jpmschweitzer/settled-reach
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(client): T-1153 — clamp _held_n at the viewer's single write site (live eyeball round blocker)

Browse Source 
The live round against a real server caught what 3518 unit tests missed:
enter_orbital stored the UNCLAMPED n (Lendel: 19139 raw circumference
cols) into _held_n while AtlasWindowRequest clamped independently to
6400 before sending — every orbital response on a real-sized body was
dropped by the viewer's own w_n != _held_n staleness check. Permanently
blank ladder. The PR #191 C1 clamp-mirror lesson, one layer up: the
mirror must apply wherever held/expected state is recorded, not only at
the request layer.

_enter_at_rung now clamps via _clamp_window_n_mirror_v2 before storing
(the file's only _held_n write site — reselect/refloat read and forward).
Regression closes the coverage hole: the oversized-orbital RESPONSE
round-trip (enter_orbital on GJ380c -> deliver echo n=6400 Region ->
assert accepted). Both new tests verified to fail against the reverted
fix before trusting green. Full suite 3520/3520.
This commit is contained in:
jpmschweitzer
2026-07-22 11:54:06 +02:00
parent 55a2909386
commit 1c87b6a712
2 changed files with 98 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
client/ui/implant/apps/atlas/atlas_window_viewer.gd
+20 -2
View File
@@ -330,6 +330,21 @@ func enter_orbital(body: Dictionary, system: Dictionary) -> void:
## no canonicalization; enter()'s does its own before calling in). Resets
## every piece of held/request state for a fresh descent, exactly as the
## pre-T-1153 enter() always did, plus the new _held_granularity_v2 tracking.
##
## **The C1 clamp-mirror lesson, one layer up (live-round finding):** `n`
## MUST be clamped via `_clamp_window_n_mirror_v2()` BEFORE it becomes
## `_held_n` — mirroring exactly what AtlasWindowRequest.request_now()
## already does to ITS OWN `_n` before storing/sending (see that function's
## own doc for the original PR #191 Tyre C1 finding). Storing the RAW `n`
## here (e.g. enter_orbital()'s full district_extent().cols, routinely tens
## of thousands at Region granularity, versus the server's clamped echo of
## at most DISTRICT_WINDOW_MAX_N_REGION=6,400) left `_held_n` permanently
## disagreeing with what the server would ever actually echo — every
## orbital-rung response was silently rejected as stale by
## _on_window_ready()'s `w_n != _held_n` check, hanging the ladder on every
## real-sized body. `_maybe_reselect_rung()`/`_maybe_refloat_window()` both
## read `_held_n` (never re-derive it), so clamping here — the ONE write
## site — fixes every downstream caller too, not just entry.
func _enter_at_rung(
body: Dictionary,
system: Dictionary,
@@ -337,17 +352,20 @@ func _enter_at_rung(
n: int,
granularity_v2: String
) -> void:
var clamped_n: int = AtlasWindowRequest._clamp_window_n_mirror_v2(n, granularity_v2)
_body = body
_system = system
_held_center = district_center
_held_n = n
_held_n = clamped_n
_held_granularity_v2 = granularity_v2
_window = null
_user_adjusted = false
_awaiting_first_window = true
_fit_and_center()
_window_request.reset()
_window_request.request_now(_dict_str(_body, "body_id", ""), _held_center, n, granularity_v2)
_window_request.request_now(
_dict_str(_body, "body_id", ""), _held_center, clamped_n, granularity_v2
)
_refresh_screen_header()
grab_focus()
queue_redraw()