Explicitly delete copy constructor and assignment in Timer (#665) ## Summary `Timer::~Timer()` calls `Report()`, logging the elapsed time to a tracer or stdout. With no explicit copy policy, a copied `Timer` would produce a second `Report()` on its own destruction — silently logging a duplicate timing entry for the same measured interval. `tag_` and `t_` are non-owning pointers (no `delete` in the destructor), so there is no memory-safety risk. The issue is purely semantic, but it is a real hazard for callers who pass or store `Timer` by value. Explicitly `= delete` the copy constructor and copy assignment to make the non-copyable intent clear. ## Test plan - [x] `basic/gpu` builds cleanly in debug mode - [x] Full debug build passes with no errors

• Summary
• integration.mpi 「1」

Refactor izstream raw zip_stream_p_ pointer to unique_ptr (#664) ## Summary - `izstream` held a raw owning `zlib_stream::zip_istream*` managed by a custom destructor, with copy/assignment only blocked by inherited `private` declarations in `irstream` — no explicit policy on `izstream` itself - Replace `zip_stream_p_` with `std::unique_ptr<zlib_stream::zip_istream>`, eliminating the manual `delete` in the destructor - Explicitly `= delete` copy constructor and copy assignment directly on `izstream` - Update all allocation/reset sites in `izstream.hh` and `izstream.cc` to use `unique_ptr::reset()` ## Test plan - [x] `utility/io` builds cleanly in debug mode - [x] Full debug build passes with no errors

• Summary
• integration.mpi 「21」
• integration.release_debug 「1」

Hotfix of T919 and T920 tests for ubuntu.clang.cxx11thread.serialization.python38.PyRosetta.unit test (#666) This PR removes a few built-in `hash()` function calls from the `T919` and `T920` unit tests, which are non-deterministic in Python-3.8.

• Summary
• integration.addsan 「1」
• integration.mpi 「1」
• integration.release_debug 「1」
• ubuntu.clang.python313.release.PyRosetta.MinSizeRel
• ubuntu.clang.python311.release.PyRosetta.Release
• ubuntu.clang.python312.release.PyRosetta.Release
• ubuntu.clang.python39.release.PyRosetta.Release
• scientific.protein_data_bank_diagnostic.cif

Fix memory leak in SimulateMPI: use unique_ptr for simulation_ ownership (#660) ## Summary - `initialize_simulation()` unconditionally assigned `new SimulateMPIData*` without deleting any previous allocation, leaking memory on re-initialization - `simulation_` was a raw owning static pointer with no cleanup path - Fixed by switching `simulation_` from `SimulateMPIData*` to `std::unique_ptr<SimulateMPIData>`, so `reset()` frees the old object on re-initialization and the static is properly cleaned up at program exit ## Test plan - [ ] Build `utility` library in debug mode: `cd source && ./scons.py -j16 mode=debug utility` - [ ] Run utility unit tests: `cd source && python test/run.py -j16 --mode=debug`

• Summary
• ubuntu.clang.cxx11thread.serialization.python38.PyRosetta.unit 「2」
• integration.mpi 「2」
• integration.tutorials 「1」
• ubuntu.clang.python312.release.PyRosetta.Release
• release.PyRosetta.conda.Release 「2」
• scientific.glycan_dock
• scientific.make_fragments.debug

Fix memory leaks in Inline_File_Provider (#659) ## Summary - `clear_input_files()` and `remove_input_file()` were leaking `std::stringstream` objects by calling `vector::clear()`/`erase()` without deleting the raw pointers - `get_ostream()` also leaked a `new std::stringstream` on every call - Fixed by switching `input_files` and `output_files` members from `std::vector<std::pair<std::string, std::stringstream*>>` to `std::vector<std::pair<std::string, std::unique_ptr<std::stringstream>>>`, so the vector destructor and `erase()` handle cleanup automatically ## Test plan - [ ] Build `utility` library in debug mode: `cd source && ./scons.py -j16 mode=debug utility` - [ ] Run utility unit tests: `cd source && python test/run.py -j16 --mode=debug`

• Summary
• ubuntu.clang.cxx11thread.serialization.python38.PyRosetta.unit 「2」
• integration.tensorflow 「1」

Clean up PyRosetta Sphinx documentation (#657) This is a very minor PR to fix a link in the PyRosetta Sphinx documentation and to properly format the PyRosetta initialization file docstrings for PyRosetta Sphinx documentation. Runtime functionality remains unchanged.

• Summary
• ubuntu.clang.cxx11thread.serialization.python38.PyRosetta.unit 「2」
• integration.mpi 「2」
• integration.tensorflow 「1」
• integration.tutorials 「1」
• maintenance.documentation
• release.PyRosetta.conda.Release 「2」
• scientific.make_fragments.debug
• scientific.mp_lipid_acc
• windows.build.debug

Update academic license information in LICENSE.md (#656)

• Summary
• ubuntu.clang.cxx11thread.serialization.python38.PyRosetta.unit 「2」
• integration.addsan 「1」
• integration.mpi 「2」
• integration.release_debug 「1」
• integration.tutorials 「1」
• maintenance.documentation
• release.PyRosetta.conda.Release 「2」
• m1.clang.cxx11thread.serialization.unit 「1」

Quick update: change which HKDF function is used by HMAC-SHA256 verification machinery in PyRosettaCluster (#655) This quick patch has virtually no effect on security or runtime functionality, it just updates which extract-and-expand hash-based message authentication code (HMAC)-based key derivation function (HKDF) is used for pseudorandom key generation in task dictionary HMAC-SHA256 verification, reflecting original intention.

• Summary
• ubuntu.clang.cxx11thread.serialization.python38.PyRosetta.unit 「2」
• mpi.serialization
• integration.addsan 「1」
• integration.mpi 「2」
• integration.release_debug 「1」
• integration.tutorials 「1」
• maintenance.documentation
• release.PyRosetta.conda.Release 「2」
• scientific.protein_data_bank_diagnostic.cif
• scientific.protein_data_bank_diagnostic.full