Igor Rončević
Igor Rončević
The changes in this PR make LSP garbage collection panic. A big thanks to @JoshuaBatty for detailed testing and finding this issue. Joshua's tests (#6665) showed that we already have...
All the GC issues reported in #6665, are now fixed on this branch. Similarly to the [transient bug explained above](https://github.com/FuelLabs/sway/pull/6613#issuecomment-2395539074), the root causes for GC related issues were already existing...
Generated IRs for both cases. Of course, since we have only a single instance of encoding of the `SomeError` type, the encoding gets inlined. In case of several encodings of...
Related to #7500.
> If the error enum instance is known at compile time, compiler can generate the slice at compile time and emit only the log with that slice, instead if encoding...
This is actually not a `__transmute` bug. The issue is, as well explained by @xunilrj, that the u32 and u16 types have size of 8 bytes. Still, we shouldn't have...
Thanks for the detailed proposal @vaivaswatha! Makes sense to me. Flattening in the AST should actually be easier than the current desugaring (which gets very complex in case of or-scrutinees)....
Initial feedback from @vaivaswatha and my remarks: > the proposal is to introduce a higher level IR instruction, init_aggr that represent a high-level concept of an aggregate initialization Just an...