feat(scripts): crash_groups.py lists the crash reports in named groups, the largest first #1288

Merged
HumanBean merged 1 commit from HumanBean/Longshot:feat/crash-groups into main 2026-11-17 00:53:46 +00:00
Owner

Type: feat · Risk: low · Tested on: Linux (a script; the game doesn't run it) · Needs: none

Why

55 open issues are titled "Longshot closed unexpectedly", and nothing says which of them are the same crash. Each one is read by hand, so we can't tell which crash to fix first or whether a fix reached the players who reported it. This is the first milestone (M0) of the renderer rework plan: every open crash report in a named group. See #81 (crash reports grouped by a signature) and #44.

What changed

python -I scripts/crash_groups.py reads the issues the crash window titled, their attached crash and unclean-end records, and the ends of their logs, and prints the reports in named groups, the largest first. It sends nothing.

  • scripts/crash_groups.py: a group's name is how the run ended (the record's reason and exception code, or the log's device lost or GPU hang), whose code the fault was in (the graphics driver, the renderer, the game's lifted code, a system library) and whether VRAM or Windows' commit was at its limit. None of the three changes between builds. Inside a group, reports sharing a build and module+offset are listed together; a fault in lifted code is listed by the game's own address of its caller, which is the same in every build. --markdown and --json for pasting and for other tools; --files groups local records and logs with no forge. Of a log over 2 MiB it fetches the first 64 KiB and the last 1 MiB.
  • tests/test_crash_groups.py: the group names for each kind of record and log end, that a name holds across builds, a device lost counted as that after an unclean end, an uncaught C++ exception at the commit limit counted as memory (as #1056 does), and that the driver modules are hos_crash_driver_module's (host/hos_crash_dialog.h).
  • docs/workflow.md: the script, beside hardware_report.py.

Measurements

Verdict: nothing to time (a script). Run on the 55 open crash reports on 2026-11-16, it puts them in 16 groups. Half are a lost GPU device (27 reports, 11 of them on Intel cards), 10 are access violations in the game's lifted code (two share a caller), and 9 ended at a memory limit: 6 with no crash record and VRAM or Windows' commit at its limit, 3 as an out-of-memory exception.

Group Reports By card
GPU device lost 21 Intel 11, NVIDIA 5, AMD 5
access violation in the game code 10 NVIDIA 6, Intel 2, unknown 1, AMD 1
GPU device lost, VRAM full 5 AMD 4, NVIDIA 1
ended with no crash record, VRAM full 3 AMD 3
ended with no crash record 2 AMD 1, Intel 1
ended with no crash record, commit limit, VRAM full 2 AMD 2
fatal error (hos_fatal) in the game code 2 AMD 2
nothing in the report says how it ended 2 Intel 1, NVIDIA 1
GPU device lost, commit limit 1 NVIDIA 1
GPU hung (no submission finished) 1 AMD 1
access violation in the graphics driver (igvk64.dll) 1 Intel 1
access violation in the graphics driver (nvoglv64.dll), VRAM full 1 NVIDIA 1
ended with no crash record, commit limit 1 Intel 1
out of memory (uncaught C++ exception at the commit limit) in the renderer 1 NVIDIA 1
out of memory (uncaught C++ exception at the commit limit) in the renderer, VRAM full 1 NVIDIA 1
out of memory (uncaught C++ exception at the commit limit) in the system library (KERNELBASE.dll) 1 AMD 1

Flags and settings

None

Testing

Tested: the script run against the forge's open issues on Linux, 2026-11-16 (the table above). Each of the 55 reports lands in a group; for two of them the group is "nothing in the report says how it ended".

Not tested: tests/test_crash_groups.py was not run on this PC (no checks are run here); the forge's checks run it. --files on a Windows path.

Must be tested:

  • None

Risks

What could break: nothing in the game. A group name can hold two different crashes that end the same way: every device lost is one group until the fault info in the logs is read too.

Turning it off: revert the commit

Compatibility: Nothing changes

Issues

See #81
See #44

Checklist

Everyone

  • One topic.
  • The title is a Conventional Commit (type(scope): description) of the status line's type.
  • Why gives the reason plainly and links the issues that fit (the tracker searched).
  • Measurements has a verdict, or one line saying why there is nothing to measure.
  • Testing names what was tested, what wasn't, and what must still be tested, where and how.
  • Rebased on current main.
  • python -I -B tests/run_all.py passes (any failure that also fails on main is named). Not run on this PC; left to the forge's checks.
  • No game files, dumps, build output or generated files.
  • third_party/ untouched (a new folder with a PROVENANCE.md instead).
  • The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled.

When it applies

  • AI assistance recorded with Assisted-by: / Co-Authored-By: trailers.
  • Docs updated.

🤖 Generated with Claude Code

**Type:** feat · **Risk:** low · **Tested on:** Linux (a script; the game doesn't run it) · **Needs:** none ## Why 55 open issues are titled "Longshot closed unexpectedly", and nothing says which of them are the same crash. Each one is read by hand, so we can't tell which crash to fix first or whether a fix reached the players who reported it. This is the first milestone (M0) of the [renderer rework plan](https://git.magicbean.dev/MagicBeanDev/Longshot/wiki/Renderer-rework-plan): every open crash report in a named group. See #81 (crash reports grouped by a signature) and #44. ## What changed `python -I scripts/crash_groups.py` reads the issues the crash window titled, their attached crash and unclean-end records, and the ends of their logs, and prints the reports in named groups, the largest first. It sends nothing. - `scripts/crash_groups.py`: a group's name is how the run ended (the record's reason and exception code, or the log's device lost or GPU hang), whose code the fault was in (the graphics driver, the renderer, the game's lifted code, a system library) and whether VRAM or Windows' commit was at its limit. None of the three changes between builds. Inside a group, reports sharing a build and module+offset are listed together; a fault in lifted code is listed by the game's own address of its caller, which is the same in every build. `--markdown` and `--json` for pasting and for other tools; `--files` groups local records and logs with no forge. Of a log over 2 MiB it fetches the first 64 KiB and the last 1 MiB. - `tests/test_crash_groups.py`: the group names for each kind of record and log end, that a name holds across builds, a device lost counted as that after an unclean end, an uncaught C++ exception at the commit limit counted as memory (as #1056 does), and that the driver modules are `hos_crash_driver_module`'s (`host/hos_crash_dialog.h`). - `docs/workflow.md`: the script, beside `hardware_report.py`. ## Measurements **Verdict:** nothing to time (a script). Run on the 55 open crash reports on 2026-11-16, it puts them in 16 groups. Half are a lost GPU device (27 reports, 11 of them on Intel cards), 10 are access violations in the game's lifted code (two share a caller), and 9 ended at a memory limit: 6 with no crash record and VRAM or Windows' commit at its limit, 3 as an out-of-memory exception. | Group | Reports | By card | |---|---:|---| | GPU device lost | 21 | Intel 11, NVIDIA 5, AMD 5 | | access violation in the game code | 10 | NVIDIA 6, Intel 2, unknown 1, AMD 1 | | GPU device lost, VRAM full | 5 | AMD 4, NVIDIA 1 | | ended with no crash record, VRAM full | 3 | AMD 3 | | ended with no crash record | 2 | AMD 1, Intel 1 | | ended with no crash record, commit limit, VRAM full | 2 | AMD 2 | | fatal error (hos_fatal) in the game code | 2 | AMD 2 | | nothing in the report says how it ended | 2 | Intel 1, NVIDIA 1 | | GPU device lost, commit limit | 1 | NVIDIA 1 | | GPU hung (no submission finished) | 1 | AMD 1 | | access violation in the graphics driver (igvk64.dll) | 1 | Intel 1 | | access violation in the graphics driver (nvoglv64.dll), VRAM full | 1 | NVIDIA 1 | | ended with no crash record, commit limit | 1 | Intel 1 | | out of memory (uncaught C++ exception at the commit limit) in the renderer | 1 | NVIDIA 1 | | out of memory (uncaught C++ exception at the commit limit) in the renderer, VRAM full | 1 | NVIDIA 1 | | out of memory (uncaught C++ exception at the commit limit) in the system library (KERNELBASE.dll) | 1 | AMD 1 | ### Flags and settings None ## Testing **Tested:** the script run against the forge's open issues on Linux, 2026-11-16 (the table above). Each of the 55 reports lands in a group; for two of them the group is "nothing in the report says how it ended". **Not tested:** `tests/test_crash_groups.py` was not run on this PC (no checks are run here); the forge's checks run it. `--files` on a Windows path. **Must be tested:** - None ## Risks **What could break:** nothing in the game. A group name can hold two different crashes that end the same way: every device lost is one group until the fault info in the logs is read too. **Turning it off:** revert the commit **Compatibility:** Nothing changes ## Issues See #81 See #44 ## Checklist ### Everyone - [x] One topic. - [x] The title is a Conventional Commit (`type(scope): description`) of the status line's type. - [x] Why gives the reason plainly and links the issues that fit (the tracker searched). - [x] Measurements has a verdict, or one line saying why there is nothing to measure. - [x] Testing names what was tested, what wasn't, and what must still be tested, where and how. - [x] Rebased on current `main`. - [ ] `python -I -B tests/run_all.py` passes (any failure that also fails on `main` is named). Not run on this PC; left to the forge's checks. - [x] No game files, dumps, build output or generated files. - [x] `third_party/` untouched (a new folder with a `PROVENANCE.md` instead). - [x] The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled. ### When it applies - [x] AI assistance recorded with `Assisted-by:` / `Co-Authored-By:` trailers. - [x] Docs updated. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
feat(scripts): crash_groups.py lists the crash reports in named groups, the largest first
Some checks failed
CI / Python source checks (ubuntu-24.04, part 1/2) (pull_request) Has been skipped
CI / Python source checks (ubuntu-24.04, part 2/2) (pull_request) Has been skipped
CI / SPIR-V fixtures (global narrow load/store, direct surface atomics, F2I immediate) (pull_request) Has been skipped
Third-party parts / The same as upstream (pull_request) Has been skipped
Website / Website on Cloudflare (pull_request) Has been skipped
CI / Conventional Commit title (pull_request) Successful in 7s
Guide screenshots / Guide images and local UI source review (pull_request) Failing after 10s
CI / Workflow lint (actionlint) (pull_request) Successful in 11s
Third-party parts / As their notes say (pull_request) Successful in 14s
Website / Website on Cloudflare (pull_request_target) Has been skipped
PR description / Description follows the template (pull_request_target) Failing after 4s
f2a06ee7c5
The renderer rework plan's M0 (https://git.magicbean.dev/MagicBeanDev/Longshot/wiki/Renderer-rework-plan): every open
crash report in a named group. A group is named by how the run ended, whose code the fault was in and whether VRAM or
Windows' commit was at its limit; none of the three changes between builds. Reads the issues the crash window titled,
their attached records and the ends of their logs; sends nothing.

Co-Authored-By: Claude Fable 5.2 <[email protected]>
Collaborator

Possibly related, for context:

  • PR #141 feat(crash): report a crash automatically from the crash window, or manually with every field — Crash reporting PR; same crash-report workflow and issue #81.
  • issue #1266 Idea: keep the field data usable once players send it: label limits that don't swallow releases, counters that survive a restart, alerts on failed pushes, and a longer retention — Open proposal on crash-report signature and telemetry data; related operational context.
Possibly related, for context: - PR #141 feat(crash): report a crash automatically from the crash window, or manually with every field — Crash reporting PR; same crash-report workflow and issue #81. - issue #1266 Idea: keep the field data usable once players send it: label limits that don't swallow releases, counters that survive a restart, alerts on failed pushes, and a longer retention — Open proposal on crash-report signature and telemetry data; related operational context.
HumanBean changed title from WIP: feat(scripts): crash_groups.py lists the crash reports in named groups, the largest first to feat(scripts): crash_groups.py lists the crash reports in named groups, the largest first 2026-11-17 00:53:42 +00:00
HumanBean deleted branch feat/crash-groups 2026-11-17 00:53:46 +00:00
Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
2 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
MagicBeanDev/Longshot!1288
No description provided.