Skip to content

Conversation

@kinke
Copy link
Member

@kinke kinke commented Apr 29, 2025

No description provided.

drpriver and others added 15 commits April 18, 2025 14:29
Fixes dlang/dmd#21244 Fixes dlang/dmd#21246 When parsing a function type, the cparser pushes a new symbols array onto the implicit scope stack, as symbols declared in a parameter-list are visible in the body of a function. Instead of an actual new symbols array it use a null pointer so that the array can be allocated lazily for the common case where it isn't needed. However, the same code is called by places that expect a type only, not a function-definition. Because of this, the implicit stack could get lost, leading to a null symbols array or just the wrong symbols array. Fix this by indicating what the caller expects in regards to this symbols stack so that it can get properly restored.
no TypeInfo generated during semantic analysis for base type of EnumDeclaration
…uses dmd to crash semanticTypeInfo() has no effect for initializers run at CTFE, so redo it if copied elsewhere
…structing a struct literal (dlang/dmd!21262)
…r_" message when initializing invalid array (dlang/dmd!21309)
Which doesn't have a toChars() implementation.
@kinke
Copy link
Member Author

kinke commented May 6, 2025

Unfortunately there's apparently not going to be a v2.111.1 upstream - stable is already at v2.112. I've merged the last stable commit before master was merged (i.e., before stable became v2.112).

The 2 'test' commits so become eager cherry-picks of dlang/dmd#21343 and dlang/dmd#21350, which fix v2.111 regressions showing up with LDC on a Symmetry codebase.

@kinke kinke enabled auto-merge May 6, 2025 19:57
@kinke kinke merged commit ef844b7 into ldc-developers:master May 6, 2025
19 of 20 checks passed
@kinke kinke deleted the merge_stable branch May 6, 2025 20:29
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

4 participants