Michael
Michael
When parsing a type-def, we currently only store the name of the referenced type in the generated `DataTypeDeclaration`. If the aliased type does not exist, it is currently not possible...
When parsing data-types, we include the initializer in the data-types source locations, leading to confusing error messages when converting the location to a slice: ``` FUNCTION main: DINT VAR x:...
``` VAR_CONFIG main.foo.bar AT %IX1.0 : STRING[100]; END_VAR ``` currently leads to a panic in `index::visitor::visit_global_var_block`, `src/index/visitor.rs:313:50` due to the custom string type not having a name. _Originally posted by...
I think we can start phasing out the @... markers since we can use pragmas at this position _Originally posted by @ghaith in https://github.com/PLC-lang/rusty/pull/1326#discussion_r1789710963_
Changes debug-info generation for struct members, which are internally generated parent-class instances, to simply have the name `SUPER` instead of `__`.
Add support for typed enums with explicit integer type specifications, supporting both IEC 61131-3 standard syntax (TYPE NAME : TYPE (...)) and Codesys syntax (TYPE NAME : (...) TYPE). Additionally,...
**Describe the bug** The compiler currently fails to detect when type names are used as values in expressions during the validation phase. This results in a late-stage codegen error that...
We might want to add an "overflow" detection for values outside of `f32` range similar to the existing integer overflow detection to warn the user about truncation, as can be...
The following example of statements currently fails during codegen: ``` PROGRAM exp VAR x, y : INT; END_VAR +x := 1 + 4; // codegen err | `UnaryExpression { op+,...
**Describe the bug** Trying to compile a self-referential struct (i.e. a linked list node) with a pointer to its own type overflows the stack when compiling. **To Reproduce** Steps to...