IdentId will now reserve the MSB for flagging whether the ident
is suffixed with a `!`.
We will use this later to constrain identifiers to be effectful or pure.
When an error message reports on a symbol that was generated during
canonicalization, use text like "This value" instead of "This `123`
value". Generated symbols use the identifier index as the symbol name,
since valid Roc variables cannot begin with a number so there's no
chance of collision. We don't want to display generated symbols to the
user, so when building the error message we check if the symbol's name
starts with a digit.
TotallyNotJson.roc now lives on the farm in virtual-dom-wip as Json.roc.
Any reference in stdlib or builtins has been removed, as well as the
last places it was used (in python/ruby-interop examples).
We will now show a warning if a builtin is imported explicitly,
since this is unncessary.
We will not show the warning if they expose functions from the builtin:
import Dict exposing [isEmpty]
However, we will show a special warning if they expose types from it:
import Dict exposing [Dict, isEmpty]
We were still passing `ModuleIds` from `load` to `can`, but now
that imports can appear in any scope, we don't know which package
an unqualified module name belongs to from the top level.
We now pass `PackageModuleIds` instead and keep a Map of `ModuleName` to
`ModuleId` in `Scope`.
This also allow us to import multiple modules with the same name from different
packages as long as a unique alias is provided.
The usize one gets used internally for things like
pattern matches. This is both more efficient (means
they don't have to do unnecessary casts) and also
less error-prone due to e.g. comparing length to
capacity, which is usize.
