Test First Contact

This guide tests the complete First Contact checkpoint across ordinary play, reload, streaming, cleanup, replay, and removal. The eleven cases are deliberately separate: a scene working once does not show that its actor, journal state, audio, or cleanup will survive every other lifecycle boundary.

completed checkpoint
  + known starting saves
  + one controlled route per case
  + visible results and fresh logs
  = a result you can diagnose and reproduce

Use the unmodified completed checkpoint. A project edited while following the authoring chapter is useful for personal testing, but it is a different build and may fail for different reasons.

Required environment

Record the versions you actually use. The WEM is already built; testing the lab does not require Wwise Console or audio conversion.

Freeze a minimal mod loadout

Run Cases 1–10 with the required frameworks and the completed Lab 5 pair. Disable unrelated quest, world, community, AI, appearance, and framework mods where practical. In particular, another mod can change the generic Tyger record’s hostility, appearance, workspot behavior, or acquisition without changing any cqa005 file.

Keep the loadout unchanged between cases. Case 11 is the exception: close the game and remove the Lab 5 pair so you can verify that the completed save does not depend on the mod remaining mounted.

Establish the site before installation

The shared center is the outdoor recycling-station cabinet row at Allen Street in Watson/Kabuki:

(-1000.02, 1497.2208, 8.3)

Before any Lab 5 candidate is installed, follow Lab 3’s ordinary control-route procedure. Confirm the route using the vanilla landmark rather than the Lab 5 pin that is under test. Do not replace an inaccessible route with a console teleport.

The two boundaries relevant to this campaign are:

The radii describe the supplied polygonal prisms, not universal safe values. Record the actual route and first/last visible behavior instead of inferring a crossing from the numbers alone.

Verify the frozen spoken-line lookup

Before building, inspect the completed checkpoint and verify this exact lookup chain:

scene locstring RUID 9638591835734011695
  -> registered subtitles\cqa005_subtitles_map.json
       -> indirect subtitles\cqa005_subtitles.json
            femaleVariant = maleVariant = "All clear. Keep moving."
  -> registered vo\cqa005_vo.json
       femaleResPath = maleResPath
         -> vo\contact_i_85c3283507e7ef2f.wem

The complete depot paths are:

mod\cqa\cqa005\localization\en-us\onscreens\cqa005_onscreens.json
mod\cqa\cqa005\localization\en-us\subtitles\cqa005_subtitles_map.json
mod\cqa\cqa005\localization\en-us\subtitles\cqa005_subtitles.json
mod\cqa\cqa005\localization\en-us\vo\cqa005_vo.json
mod\cqa\cqa005\localization\en-us\vo\contact_i_85c3283507e7ef2f.wem

The subtitle map and VO map are registration roots. The subtitle entries and WEM are archived indirect dependencies. Renaming one file or registering the entries directly changes the candidate.

Use the supplied WEM unchanged for these tests.

Prepare test saves

Create and preserve two manual-save originals before the first Lab 5 installation. Never use an autosave or checkpoint save, and keep both source slots outside the game’s rotating autosave set.

1. Make CQA005 ORIGINAL OUTSIDE SETUP outside the setup volume on the verified ordinary approach.
2. Make the separate CQA005 ORIGINAL NEAR SETUP near but still outside setup for the immediate-arrival race. Reach it through ordinary game travel, including the verified vanilla fast-travel approach if used; do not teleport.
3. Record each complete slot-directory name in your test notes.
4. Confirm neither save has ever loaded any CQA tutorial candidate: no Lab 1–4 start/completed project and no Lab 5 start, completed, or older cqa005 candidate. Labs 3 and 4 share this site, and removing their archives does not erase saved quest, journal, trigger, or world state. If you already followed those labs, restore a separately preserved pre-tutorial save or create a genuinely new clean lineage; a disclosed prior-lab history is not sufficient.
5. Exit the game completely before cloning either complete source slot into a run-specific directory. Never execute a case from the immutable source directory itself.
6. Case 1 must continue one unbroken ordinary route through completion and create exactly three additional manual-save captures at the frozen moments below. A checkpoint or autosave does not satisfy a capture row.
7. After Case 1, exit the game completely before cloning any complete seed slot for Cases 3–9 or 11. Never edit a clone, reset a fact, or copy only sav.dat without the rest of its slot directory.

The test uses five named save captures:

Keep all five source slots unchanged. With the game closed, copy the entire appropriate slot into a separate working directory for each case, including its metadata files. Check that its sav.dat matches the source copy. Never run a case directly from the preserved source slot.

Cases 3, 4, and 7 use copies of the same pre-scene capture outside setup. There is no save window between child checkpoint node 15 and scene node 16: they are connected directly. Cases 5, 6, and 8 share the post-contact capture; Cases 9 and 11 share the completed capture; Cases 1 and 10 share the outside original. Case 2 alone uses the separate near-site original.

Do not reset cqa005_completed to prepare a case. Journal, scene, community, actor, phase, and streamed-world state can outlive that fact.

Build, install, and bind the canonical candidate

Use the unmodified completed checkpoint:

project: CQA_Lab05_FirstContact.cpmodproj

Open all eleven CR2W resources in WolvenKit before testing. Resolve missing files or conversion errors before installing the project.

1. Close the game.

2. Open the completed project in WolvenKit and use its normal build/package and install workflow.

3. Confirm the start checkpoint, every older cqa005 archive, and every Lab 1–4 tutorial candidate are absent. Labs 3 and 4 deliberately reuse this Allen Street test area; leaving them installed can add overlapping markers, triggers, objectives, and log traffic even though their depot namespaces differ.

4. From the game directory, run:

   Get-ChildItem .\archive\pc\mod\CQA_Lab0*.archive* |
     Select-Object -ExpandProperty Name
   

   The complete output must be exactly these two names, with no Start variant and no CQA_Lab01_* through CQA_Lab04_* pair:

   CQA_Lab05_FirstContact.archive
   CQA_Lab05_FirstContact.archive.xl
   

5. Confirm the candidate contains the twelve depot paths listed in the resource inventory, including the indirectly reached child, scene, subtitle entries, WEM, and both sectors.

6. Preserve byte-identical copies of both installed files until the complete campaign and log audit finish.

Keep the installed pair unchanged between cases. If you rebuild it, repeat the affected tests from their starting saves.

Capture every execution consistently

For each run, keep a dated note containing your game/framework versions, installed filenames, starting save label and slot directory, the source save it was copied from, and what you observed. Mark cases you did not run as pending. Keep a fresh copy of all four logs listed below before the next launch can overwrite them.

Use a separate working slot and log folder for every case. Reusing one run’s logs or observations cannot establish the result of another run.

Capture a short continuous video with game audio for timing, replay, visible spawn, scene, and cleanup claims. Add focused screenshots or notes for journal and pin states. Record the first and last visible contact behavior, the route, any combat stimulus, and whether a loading screen occurred. Do not use the video alone for log or save history.

Keep saves, logs, and captures in your own test folder. If you contribute a result to the book, follow the repository contribution guide for sanitizing and submitting evidence. Do not publish private saves or extracted vanilla resources.

Case 1: clean-ordinary-passive-spawn

Acceptance precondition: Install the canonical candidate, load a closed-game full-slot clone of the untouched outside-setup original, approach by ordinary movement, and continue the same unbroken route through ordinary completion while making the three named manual seed saves.

Required result: The root and child activate once, the passive community contact appears through the community/AI-spot mapping, and the meet objective and pin appear once without a spawn or NodeRef error. The ordinary route completes once, and exactly the pre-scene, post-contact, and completed seed captures are made at their frozen observable states.

Walk the verified route without fast travel, teleporting, combat, or waiting at the boundary to manipulate streaming. Begin capture before the first Lab 5 objective. Record the meet objective and pin appearing once, the contact at the authored spot, and the absence of hostile/combat behavior when unprovoked.

Continue this same run and make exactly these three manual saves, never a checkpoint or autosave:

1. CQA005 SEED PRE SCENE OUTSIDE SETUP, after the contact is spawned and passive but while V remains outside setup. Meet and pin are Active, the child is active, checkpoint cqa005_first_contact has not executed, the scene has not started, and cqa005_completed is 0.
2. CQA005 SEED POST CONTACT INSIDE CLEANUP, after contact_done returns but before leaving cleanup. Meet is Succeeded, pin Inactive, Leave Active, child/contact active, and the fact remains 0.
3. CQA005 SEED COMPLETED, after cleanup and root return. Leave and quest are Succeeded, pin Inactive, contact deactivated, child/root returned, and the fact is 1.

Keep the route capture, state screenshots, and fresh logs. Exit the game before copying the three complete seed directories. There is no save window between checkpoint node 15 and scene node 16.

Case 2: fast-arrival-race

Acceptance precondition: Load a closed-game full-slot clone of the separate untouched near-setup original and enter immediately before waiting for streaming or spawn.

Required result: The spawned-character wait resolves before the setup wait advances; the scene begins once only after the contact exists, with no missed trigger or deadlock.

This is the fast-travel/immediate-sprint setup-race case. Use the independently preserved near-site save, load it, and sprint into the 25 m setup volume as soon as player control returns. Do not pause nearby to let the actor appear first. If fast travel was used to prepare the pre-install route, record that fact and the vanilla destination; never replace it with a console teleport.

Capture from the loading screen through contact appearance and scene start. The decisive ordering is actor visible/available before the scene performs, one scene start, and eventual forward progress. A scene that never starts, a scene that starts without its contact, or a missed already-true setup condition fails the case even if a slower retry works.

Case 3: slow-contact-av-once

Acceptance precondition: Load a byte-identical full-slot clone of the named pre-scene outside-setup seed, approach slowly, remain in range for the full scene, and do not skip or interrupt it.

Required result: The exact subtitle All clear. Keep moving. and the supplied WEM play once; the contact is acquired, the line does not duplicate, and lipsync/audio errors are absent from retained logs.

Start from clean-derived-before-setup, a distinct execution-slot clone of seed-pre-scene-outside-setup. Keep the contact in view, enter at a walking pace, and remain until the scene completes. Capture the exact subtitle, audible line, contact performer, and mouth/facial response in one continuous recording. Count subtitle display and audible playback separately; each must be exactly one.

Check the line with RUID 9638591835734011695. A subtitle without sound, sound without the exact subtitle, or duplicate playback fails this test.

Case 4: named-exit-only

Acceptance precondition: Load a byte-identical full-slot clone of the same named pre-scene outside-setup seed used by Cases 3 and 7, then observe ordinary completion. This is not a save between checkpoint node 15 and scene node 16.

Required result: Only named exit contact_done advances the child to meet Succeeded; Default INT/RET remain unwired and no fallback continuation is observed.

Start from clean-derived-before-scene, another distinct execution-slot clone of seed-pre-scene-outside-setup, not a new capture. Record the meet objective before crossing setup, the scene’s ordinary terminating end, and the transition to meet Succeeded/leave Active. Do not invoke an interrupt or cut path in this case.

Pair the runtime capture with the canonical scene/questphase structural record showing contact_done wired to child node 17.Succeeded and Default INT / Default RET unwired. Runtime evidence must show no fallback or double continuation; graph inspection alone cannot pass the case.

Case 5: no-replay-inside-lifetime

Acceptance precondition: Load a byte-identical full-slot clone of the named post-contact inside-cleanup seed, then cross the setup boundary or revisit the contact point before leaving cleanup.

Required result: The scene, subtitle, and WEM do not replay during the same child lifetime; the leave objective remains the only active progression state.

Use clean-derived-after-contact-inside-cleanup, a distinct execution-slot clone of seed-post-contact-inside-cleanup. Stay inside the 110 m cleanup volume, and walk out of and back into the 25 m setup area or revisit the contact point without crossing cleanup. Remain long enough that a duplicate line would be obvious.

Capture the complete revisit and the journal state. The contact may remain as part of delayed cleanup, but the scene, subtitle, and WEM must not restart and the meet objective/pin must not reactivate.

Case 6: post-scene-reload

Acceptance precondition: Without changing the installation, load a byte-identical full-slot clone of the named post-contact inside-cleanup seed.

Required result: Meet stays succeeded, the pin stays inactive, leave stays active, the contact line does not replay, and crossing cleanup continues once.

Hash and load the contact-done-inside-cleanup execution clone, verifying it matches seed-post-contact-inside-cleanup, with the exact installed pair unchanged. Capture journal and pin state immediately after load and after walking continuously outside cleanup. Record that the line does not replay and that leave/child progression occurs once.

A run that repairs itself only after re-entering setup is a failure: the saved post-scene owner state must resume directly.

Case 7: stream-away-return

Acceptance precondition: Load a byte-identical full-slot clone of the named pre-scene outside-setup seed, travel by ordinary movement beyond the finite Quest descriptor box before entering setup, return, and finish.

Required result: The community and child remain coherent across streaming; returning resolves the same NodeRefs and advances the scene exactly once without duplicate contact entries.

Start from child-active-before-setup, a third distinct execution-slot clone of seed-pre-scene-outside-setup. Do not enter the setup volume. Travel by ordinary movement beyond the Quest descriptor’s finite box, whose supplied X/Y range is -1300.02..-700.02 by 1197.2208..1797.2208, then return by the same recorded route. Do not use fast travel for this case.

Capture departure, return, contact materialization, scene, and objective progression. Retain logs from both streaming transitions. Two contacts, a stale spot, unresolved NodeRefs, lost child state, or a second scene execution fails the case.

Case 8: cleanup-boundary-no-pop

Acceptance precondition: Load a byte-identical full-slot clone of the named post-contact inside-cleanup seed, then walk outward continuously through the cleanup boundary more than 110 metres from the shared center.

Required result: IsOutside resolves once, whole-community deactivation occurs only at cleanup, no visible contact pop occurs inside the cleanup volume, leave succeeds, and child Out1 returns once.

Load a third distinct contact-done-inside-cleanup execution clone and confirm it matches seed-post-contact-inside-cleanup. Walk outward continuously; do not fast travel, teleport, reload again, or double back. Keep the meeting point in view when the route permits and record distance/route landmarks, the last visible contact state, the leave-objective transition, and quest completion.

The actor must not visibly disappear while V is still inside the authored cleanup prism. The expected deactivation command and return are proved by the ordered progression and absence of errors, not by claiming that Spawn Manager guarantees same-frame visual removal.

Case 9: completed-reload

Acceptance precondition: With the same exact candidate installed, load a byte-identical full-slot clone of the named completed seed.

Required result: cqa005_completed bypasses the child; objectives, pin, community, scene, subtitle, and WEM do not reactivate or replay.

Load the completed execution clone, confirm it matches seed-completed, and remain near or revisit the site long enough to observe the one-shot guard. Record the absence of all six reactivation surfaces named in the required result. A silent scene with a reactivated objective or actor still fails the case.

Case 10: untouched-replay

Acceptance precondition: With the canonical candidate still installed, load a byte-identical full-slot clone of the same untouched outside-setup original used by Case 1 and repeat the ordinary route.

Required result: Spawn, acquisition, line, named exit, cleanup, child return, and root completion reproduce once with the same visible and logged behavior.

Load a working copy of the original untouched-preinstall-outside-setup slot and repeat the complete route with the same build. Compare spawning, dialogue, cleanup, and completion with Case 1.

This is clean reproducibility. Resetting cqa005_completed on the completed save would test a different, dirty state and cannot pass this case.

Case 11: removal-isolation-diagnostic

Acceptance precondition: With the game closed, clone the named completed seed into the disposable execution slot, remove both exact candidate files before launching and loading that clone, and revisit the site without resetting any fact.

Required result: No mod-owned resource remains mounted, no Lab 5 objective or pin reactivates, no cqa005 contact appears, fresh logs show no cqa005 registration, and the packaged fixture contains no TweakDB mutation or override.

With the game closed, copy the complete seed-completed source slot to a separate disposable slot. Preserve the installed pair, then move these two files out of the game directory:

archive\pc\mod\CQA_Lab05_FirstContact.archive
archive\pc\mod\CQA_Lab05_FirstContact.archive.xl

Confirm there is no second Lab 5 archive or ArchiveXL file under another name. Do not edit the clone, reset cqa005_completed, or use a save editor or console. Archive the prior four logs, launch the disposable clone, revisit the site, and then retain the fresh four-log bundle.

Confirm that the fresh ArchiveXL log contains no registration for mod\cqa\cqa005, no Lab 5 objective or pin reactivates, and no mod-owned contact appears. Completed journal history already stored in the save is not a mount or reactivation and is not grounds for failure by itself.

The exact packaged inventory contains no r6\tweaks, YAML tweak, script, or other TweakDB mutation source. Its Character.jpn_tyger_claws_biker2_ranged2_copperhead_wa value is a read-only characterRecordId reference. That source-level absence is structurally validated; this diagnostic deliberately does not claim to prove every vanilla field or exclude unrelated installed mods. Disable unrelated mods rather than trying to edit or compare the generic record.

Label every artifact removal-isolation-diagnostic. Restore the unchanged candidate pair only after the game and framework processes have fully exited.

Retain and audit four logs per execution

Before starting another run, copy these logs into that run’s test folder:

red4ext\plugins\ArchiveXL\ArchiveXL.log
red4ext\logs\red4ext.log
red4ext\logs\game.log
r6\logs\redscript_rCURRENT.log

Copy the logs before the next launch overwrites them. Keep them in a separate folder for each run.

Search all four logs for:

• cqa005, all twelve depot paths, and the installed project name;
• #cqa005_pr_first_contact, the community, spot, scene marker, and trigger NodeRefs;
• RUID 9638591835734011695 and contact_i_85c3283507e7ef2f.wem where the subsystem logs expose them;
• ArchiveXL registration, missing resource, NodeRef, streaming, scene, community, localization, audio, RED4ext, and redscript errors.

Record both relevant matches and the absence of expected failures. Do not claim that silence proves A/V playback: Case 3 still requires direct subtitle and audio observation.

Failure triage

Preserve the failed run, save name, candidate pair, video, and all four logs before editing anything.

Lab sequence: All labs · Previous: Author First Contact in WolvenKit · Next: Troubleshooting.