[MarcZX10]

What

TraceFile::ReplaceGzip writes the gzip to <name>.tmp and renames it over the benchmark file. The write sat outside any handler, so a write that failed half way (a full disk, for example) threw out of ReplaceGzip and left the partial temporary file behind; only a failed rename removed it. FlushInterval calls it every interval, so the stale file stays until a later write truncates it.

The write is now guarded: the temporary file is removed (unless something else made that name a directory) and the exception is rethrown.

Two commits: the fix, then the test (ReplaceGzipFile in PerformanceTrace.cpp: on Linux, <name>.tmp is a symlink to /dev/full, which accepts the open and fails the write with ENOSPC; ReplaceGzip has to throw and leave no temporary file).

Tested

Linux x86-64, a throwaway archive build with the host CPU gate patched. agc_performance_trace passes. With the guard removed the same test fails with "a replace whose write failed left its temporary file behind". The case is skipped when /dev/full does not exist and does not run on Windows. tools/check_conventions.py --base upstream/main and git diff --check pass.

Checklist

• Based on current main; no other open PR implements the same functions
• One topic per PR; follow-ups go in a new PR
• No comments in code except technical debt
• Unimplemented paths throw (NotImplemented_nid_no_patch); silent stubs are listed under Silent stubs in their library's technical debt file
• No notes, investigation .md files or images added to the repository (attach them to this PR)
• New third-party code is a submodule built from source
• Generic behaviour, not specific to one title
• Depends on: none
• Fixes: none
• Unblocks: none
• GPUs tested: n/a
• AI-assisted: yes

[maskym]

Windows + NVIDIA result for head 5e96d65b83, merged into main 3b979d42a9 (the PR's GPUs tested line names no NVIDIA run on Windows).

This PR's own tests: agc_performance_trace (0.16 s), all passed.

Host: Windows 11 26300, Intel i7-12700K, NVIDIA GeForce RTX 2070 SUPER (vendor 0x10de, device 0x1e84; Vulkan loader 1.4.357 from the NVIDIA driver 621.01), Khronos Validation Layers nightly 1.4.366 (b2b44c0d37) through VK_ADD_LAYER_PATH, WinLibs GCC 15.2.0 (x86_64-ucrt-posix-seh, r6), Release, BUILD_TESTING=ON, Ninja. Build clean.

Same focused run on main 3b979d42a9 and on this tree, same machine and GPU: VK_INSTANCE_LAYERS=VK_LAYER_KHRONOS_validation, ANYPS5_NO_SHADER_CACHE=1, ANYPS5_REQUIRE_VULKAN=1, VK_LOADER_DEBUG=error,warn,layer, ctest --test-dir build -R '^(agc|guest_video_out|video_out)' --parallel 4 -V.

• main: 1143/1144 passed, the failure being agc_driver_mesh_barycentric_emulation (AGC: agc_driver_mesh_barycentric_emulation fails since #9193 on devices with mesh shaders and VK_KHR_fragment_shader_barycentric #12121, fixed by the open test(agc): disable barycentric built-ins in the mesh barycentric emulation test #12062).
• This tree: 1143/1144 with the same single failure; every other pre-existing test gives the same result. The loader inserted VK_LAYER_KHRONOS_validation in every test (1011 insertions in each log), neither log contains was not found but was requested by env var VK_INSTANCE_LAYERS, and both report exactly one VUID, VUID-VkPipelineColorBlendStateCreateInfo-pAttachments-00605 18 times (pre-existing on main, AGC: pipelines use differing per-attachment blend states while independentBlend is never enabled (VUID-VkPipelineColorBlendStateCreateInfo-pAttachments-00605) #11990 / fix(agc): enable independentBlend for color targets with different blend states #6782): this PR adds no new VUID- error.

The tree also carried the open PRs #12298, #12272, #12262, #12256, #12250, #12228, #12211, #12210, #12209, #12208, #12207, all merged cleanly, so the run is a joint one. Not run on a PS5.

AI-assisted: yes (Claude Code), the runs and the numbers above are from this machine.