No reviewers
Labels
No labels
area/accessibility
area/audio
area/input
area/install
area/menu
area/renderer
area/saves
area/tooling
area/upscaling
beanstalk: approved
beanstalk: failed
beanstalk: needs approval
beanstalk: running
beanstalk: tested
ci: approved
good first issue
gpu/amd
gpu/igpu
gpu/intel
gpu/nvidia
help wanted
kind
bug
kind
chore
kind
docs
kind
feature
kind
performance
kind
question
needs measurements
platform/linux
platform/macos
platform/steam-deck
platform/windows
preview
priority
critical
priority
high
priority
low
regression
risk
high
risk
low
risk
medium
status
blocked
status
confirmed
status
duplicate
status
invalid
status
needs-info
status
wontfix
ci: approved
No milestone
No project
No assignees
4 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
MagicBeanDev/Longshot!1378
Loading…
Reference in a new issue
No description provided.
Delete branch "GummiNaomi/Longshot:fix/tested-one-row-a-file"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Type: fix · Risk: low · Tested on: Windows 11 (the checks; no game needed) · Needs: nothing
Why
My mistake, in #1255 and #1256: four of the files they added under
docs/tested/hold several rows (two each for the dead targets' runs, six each for the shadow-map slots' runs). A fragment is the header, the separator and exactly one row (docs/tested/README.md, #96), so since they were mergedscripts/tested_configurations.pystops withand
tests/test_tested_configurations.pyfails on main for everyone. The combined catalogue is rendered by the same script. I did not see it because that test already failed on my PC for another reason (its pipe check, with Windows line endings), and I read the failure as the old one.How to reproduce
Seen on: any OS, no GPU involved; main @
544229a2Settings and switches: defaults
Steps:
python -I -B tests/test_tested_configurations.py, orpython -I scripts/tested_configurations.pyExpected: the test passes and the script prints the catalogue
Actual: the
ValueErrorabove, fromread_rowWhat changed
Each row of those four files is a file of its own; no row's text is changed.
2026-11-16-dead-targets-every-present-gumminaomi-old.mdand-new.mdbecome-old1.md,-old2.md,-new1.md,-new2.md;2026-11-16-shadow-map-slots-gumminaomi-off.mdand-on.mdbecome-off1.mdto-off6.mdand-on1.mdto-on6.md. Each row already names its run (prdn-...,prsl-...), so the rows stay distinguishable.Measurements
Nothing to measure: sixteen rows moved to sixteen files.
Flags and settings
None.
Testing
Tested: Windows 11:
python -I scripts/tested_configurations.pyprints the catalogue with this change and raises theValueErrorabove on main @544229a2.python -I -B tests/test_tested_configurations.py: 24 of 25 pass with this change (23 on main); the one left,test_cli_ascii_pipe_emits_the_arrow_containing_source_as_utf8, fails on this PC on main too (the script's output through a pipe has Windows line endings here).python -I -B tests/run_all.pywith clang here, this PR @c6189282: 393 passed, 107 failed, 221 skipped; main @544229a2in the same session: 392 passed, 108 failed, 221 skipped. Nothing fails here that does not fail on main (this PC's toolchain: no gcc, Windows line endings, fixtures needing tools not installed); the one more on main wastests/test_gpu_hang_watch.pyin that run.Not tested: Linux, where the remaining check should pass too; the website's build.
Must be tested:
python -I -B tests/test_tested_configurations.py, expecting all 25 to pass (by hand, or the pull request's own checks: look at that test's line)Risks
What could break: nothing I can name: no row's text changes, only which file holds it.
Turning it off: revert the commit.
Compatibility: Nothing changes. Links to the four old file names, if any exist outside the repository, stop working.
Issues
None found for this. #1255 and #1256 added the files; #96 is the rule.
Checklist
Everyone
type(scope): description) of the status line's type.main.python -I -B tests/run_all.pypasses (any failure that also fails onmainis named).third_party/untouched (a new folder with aPROVENANCE.mdinstead).When it applies
Assisted-by:/Co-Authored-By:trailers.docs/tested/for each new measured run (format indocs/tested-configurations.md).🤖 Generated with Claude Code
Possibly related, for context:
Each of the sixteen measured rows is now its own file, and the row text is unchanged. I read head
c6189282. It applies onto current main. This can merge as it is.docs/tested/allows a header, the separator and one row. The four files from #1255 and #1256 held two or six rows, so the catalogue script stopped on main. The sixteen files are that shape, in LF, and each name carries the row's date. Nothing else in the tree still names the four old files. The Windows pipe check that still fails on this PC is the one that already fails on main, and this change does not touch it.Linux check for the item in “Must be tested”, at exact head
c6189282dcff4254764ad9b0c3cfe0ed75854a17:python -I -B tests/test_tested_configurations.py: 25 tests, OK, exit 0.544229a260c5f5008f241b9cf48ae6b01c1602b9, with the same interpreter/environment policy: 25 tests, one error, exit 1. The error is the committed multi-row fragment rejected byread_row, not a parser defect.scripts/tested_configurations.pyCLI, both stdout and--output, exits 0 at this head; their UTF-8 output bytes agree.This is a focused source-only replay pinned to the public Git tree/blobs, including the complete tested-fragment corpus, frozen source page, script and test. Those inputs remained unchanged during execution. It is not a full checkout/full-suite run, website build, Windows test, game run or validation of the historical performance measurements. No branch or source changes made.