Upgrade front-end & libs to v2.086.0 #3062
Conversation
The _d_switch_error function has been removed from druntime.
|
lit-test |
The `__ArrayCast` lowering isn't (always?) done when casting a static array to a slice with different element size. So exploit constant source lengths and compute the target length at compile-time. This fixes compilable/ldc_github_421.d - a float[16] can be cast to a float4[] with constant length 4.
Invoke _d_newThrowable for exception allocation & initialization, and increase the reference count in LDC-specific _d_throw_exception for MSVC targets. Fixes runnable/test19317.d.
By converting the real_t to a `double` and repainting that value. E.g., this fixes druntime-test-hash for MSVC targets; core.internal.hash.hashOf!(real) reinterprets a 64-bit (target) `real` as `ulong`, and that code path is taken at CTFE too.
|
Looking pretty good. Remaining issues:
|
|
Is there an ETA on this? No hurry, but ideally I would like to be able to plan around a rough time when a beta will be available. |
|
The union regression was fixed upstream. |
kinke
changed the title
|
The |
kinke
changed the title
expressionSemantic() for a ThisExp doesn't just set its type (and do nothing if it's already set), but may also set the ThisExp.var VarDeclaration. This crucial step was previously skipped after syntaxCopying an already analyzed ThisExp, as a later expressionSemantic() for that copy was a no-op and `var` still pointed to the original one. This fixes runnable/test19731.d, as dlang/dmd#9467 may lead to a syntaxCopy of a sema3'd FuncDeclaration, and that copy replacing the original function. A ThisExp's `var` still pointed to the variable in the original function, so LDC assumed it was a nested variable and failed to find a matching parent function.
kinke
changed the title
No description provided.