• Mohamed El Shorbagy (10/10) Jan 12 # Summary of Progress (Jan. 05 - Jan 12)
• Paul Backus (7/13) Jan 13 Have we really lowered our standards so far that we're accepting

Mohamed El Shorbagy <mohrizq895 gmail.com> writes:

This week I continued work on refactoring AST nodes to break the 
dependency on semantic, backend helpers and special output such 
as `doc`. The better approach, of course, would be to break the 
dependency of files such as `doc`, `dcast`, etc., on semantic 
analysis, but I assume that would be very time-consuming and 
complicated. Here are some PRs:

- https://github.com/dlang/dmd/pull/22356
- https://github.com/dlang/dmd/pull/22389
- https://github.com/dlang/dmd/pull/22390

Paul Backus <snarwin gmail.com> writes:

On Monday, 12 January 2026 at 23:21:03 UTC, Mohamed El Shorbagy 
wrote:
 The better approach, of course, would be to break the 
 dependency of files such as `doc`, `dcast`, etc., on semantic 
 analysis, but I assume that would be very time-consuming and 
 complicated. Here are some PRs:

 - https://github.com/dlang/dmd/pull/22389
 - https://github.com/dlang/dmd/pull/22390

Have we really lowered our standards so far that we're accepting 
PRs that replace typed pointers with `void*` and `cast`?

Surely even a solution as crude as moving the `DocComment` into 
an external hash table, keyed on the object's `this` pointer, 
would be preferable to what was actually merged.