Create a seperate runSetup function
This does the `npm install` (if required) and the database setup.
Since this is now happening in the generation phase, this means we don't need
`compileAndSetup` anymore.
Since compile now sets up the dependencies, we only need to compile in any db command
Reworked the naming and types of combining the various dependencies.
This is the base operation used now to combine package dependencies. It
knows about dev dependencies too. I've adjusted the code for
both web app and server to use this.
I've factored out commonalities in error handling into PackageJsonGenerator
We now support devDependencies for web app and in the deps command
We also have nicer functions to generate the package json entries
directly from the PackageJsonDependencies structure
Create a new resolveDependencies
Use Map to do differences between Wasp and user package.json dependencies.
This is now used internally in resolveNpmDeps.
Also fixes a bug in the conflict error message, where the wasp and user dependency
were previously not properly distinguished.
Add .vscode and move everything that appears text editor related to a section
Allows a user to specify a redirect target in case of successful authentication.
This used to be hardcoded to /, and the current implementation will fallback to that if the property is not specified.
Change-type: minor
Signed-off-by: Giovanni Garufi <nazrhom@gmail.com>
* Introduced AppSpec, replacement for previous Wasp.hs as a core IR in Wasp.
* Organized AppSpec into Core and Domain.
* Removed Domain from AppSpec.
* Moved ExternalCode.File to AppSpec.
* Added some TODOs.
* Added IsDecl to AppSpec.
* Removed ExternalCode File reexports.
* Added comments.
Code is probably not great, but passing tests for generating haskell
data structures exist!
Still have to hand write IsDeclType and IsEnumType instances, but it's cool
Move code into Analyzer.TypeChecker.Internal to clean up interface, change lots of
names to be more readable, add comments to explain the complex parts of typechecker
code.
Instead of just a string, now the possible errors are enumerated. Recursive
error structures are used for dictionary/list type errors. Tests are updated
to check behavior of the new type errors.