model-writtenrun 1081 - may be wrong, links go to the source messages. Commit-by-commit LLM analysis is in the Commits dialog.
- Sami Imseih proposed dshash_get_dsa_area() so users of GetNamedDSHash() could dynamically limit the underlying DSA area's size, preferring an accessor over a creation-time size argument; he also suggested exposing max_size in pg_dsm_registry_allocations without supplying a patch for that extension #1.
- jie wang reviewed the patch and suggested asserting that the hash-table pointer is non-null #2.
- Michael Paquier proposed a wrapper that would call dsa_set_size_limit() directly instead of exposing the DSA area #3.
- Sami retained the original design, explaining that existing dshash functions already require non-null inputs and that exposing the area keeps size-limit handling consistent with GetNamedDSA() #4.
- Michael committed the original patch unchanged as 762e329e83 and announced its application, accepting the API consistency argument and identifying allocation of additional objects in the same area as another benefit; the accessor was complete, while the proposed max_size display remained unimplemented in this thread #5.