYT?
After looking into this it seems that it could be a while before it’s fixed in the Crystal library. Is it worth putting a work-around in for now. E.g. if a 302 response is found send a second request to the redirect location?
I realise this isn’t an efficient solution, but since this command is only ran during development (and only once per project at that) I don’t feel it would be too much of an issue.
If this is an ok solution I can create a fix for this and do a pull request 🙂
@acoustep I think that's whats going to end up in stdlib anyway, at some point, so that sounds good to me.
I think we can just re-open this PR and merge. https://github.com/amberframework/amber/pull/900
@acoustep Will this fix the issue?
@drujensen yes, that looks about the same as how I would have done it
I am now out of contract so I will have some time to create a PR to install from shards
@acoustep I merged #900 to master. Let me know if this fixes the issue.
@drujensen Can confirm it's working 🙂 (This is @acoustep - wrong account!)
Most helpful comment
I think we need to move to shards based install asap.