zzfy — 郑州风云3 (Zhengzhou Fengyun 3)

Archive: zzfy (full).rar. Port: 40025. Status: done (boots clean, full registration flow verified end-to-end including a real Chinese name).

What this is

Same 风云 (Fengyun) lineage as fy2/fy2005/fengyun434 (archives #14, #16, #17) — adm/obj/master.lpc/adm/obj/simul_efun.lpc layout, same adm/daemons/chinesed.lpc + adm/simul_efun/chinese.lpc shape. Mudlib root at zzfy/fy3/, config at zzfy/config.fy3. readme.txt says this build ("郑州风云3") is provided by "长孙无情", described as one of the more stable/established 风云-family sites, admin access via ID xgchen. ~10,293 raw files.

Registration flow here is architecturally simpler than the TMI-2/nitan family processed earlier: there's no separate new keyword branch in get_id() — every English id typed goes straight through check_legal_id(), and if no save file exists for it, straight to the "create new character?" confirmation. "new" is itself in the banned_id list (line 20 of logind.lpc), so typing literal new is correctly rejected as an illegal id, not a special command — don't mistake this for a bug when retesting.

Fixes applied

AGENTS.md §15h (GBK byte-range Chinese detection), applied proactively before first boot, same shape as the other fengyun-family libs:

• adm/simul_efun/chinese.lpc's is_chinese(): GBK lead-byte range check (str[0] > 160 && str[0] < 255) → CJK Unicode codepoint range check (str[0] >= 0x4e00 && str[0] <= 0x9fff).
• adm/daemons/logind.lpc's check_legal_name(): byte-count length bound (< 2 || > 12, "1 到 6 个中文字") halved to character count (< 1 || > 6); removed the i%2==0 even-byte-offset gate so every character position is actually checked (not just every other one).

§4 (lazy security-daemon load): not needed — master.lpc's valid_write/valid_read already gate load_object(SECURITY_D) behind find_object() in this build.

Interactive test result — full registration flow

Verified the complete registration path in one continuous connection:

1. Connect → banner → English-id prompt. 2. zzfytest → passes check_legal_id, reaches "使用 zzfytest 这个名字将会创造一个新的人物，您确定吗(y/n)？". 3. y → reaches the Chinese-name prompt. 4. Real Chinese name 萧峰 → accepted by the fixed check_legal_name (no rejection message), proceeds straight to "请设定您的密码：" — the actual proof the §15h fix works, not just that the prompt renders.

Noted, not a bug: get_name() has a pre-existing printf("%O\n", ob); debug line (prints the login object's default toString, e.g. /obj/login#0) that fires right before the password prompt on a successful name — leftover debug output from the original author, harmless, left as-is (out of scope for a UTF-8 compatibility pass).

lpcc sweep

9223 files, 8251 pass / 972 fail (89.5%). Failure tail is the usual shape (missing daemons like STORY_D/TOP_D, undefined globals like prepare_skill/NewRandom, a handful of genuine syntax typos) — not triaged individually per AGENTS.md §6b/§13, boot + full interactive registration test is the verification gate. Memory stayed healthy throughout the sweep (~17-18GB free the whole time, no pressure).

Re-verification pass: driver rebuild + formatter + WASM (2026-07)

• LPC formatter applied to all .lpc under work/ (9,223 total, 9,197 written, 10 unchanged, 16 self-checked errors left untouched).
• Native re-test against the freshly rebuilt driver (~/src/fluffos/build-debug/src/driver): clean boot, zero errors of any kind in debug.log. Full registration flow re-verified end-to-end with a fresh id and real Chinese name (zzfysix/萧峰六): id → confirm y → Chinese name → password ×2 (this lib requires upper+lower+digit/symbol, ≥6 chars) → email → gender → ethnicity (0=汉族) → landed in 凤求凰客栈; score produced the full real attribute sheet, quit clean. Same live-clock-prompt --idle 0.4 pacing adjustment as other libs in this batch was needed here too.
• WASM test (scripts/wasm_client.js against build-wasm/src): boots cleanly (only the expected non-fatal ftpd/dns_master sockets-unavailable preload errors). Full registration + gameplay flow completed successfully under WASM — real Chinese name 秦网七 reached 凤求凰客栈, score showed the correct full attribute sheet, quit worked cleanly. This lib's registration path does not gate on query_ip_number()'s format. Only a cosmetic artifact observed: the "现在共有一位玩家从你的站点（）连线" line shows an empty site name (known WASM query_ip_number() limitation, driver-side, not a mudlib bug) instead of a real hostname/IP — does not affect login or gameplay.

WASM-enablement pass (2026-07): loopback gates + admin seeding

Standard pass per AGENTS.md §1.3b/§1.3e/§1.5:

• adm/daemons/band.lpc: new is_local_site(site) helper (loopback / empty / malformed IP ⇒ local); is_banned() returns 0 for local sites. (logind calls it with BOTH query_ip_name() and query_ip_number() — a resolved hostname passing the local check is acceptable for local play.)
• adm/daemons/logind.lpc:

- logon(): iplimit > 15 per-IP multi-login cap — loopback exempt. - get_passwd() relogin path: 20-second "减轻系统负担" relogin load throttle — loopback exempt (the 1-hour kickout penalty is punishment/game design, kept).

• No uptime() startup-grace gate or registration throttle exists in this lib — nothing else to bypass.
• Admin seeded: fluffos / Mud@2026 / 浮浮, rank (admin) via the lineage's securityd SAVE DATA (adm/daemons/securd.o — this fengyun build stores wiz_status in the daemon's own save file, per its "在wizlist属性里储存，而不再在/adm/etc/wizlist里存储" comment; added "fluffos":"(admin)" to the wiz_status mapping alongside the original xgchen admin). No anti-steal password gate in this lineage — registration with the id already granted admin worked fine. Verified: relogin + update /d/fy/fqkhotel.lpc → 成功 (wizard-style room path display also confirms wizhood).
• Save file paths (note the extra per-id directory level in this lib): data/login/f/fluffos/fluffos.o + data/user/f/fluffos/fluffos.o.
• Retest: fresh normal registration (regtest/秦风测, password rule upper+lower+digit ≥6) end-to-end into 凤求凰客栈 with look/score/quit correct; test saves removed. debug.log clean (no error lines).

WASM long-sit boot-watch pass (2026-07)

200s scripts/wasm_boot_watch.sh sit: no new findings beyond the already-documented sockets-absent preload compile errors (ftpd.lpc, adm/daemons/network/dns_master.lpc — caught by master.lpc's preload catch(), cosmetic). Proactively fixed adm/simul_efun/object.lpc's file_owner() (return name → return dir) as part of a repo-wide port of a bug found live on sibling zhonghua2 (misattributes 3-level-deep /u/<wiz>/<subdir>/<file> log_error writes to a bogus path); didn't fire in this lib's own sit, fixed proactively since it's the identical shared file. Retest: fresh registration (id zzfsanty) through look/score/quit, clean.

深度功能测试 / Deep functional test (2026-07-24)

First real *playthrough* pass on this lib (all prior passes verified only registration + look/score/quit + admin login, or watched boot output). Played as an ordinary new player through most systems, native driver (build-debug), one continuous session per §10.7 with several short reconnects woven in deliberately to exercise net-dead/reconnect and quit/relogin paths. Found and fixed four real bugs, one of them a new bug class; found and fixed one minor pre-existing wizard-command bug along the way; one further pre-existing gap (15 vendor NPCs referencing an undefined F_VENDOR_SALE macro) is a genuine archive content gap, documented but deliberately not "fixed" by fabricating the missing feature.

Test characters (kept, not cleaned up)

• suwan / password Suwan2026 / Chinese name 苏晚照, female, 汉族. Notable state: joined 权力帮 (Power Gang) as a 3rd-generation disciple via apprentice sqx; learned unarmed from 查猛 (an unrelated NPC) via the organic learn ... from ... with ... path; died once (to a wildly overpowered NPC, see below) and respawned; bought and consumed a wineskin and dumpling from the starting inn's vendor; sparred (non-lethally) with 登徒子(dtz) via fight. Currently at 南城客栈 (/d/city/kezhan, a valid_startroom-flagged inn — her startroom was explicitly re-saved there with save as part of this test, see the startroom bug below). Save files: work/data/user/s/suwan/suwan.o, work/data/login/s/suwan/suwan.o.
• linruoxi / password Linrx2026 / Chinese name 林若溪, female, 汉族. A second, untouched fresh-registration character (registered AFTER all fixes below were applied and the driver rebooted), used only to re-confirm the full registration→look→score→i→quit flow end-to-end post-fix. At 凤求凰客栈 (start room), unmodified otherwise. Save files: work/data/user/l/linruoxi/linruoxi.o, work/data/login/l/linruoxi/linruoxi.o.
• Admin fluffos/Mud@2026 (pre-existing, §1.5) was used throughout for reconnaissance (goto, goto -i, summon, clone, give, update) — reading room/NPC source ahead of time and cross-checking suspicious results against a debug instrumentation command (written, tested, then deleted — no scratch files left behind) was what actually root-caused the two feature/vendor.lpc/std/room.lpc bugs below.

What was tested and confirmed working

• Registration: real Chinese names (苏晚照, 林若溪) through the full id→confirm→name→password(2x)→email→gender→ethnicity flow, landing in 凤求凰客栈 correctly both times.
• Movement/exploration: walked from the start inn through 风云城's street grid (d/fy/swind1→fysquare→nwind1..5→goldlion→ghall) to 金狮镖局 (a escort-agency zone) and further east through ecloud1..5 to 风云天骄门(d/fy/hall) — multi-room, multi-zone navigation all correct, room descriptions/exits/NPC-presence all rendered correctly at every stop. Read room .lpc source directly (set("exits", ...)) to plan routes rather than guessing, per §10.7.
• score/hp/i all correct at every state change (fresh, post- sect-join, post-skill-learn, post-death/respawn, post-shopping, post-relogin).
• Skill learning (organic teacher path): d/fy/npc/gmaster.lpc (查猛, id meng, at d/fy/ghall) teaches ANY player (recognize_apprentice() only rejects players with a specific quest flag already set) via learn unarmed from meng with 20 — verified live, skills correctly showed the new skill at 初学乍练/1 afterward. Note: apprentice meng itself never actually completes (查猛 has no attempt_apprentice() override, so the default one silently no-ops — not a bug worth fixing, just means he's a *teacher* NPC, not a *sect founder* NPC in this build; learn doesn't require formal apprenticeship anyway, matching learn.lpc's own recognize_apprentice gate).
• Sect/faction join: d/fy/npc/sqx.lpc (孙青霞, id sqx, 权力帮's 2nd-generation 天王, at d/fy/hall) DOES implement a real attempt_apprentice() that auto-recruits — apprentice sqx instantly and correctly joined 权力帮 as a 3rd-generation disciple (title changed from 普通百姓 to 权力帮第三代弟子 in score, family/master_name set). Confirms the organic sect-join mechanism works when pointed at an NPC that actually implements it — the two first-tried NPCs (查猛, and separately 天机老人/half_god.lpc) both have a stubbed/no-op attempt_apprentice() by original design (确认 码/questgated in 查猛's case, deliberately "extremely picky" in 天机老人's own flavor text) — not a bug, just means the FIRST NPC tried isn't always the right one; had to read several NPC files' actual attempt_apprentice() bodies to find one that really completes the flow.
• Safe sparring: fight <target> is gated through accept_fight() for any NPC with can_speak truthy (effectively all ordinary human NPCs, via std/char/character.lpc's default_ob fallback to adm/daemons/race/human.lpc, which sets can_speak=1) — declines for attitude:"friendly" NPCs unless already fighting, and for anyone not near-full on resources; adm/daemons/combatd.lpc's per-round damage resolution auto-halts a fight (never a kill) once either side drops under 50% kee. Verified live and working correctly against d/fy/npc/dtz.lpc (登徒子, combat_exp 1000, attitude:"heroism" — auto-accepts, roughly newbie-appropriate per help newbie's own "look for someone with comparable attack power" guidance): the fight properly auto-stopped at 31/300 kee (~10%) instead of continuing to death, no crash, correct combat narration throughout. This lib's own d/city/obj/muren.lpc (a accept_fight()-stat-mirroring training dummy in the exact bxsj-style pattern the checklist calls out) is NOT placed in any live room anywhere in the archive — orphaned content, an archive gap, not fixed (would require inventing which room it belongs in). dtz fills the same "safe first sparring target" role well enough in practice.
• Shop purchase: bought a wineskin (20文) and (attempted, correctly rejected for insufficient funds after) a dumpling from the starting inn's 店小二(waiter) via list/buy <item> from waiter, then drink wineskin — full purchase→consume loop confirmed working end-to-end (only after the feature/vendor.lpc fix below; broken before it).
• Combat/death: reached real death (see bug #1 below — accidentally, by fight-ing an NPC far too strong for a fresh character) and full automatic respawn via d/death/npc/panguan.lpc's death_stage() — ghost dialogue sequence, reincarnate(), teleport to a revive room, score's 被杀了X次 counter incremented, stat penalties applied (potential 299→141, exp 2000→1968) — the whole death/respawn cycle works mechanically. (Getting there was NOT the intended safe-sparring test — see bug #1's writeup for why, and don't repeat the mistake of fight-ing a title-bearing high-rank sect NPC as a "safe" spar target next time.)
• Persistence — both layers: net-dead (not quit) reconnect, both within seconds AND after a genuine ~150s wall-clock gap (real sleep, not simulated) — location, HP (regenerated correctly during the gap), inventory, sect membership all intact both times, world time-of-day flavor text advanced correctly across the gap. Clean quit + a real ~90s wall-clock gap + fresh full login — same full state persisted correctly (skills, sect title, stats; only inventory changed, and correctly so — see the item-drop-on-quit note below, matching the bxsj-family design already documented elsewhere in this project). No §7.20/§7.21-style void-stranding or wizard-blocking bug found — obj/user.lpc's net_dead() leaves the player in their REAL room (no VOID_OB parking at all in this lib's design) and its reconnect() is a plain re-link with no location logic needed; there is no mandatory input_to()-driven setup wizard blocking commands after registration either. This lib's net-dead design is structurally immune to both of those bug classes.
• Not verified live, explicitly: the full 900-second (NET_DEAD_TIMEOUT) auto-force-quit path (obj/user.lpc's user_dump(DUMP_NET_DEAD) → command("quit")) — code-reviewed only (it runs quit from within the player's real room, and with this session's other fixes in place there is no known reason it would misbehave, but a real 15-minute wait was judged not worth the time budget for this pass). Also not reached: a wizard-gender/ethnicity-branch second registration pass (both test characters are 汉族/female) — the registration code path itself was already fully verified in earlier passes per this file's own history above, so not repeated here.

Bugs found and fixed

1. d/death/npc/panguan.lpc's death_stage() permanently overwrites the player's LOGIN location on every death — new bug class, not previously in AGENTS.md.

• Symptom: reconnecting (even via a completely clean quit + fresh login, not just net-dead reconnect) after your character's first-ever death lands you in one of two out-of-the-way "death antechamber" rooms instead of your real home — forever, every session, until manually fixed. Reproduced live: suwan died fighting 孙青霞(sqx, a 权力帮 2nd-generation 天王 with combat_exp 1,000,000 and force 10000 — picked as a fight target by mistake, since she was the same NPC used moments earlier for a successful, intended apprentice sect-join; her attitude is "friendly" like every other safely-decline-able NPC, but the single hit that killed suwan outright happened in the SAME combat round accept_fight() was evaluated in, before the 50%-kee auto-halt in combatd.lpc got a chance to apply — a fight against a wildly stronger NPC can still one-shot-kill, fight's non-lethality guarantee only protects against a slow multi-round grind, not a single overwhelming blow. This is arguably a real, if harsh, design property the game already accounts for — help newbie explicitly warns "只有 通过和你战斗水平差不多的人实战" — so it's noted here as the trigger, not filed as a bug in itself). After the death sequence completed, every subsequent login (clean quit+relogin AND net-dead reconnect) placed suwan in /u/guanwai/tower — a small, mostly self-contained "关外" zone with no short path back to the main city map (its only exit chain leads further outward toward /u/resort//u/ghost, not back toward d/fy).
• Root cause, d/death/npc/panguan.lpc:55-56 (before fix):

``lpc ob->move(revive_loc[random(sizeof(revive_loc))]); ob->set("startroom", base_name(environment(ob))); ` where revive_loc = ({ "/d/fy/church", "/u/guanwai/tower" }). This runs unconditionally at the end of every player's death sequence (death_stage(), call_out-chained from panguan.lpc's init(), which fires whenever a dying player enters 鬼门关). query("startroom") is the SAME field adm/daemons/logind.lpc's enter_world() reads on every future FULL login to decide where to place the player — so this silently and permanently hijacks it. Neither /d/fy/church nor /u/guanwai/tower carries the valid_startroom flag that cmds/usr/save.lpc` (the only OTHER place in the whole codebase that legitimately updates this field) requires before doing the exact same thing — strong evidence this was meant to be a temporary "you glimpse a wandering ghost" placement for the death/limbo flavor scene, not a permanent relocation of the player's home. No player-facing message indicates anything about their login location changing.

• Fix (d/death/npc/panguan.lpc): deleted the ob->set("startroom", ...) line entirely — the move() above already handles correct immediate post-death placement; the player's real, previously-chosen login location is left untouched.
• Verified: reproduced pre-fix (confirmed via direct save-file/behavior inspection that suwan's startroom had become /u/guanwai/tower); applied the fix, rebuilt nothing (LPC-only), full driver restart, re-tested a FRESH death is no longer reachable to re-verify cheaply without repeating the mistake, so verified indirectly instead: the file compiles clean (update succeeded) and manual code review confirms no other line touches startroom. suwan's own already-corrupted save was separately, manually repaired for this writeup by walking her to a real valid_startroom room (/d/city/kezhan, itself only reachable/populate-able after bug #3's fix) and running save there — confirmed via the "当你下次连线进来 时，会从这里开始" message and a subsequent clean relogin landing there correctly.
• New AGENTS.md bug-class draft (see final report — not added to AGENTS.md directly per instructions, the orchestrating session owns that file).

2. feature/vendor.lpc — a single bad vendor_goods path crashed list/buy for the ENTIRE shop, and 13 NPC files across the archive carried an identical bad path (matches AGENTS.md §7.18's "stale hardcoded path" shape, generalized to a shared vendor helper — worth folding into that entry or its own note, orchestrating session's call).

• Symptom: list at the STARTING INN (d/fy/npc/waiter.lpc, present from the very first room every player sees) crashed with a caught but real runtime error, printing NOTHING useful to the player ("执行时段错误：*call_other() couldn't find object '/obj/example/wineskin'." from /feature/vendor.lpc:73) instead of the intended goods list — the entire shop was unusable for every vendor sharing this bug, not just missing the one bad item.
• Root cause: d/fy/npc/waiter.lpc's (and 12 other NPCs') vendor_goods mapping references /obj/example/wineskin, which does not exist in this archive (the real file is /obj/food/wineskin.lpc — confirmed by the archive shipping both /obj/example/chicken_leg.lpc and /obj/example/dumpling.lpc, the mapping's OTHER two entries, which really do live under /obj/example/, only wineskin was misfiled/ mis-referenced). feature/vendor.lpc's do_vendor_list()/ buy_object()/compelete_trade() all call methods directly on the mapping's string KEYS (name[i]->name() etc.), which lazily load_object()s the path — throws uncaught on the first missing entry, aborting the ENTIRE loop (not just that one item).
• Fix, two parts: 1. Repointed all 13 files' /obj/example/wineskin references (both vendor_goods mapping keys and bare carry_object(...) calls) to the real /obj/food/wineskin: d/laowu/npc/{dang,drunk,waiter}.lpc, d/kaifeng/npc/waiter.lpc, d/chenxiang/npc/oldman.lpc, d/tieflag/npc/xiaofan.lpc, d/fy/npc/{waiter,biaotou,drunk,waiter_bak}.lpc, d/songshan/npc/waiter.lpc, u/tangmen/npc/waiter.lpc, u/wudang/npc/waiter.lpc. 2. Hardened the SHARED feature/vendor.lpc itself (all three functions) so any OTHER currently-undiscovered bad vendor_goods reference, in any of the remaining ~114 vendor files in this archive, degrades to "skip that one item" instead of crashing the whole vendor:

``lpc // BEFORE (do_vendor_list, same shape in buy_object/compelete_trade): for (i = 0; i < sizeof(name); i++) list += sprintf("%-10s %-20s ：%s\n", name[i]->name(), name[i]->query("id"), price_string(name[i]->query("value"))); // AFTER: for (i = 0; i < sizeof(name); i++) { if (catch(load_object(name[i]))) continue; list += sprintf("%-10s %-20s ：%s\n", ...); } ` Used catch(load_object(...)), NOT file_size(...) — tried file_size() first and it silently hid EVERY item (not just bad ones), because file_size() does NOT do the .lpc/.c extension fallback that load_object()/call_other() do (AGENTS.md §4.2), so a bare file_size("/obj/example/dumpling") (no extension) always returns -1 even for files that genuinely exist as dumpling.lpc`. Caught this immediately by re-testing live rather than trusting the first fix — worth calling out since it's an easy trap to fall into silently (the "fixed" version LOOKED successful — no crash — but was actually worse, hiding the whole shop instead of just the one bad item).

• Verified: reproduced pre-fix (list crash, exact error above); applied fix 1, retested — list then printed correctly but was EMPTY (fix 2's file_size() version, the trap described above); fixed fix 2 to use catch(load_object(...)), retested — list now shows all 5 real items including 牛皮酒袋(wineskin); buy wineskin from waiter → drink wineskin completed a full purchase→consume cycle live, correct coin deduction, correct insufficient-funds rejection on a second purchase attempt.

3. std/room.lpc's make_inventory() — an unguarded new(file) + ->move() chain crashed a room's population the first time ANY of its listed NPCs/objects was missing OR failed to compile; 36+ "objects" mapping entries across the archive reference paths that don't exist, one specific one (d/city/npc/xiaoer.lpc) exists but fails to compile. Matches AGENTS.md §7.14's "Factory calls chained without a check" class, generalized to the shared room-population helper (not previously documented at that scope).

• Symptom: goto /d/city/kezhan (南城客栈, an ordinary city inn, its room ITSELF has valid_startroom:1 — i.e. intended as a legitimate home base) threw an uncaught error reaching all the way up through move()/the calling command, on the room's first-ever population this boot: *Bad argument 1 to EFUN call_other(). Expected: object, string, array, Got: int(0). at /std/room.lpc:19 (ob->move(this_object()) where ob was the 0 returned by a failed new(file)), immediately followed (after the first fix pass) by a SECOND, different failure mode from the SAME line: *No program in object '/d/city/npc/xiaoer'! — new() doesn't always fail gracefully; when the target file EXISTS but fails to COMPILE, new() throws instead of returning 0, so a plain objectp(ob) post-check isn't enough on its own.
• Root cause, two independent flavors found live in the SAME room: 1. d/city/kezhan.lpc's objects mapping lists /u/bbrbbt/seng, /d/ourhome/npc/bigeye, /u/yudian/npc/new_jing — none of these files exist anywhere in the archive (stale wizard-workspace references, an archive gap). A repo-wide scan of every set("objects", ([...])) mapping across the whole lib found 36 more such missing-path entries in 20+ other rooms (d/xiakedao/, d/wiz/, d/shaolin/, d/city/ zones especially) — this is a systemic archive gap, not a one-room fluke. 2. d/city/npc/xiaoer.lpc:1 — inherit F_VENDOR_SALE; — and 14 OTHER d/city/npc/*.lpc files — reference a macro that is never #defined anywhere in include/*.h. No F_VENDOR_SALE-equivalent feature file exists in this archive either (only plain /feature/vendor.lpc, mapped to F_VENDOR, exists) — and xiaoer.lpc's own vendor_goods mapping uses the OPPOSITE key/value convention from feature/vendor.lpc (id-string keys → path values, vs. feature/vendor.lpc's path keys → price values), confirming this was meant to inherit a genuinely different, richer vendor feature that simply isn't present in this archive — a real content gap, not a typo'd path to an existing file, so NOT "fixed" by aliasing the macro to feature/vendor.lpc (that would compile but behave wrong at runtime). Left as a documented gap; these 15 d/city/npc/*.lpc files (boss, yang, laowei, jin, liu, teawaiter, xiaoxiao, laosun, tiejiang, jia, xiaoliu, lu, xiaoer, bookseller, pig_counter) simply never compile/populate in this build — an entire zone's worth of vendor flavor NPCs, silently absent.
• Fix (std/room.lpc, the SHARED base every room in the mudlib inherits): wrapped make_inventory()'s construction in catch() and added objectp() guards at its two call sites in reset():

``lpc // BEFORE: object make_inventory(string file) { object ob; ob = new(file); ob->move(this_object()); ob->set("startroom", base_name(this_object())); return ob; } // AFTER: object make_inventory(string file) { object ob; if (catch(ob = new(file)) || !objectp(ob)) return 0; ob->move(this_object()); ob->set("startroom", base_name(this_object())); return ob; } ` plus objectp(ob[list[i]])/objectp(ob[list[i]][j]) guards before the two ->is_character()/->return_home() call sites in reset() that consume make_inventory()'s result. Also (separately, same first-visit-crash shape, found while investigating this) wrapped 13 rooms' bare call_other("/obj/board/<name>", "???") force-load calls (a per-room message-board lazy-load idiom used in 14 rooms archive-wide) in catch() — 5 of the 11 unique referenced boards (idle_b,nancheng_b,advise_b,wizard_l,zhuzi_b) don't exist either, same missing-content shape, same uncaught-crash risk, same fix pattern: d/city/{kezhan,misc/idleroom,misc/advise,club3}.lpc, d/wiz/workroom.lpc`.

• Verified: reproduced pre-fix (exact errors above, twice — once before the catch() was added to make_inventory(), once after that fix when the SEPARATE xiaoer.lpc compile failure surfaced as a new crash from the same line); applied both fixes, full driver restart, goto /d/city/kezhan now loads cleanly (room description/exits render correctly, no crash reaches the player or the command dispatcher) — the caught errors are still logged to debug.log (visibly marked "错误讯息被拦截"/"error message intercepted", i.e. genuinely caught, not propagating) for future maintainers to notice, matching this project's established convention for cosmetic-but-informative caught errors. Re-ran the full registration→look→score→i→quit flow with a brand-new character (linruoxi/林若溪) after this fix — clean, zero new debug.log errors of any kind beyond the two intentionally-caught ones from this room's own inherently-missing content.
• This is a NEW bug class draft — see final report.

4. cmds/wiz/summon.lpc — minor, admin-only: null-check ordering bug (not part of the main playthrough, found incidentally while using summon for test-character positioning).

• ob->query("id") == "xgchen" was checked BEFORE if (!ob) — summoning any currently-OFFLINE player id crashed with *Bad argument 1 to EFUN call_other() instead of printing the intended "咦... 有这个人吗?" rejection.
• Fix: reordered the two checks (null-check first).
• Verified: reproduced live (summoning suwan while she was offline), fixed, retested — correct rejection message, no crash.

Lineage note

zzfy shares its std/, feature/, and adm/daemons/logind.lpc architecture with fy3xd/fy3dz (same 风云3 engine, AGENTS.md §11) and, one level further out, the whole 风云 family (fy2/fy2qh, fengyun434/fy2005, moniHuafu). Bugs #2 and #3 above live in shared engine files (feature/vendor.lpc, std/room.lpc) that are very likely byte-identical or near-identical across at least the 风云3 siblings, and possibly the wider family too — worth a proactive port-and-diff pass on those two files in fy3xd/fy3dz (out of scope for this task; not done here). Bug #1 (d/death/npc/panguan.lpc) and bug #4 (cmds/wiz/summon.lpc) are smaller, single-file changes — check whether the siblings carry the identical file before assuming the fix is needed there too, rather than porting blind.

深度功能测试（2026-08-13，round two，新驱动重测）

针对驱动升级（quest_times/win_times %-operator 修复 + Warning/warning 大小写回退兼容）做的重测。上面记录的四个修复逐项核对代码仍然生 效；log_error()（实际生效的 adm/obj/master.lpc）确认已经有 strsrch(message, "arning:") == -1 判断，win_times 也已用 to_int(query("win_times")) % 5（仅存在于 u/wiz/u/xxy/{city2,huashan}/npc/refereew.lpc 这份巫师个人工作目 录副本里，已确认修复覆盖到位）。

本轮新发现并修复的 PROGRAMMING bug

1. 管理员账号播种从未真正生效——securityd.lpc 的 wiz_status 是 nosave，每次开机只硬编码授予 xgchen（AGENTS.md 已归档的 "wiz_status nosave 硬编码单一 bootstrap id"类）：fluffos 早 在 2026-07-24 那轮深挖里就已经真实注册过（存档 data/{login,user}/f/fluffos/fluffos.o 一直存在，密码已设 定），但从未被授予任何权限——wizlist 指令显示"本游戏没有管理 巫师"（后来确认这条指令本身是硬编码的固定文案，不反映真实状 态，是干扰项），改用 update 指令直接验证：修复前从未测试过写 权限。已在 restore_list() 里 set("wiz_status/xgchen", "(admin)") 后面加一行 set("wiz_status/fluffos", "(admin)")， 不改动既有的 xgchen 硬编码。Live 验证：重启后 fluffos 登录， update /adm/simul_efun/file 显示"重新编译 /adm/simul_efun/file.lpc：成功！"，确认拥有 (admin) 级写权 限。 2. log_file()（adm/simul_efun/file.lpc）本身缺少 assure_file() 保护：已加上前向声明 + assure_file(LOG_DIR + file);。 3. cat()（同一文件）对不存在文件的空指针式崩溃，主动加固： 未在本档案现场触发，属主动加固，改成 write(read_file(file) || "");。 4. get_resp()/get_name()（adm/daemons/logind.lpc）各有一处 调试残留 printf("%O\n", ob)（AGENTS.md §7.34-class）：紧跟 在中文名字确认之后，把连线桩物件的原始引用直接回显给正在注册 的新玩家。已删除两处。

Proactive checks（无需改动）

• §8.9 食物/饮水初始化不适用：init_new_player() 之类的流程直接 无条件 user->set("food", user->max_food_capacity())，不涉及 ob/user 混淆读取。
• feature/dbase.lpc 未发现密码写保护，不适用 tybxjh/wlhd 那一类 bug。

实测过程

管理员 fluffos（此前深挖已注册、密码 Mud@2026，但从未真正拿到 过权限）用真实密码重新连线两次（第一次先用 update 验证写权限， 第二次单独验证密码重连本身）：均成功登录，update /adm/simul_efun/file 确认 (admin) 权限生效，存档数据一致。全程 debug.log 无运行时错误。驱动按精确 PID 结束；测试期间产生的存档 时间戳增量已 git checkout -- 还原。

WASM 修复摘要（迁移自 meta.json 的 group_note）

风云III 引擎基础版（郑州风云3）。

§7.86 跨库扫描修复（留言板 post 崩溃）

• BULLETIN_BOARD inherit + 多余 replace_program() 致命形状（AGENTS.md §7.86，post 命令崩溃）：全档案 34 处命中，已删除多余的 replace_program(...) 调用（保留 inherit），逐文件保留原有行尾格式（CRLF/LF 按文件原样）。本次为跨库 §7.86 扫描修复（触发原因：该 bug 已在 6+ 个互不相关的血统家族独立确认，属于近乎普遍的拷贝粘贴模式），仅做编译检查（驱动干净启动、端口正常监听），未做完整 §10.7 深度游玩测试。

§7.100 房间基类 replace_program() 扫尾修复（2026-08-19）

ROOM 宏（/std/room）在本档案 1,967 处房间文件的 create() 里 紧跟 inherit ROOM; 之后又多余调用了一次 replace_program(ROOM); ——AGENTS.md §7.100 记录的同一个休眠 bug。用 fix_710_room.py 先 扫过全部 work/，删除 1,817 处标准形状；剩余 150 处存活是三种此 前批次已知的不规则形态：（1）u/wiz/u/heart/xuedao/ 目录本身混 了标准形状（另用脚本单独对该子目录再扫一次，删除 40 处）和一种 不规则形状——143 个文件带有"如果没有 init 函数请不要删除这句话" 尾随 UTF-8 注释（和 tiexuejianghu 血统同源），按 tiexuejianghu 那次的既有先例整行删除（含注释）；（2）14 个文件、18 处帮派/房间 建造工具的字符串拼接变体（cmds/adm/roommaker.lpc 一份文件里就 有 3 处不同写法、d/wiz/xgchen/roommaker.lpc、obj/wall.lpc、 u/wiz/u/panguan/{room,rmmaker}.lpc），手工改成 str += "... setup();\n}\n";／room_code += "...setup();\n}\n";。合计 1817+40+143+18=2018 处操作中有 40 处和标准扫描的一部分重叠计数， 实际净删以 git diff --stat 为准：显示 1963 个文件净删 1967 行、 增 7 行，与 survey 记录的存活总数（1,967）完全吻合。work/data/ 下没有真实 .lpc 源码命中。

驱动干净启动（零新增编译错误、端口 40025 正常监听、debug.log 无任何"cannot replace"/"cannot bind"行）。管理员 fluffos/ Mud@2026 实机登录成功，look/score/quit 均正常。验证过程 中撞上一个纯测试脚本层面的假象，记录以免下次重复踩坑：这份档案 登录后有一个每秒刷新的 HH:MM:SS> 提示符，"等到静默"式的 recv 循环会因为提示符持续刷新而永远等不到静默，进而误判为驱动 挂起（本次先后被两个提前用 SIGKILL/SIGTERM 粗暴杀掉的测试连接搞 成"账号仍被视为在线"的僵尸状态，进一步放大了误判）——换成"固定 总时长"而非"等到静默"的读取策略后，登录/look/score/quit 全部正 常完成，debug.log 全程干净，确认不是本次改动引入的回归，也不 是真实的驱动 bug。管理员存档的时间戳漂移已用 git checkout HEAD -- 还原，未提交。驱动按精确 PID 结束。

§7.30 uninitialized-mapping accessor sweep (2026-08-20)

Corpus-wide mechanical sweep of the feature/skill.lpc shared-lineage bug (confirmed independently on xiakexing2017/jqxz2015/haiyang2 via round-four testing): 3 accessor(s) in this file returned a raw never-initialized mapping instance variable (defaults to int 0, not ([]), until first assigned), crashing any unguarded keys()/sizeof()/indexing caller for a fresh/untrained character. Fixed at the accessor level (mapp(x) ? x : ([])) per the documented remedy. Verified via lpcc --batch static compile check only (not a live boot) as part of a large mechanical sweep; not individually functionally re-tested live on this lib.

深度功能测试（§10.7，2026-10-04）— negative range ends

scripts/lpc_fix_negative_ranges.py: 17 line(s) in 5 file(s) count a range from the end with < (x[a..-1] -> x[a..<1], x[-2..] -> x[<2..]; KB 06 §7.209). In this driver a negative constant end gave "" / ({ }) and a negative start the whole value, so each of these returned the wrong slice; the compile warning A negative constant as the second element of arr[x..y] is gone from the files that compile. Range lvalues (s[0..-1] = text, the prepend idiom) are left alone. A paired compile of the changed files at HEAD and in the working tree showed no new error.

深度功能测试（§10.7，2026-10-04）— undefined macro and simul_efun gaps

• 11 file(s) inherit F_SSERVER; and no header defines it (11 did not compile: syntax error, unexpected L_IDENTIFIER, expecting L_STRING or '('). #define F_SSERVER SSERVER in include/globals.h (KB 04 §6.2, scripts/lpc_alias_macro.py). 1 still fail.

深度功能测试（§10.7，2026-10-05）— here-document terminators

scripts/lpc_fix_heredoc_terminator.py --all --apply: 3 here-document block(s) repaired (3 with the terminator glued to the last text line, 0 with it indented), so End of file in text block no longer hides the room or object behind them: u/wiz/u/xxy/room/heimuya/shenggu.lpc, u/wiz/u/xxy/room/heimuya/tang.lpc, u/wiz/u/xxy/room/heimuya/npc/tang.lpc. The terminator must start its line (the lexer's own test, KB 04 §6.10); a glued one lost its newline in the GBK -> UTF-8 conversion together with a half character, which stays out.

深度功能测试（§10.7，2026-10-05）— tail() simul_efun

scripts/lpc_add_tail.py --apply: cmds/adm/tail.lpc, cmds/wiz/tail.lpc call tail(file), an efun of the MudOS this archive ran on that FluffOS does not have (Undefined function tail: the command did not compile). adm/simul_efun/file.lpc now defines void tail(string file): the last 10 lines of the file, written to the player. Load check (HEAD against the working tree): 2 of the 2 tail command file(s) failed to load at HEAD, all load now, and the simul_efun file loads. KB 04 §6.2.

深度功能测试（§10.7，2026-10-05）— 任务表里的 .c 路径与任务守护进程的防护

scripts/lpc_dynamic_rows.py：quest/dynamic_location、tongji/dynamic_location 里 2308 行写的是原档案的 .c 路径。转档时文件已改名 .lpc，而显式后缀按原样解析，load_object("/d/x.c")、new("/d/obj/quest/x.c") 返回 0 且不报错（KB 03 §4.2，KB 06 §7.220），所以 adm/daemons/questd.lpc、adm/daemons/tongjid.lpc、d/meiyuan/npc/zhuge.lpc、d/yihua/npc/yaoyue.lpc、tongji/npc/zhuge.lpc 读到的每一行都加载不出来：随机任务和任务物品从未生成，在 create() 里直接对结果调用 tar->set() 的版本还会崩溃，连带加载它的 cron 守护进程和 give 的任务钩子。这些行已改为 .lpc；另有 43 行指向的文件两种后缀都不存在（档案缺内容），保持原样。 scripts/lpc_dynamic_quest_guard.py：1 处 spread_quest() 调用（init_dynamic_quest() 的循环里）包进 catch()，缺失的任务物品或编译不过的房间只跳过那一个任务，守护进程照常加载、其余任务照常生成。 验证：adm/daemons/questd.lpc、adm/daemons/tongjid.lpc、d/meiyuan/npc/zhuge.lpc、d/yihua/npc/yaoyue.lpc、tongji/npc/zhuge.lpc和 cron 守护进程在新进程里加载（HEAD 一次，工作树三次，选房是随机的）：HEAD 通过 5 个，工作树三次分别通过 5、5、5 个（共 6 个），无回退；d/meiyuan/npc/zhuge.lpc 在 HEAD 与工作树都加载失败，另有原因，与本次无关。

深度功能测试（§10.7，2026-10-06）— 与 zhonghua2 同源的三处编译警告

• include/npc/indian.lpc 与 masia.lpc 共用全局名 ma_sname/ma_pname；两个头文件被同一个守护进程（npcd 等）一起包含时，后包含的印度名表覆盖马来名表，generate_ma_name() 生成的是印度名字。印度表改名 in_sname/in_pname（KB 04 §6.10）。

验证：改动的文件和包含 npc/indian 的守护进程在 HEAD 与工作树分别加载（scripts/lpc_listcheck.sh，新进程），无回退。

深度功能测试（§10.7，2026-10-07）— 编译警告整理

先用 scripts/lpc_twin_port.py adopt wqfy zzfy 把 wqfy 警告整理提交（0dd906c83e7）里的改动搬到约 525 个修改前与它一致的文件（逐文件核对非 ASCII 文字：除 wqfy 自己删掉的以外没有丢失或改动），剩下的再跑 scripts/lpc_warnings.py zzfy --fix 加 scripts/lpc_fix_no_effect.py --apply，共改动 816 个 .lpc 和若干 .h。 与 wqfy 相同的两处真错误：adm/daemons/combatd.lpc 易筋经反震"内伤"800–1600 一档少了 return，描述算出来就扔了；d/kaifeng/npc/huzhou.lpc 的 command("follow %s", ob->query("id")) 在本库是编译错误，护舟这个 NPC 整个加载不了——改成 command("follow " + ob->query("id")) 后能加载了。 HEAD 与工作树分别加载（新进程）：745 -> 746 通过（多出的就是 huzhou），无回退；改过的头文件的 25 个包含者 24 -> 24 通过。

启动时警告 96 -> 0。

• 2026-10-10：编译警告第二轮（按驱动报告逐条修）：按位 & 改逻辑运算 0 处，返回类型/原型 9 处，NPC+F_UNIQUE 的 init() 0 个，带初始化的未用局部变量 1 个，只在 #if 分支用的声明移进分支 1 个文件；改动 10 个文件，HEAD 与工作树各载入一次无回退（10 objects; HEAD PASS 5, tree PASS 5; regress 0; newly loading 0），这些文件的警告 11 -> 1。

深度功能测试（§10.7，2026-10-11）：“无副作用表达式”警告复查

• u/wiz/u/lxh/emoted.lpc：删去空的 if (!channel_emote) {}（下一句才被报警）。载入无回退。

编译警告收尾（2026-10-12）

scripts/toolkit_fix.sh 一轮加手工两处，能编译的文件警告清零。改动的 31 个对象 HEAD 与工作树各载入一次：HEAD PASS 11，tree PASS 12，regress 0，新载入 1。

• 冗余继承 28 处（lpc_diamonds.py）。其中 d/fy/fish.lpc 用 room.h 的 CREATE_CHAT_ROOM 宏（宏里已 inherit ROOM）之外又写了 inherit ROOM;，驱动报 Illegal to redefine 'nomask' function '_query' 等，这个房间从来载不进来；删去后能载入。
• cmds/adm/roommaker.lpc：special_props 声明两次，第二次（多一个 "file_name"）的初始化在后、本来就生效；删去第一次。
• u/wiz/u/lxh/globals.h（topd.lpc、test.lpc 在全局 globals.h 之外再包含它，有意把排行榜标题换成“铁血江湖…”）：RICH_B/PKER_B/EXP_B 前加 #undef，取值不变；BANNED_SITES 末尾多余的 ; 去掉（同 fys）。

• 2026-10-14：新驱动（FluffOS 970d9f09）报 Float value truncated to int（#1426）。2 个文件 2 处：赋值、初始化、return 改写成 to_int(...)（值不变，只是写明截断）。改动的对象新进程各载入一次，无回退（scripts/lpc_fix_float_trunc.py，KB 04 §6.10）。
• 2026-10-14：技能文件里 int power_point()/int hard_point() 返回 1.5、1.75 之类的浮点，经 int 返回被截断；本库没有调用者（练功走 practice_bonus/learn_bonus），改成 float 返回值，保留原数值。

• 2026-10-14：3 个文件因 Illegal to inherit after defining global variables 整份不编译：FluffOS 要求所有 inherit 写在第一个全局变量之前，MudOS 不管。scripts/lpc_fix_inherit_order.py 把写在 inherit 之前的全局变量（以及声明全局变量的头文件的 #include）挪到最后一个 inherit 之后（KB 04 §6.3）。

• 2026-10-14：1 个文件用了颜色宏（HIW、NOR）却没有 #include <ansi.h>（首个错误 syntax error, unexpected L_STRING）；开头补上；1 个文件的函数声明为 int（或 void）却返回字符串、对象（int query_seal() { return "…"; }），旧驱动照样把值交给调用者；改为实际返回的类型（scripts/lpc_fix_void_return.py）（KB 04 §6.1、§6.3）。新编译出来的文件露出的警告一并清掉（删去未用局部变量，有返回值函数里的空 return; 改为 return 0;）。新进程比对 HEAD 与工作树：4 个对象，新增载入 4，回退 0。

• 2026-10-14：讲故事发奖励系统（adm/daemons/storyd.lpc 与 clone/gift/ 下 20 个故事）照 obj/gift/readme.txt 应在 include/globals.h 定义 STORY_D、STORY_DIR，安装时漏了，这 23 个文件整份不编译；补上两个宏后全部载入（新进程比对，回退 0）。readme 还要求把 storyd 加进 preload，那是开不开这项玩法的决定，没有动。u/wiz/u/xxy/city2/library.lpc 自己的 STORY_DIR 前加 #undef。

其他体系的属性名（2026-10-15，KB 06 §7.241 第三类，--any-file）

10 个文件（技能、绝招、兵器、书籍、区域 NPC 等）来自侠客行体系（qi/jing/neili），而本库人物是 gin/kee/sen/force：receive_damage("qi", ...) 报 伤害种类错误，query("neili") 永远为 0。按 qi→kee、jing→sen（本库学习/读书指令消耗的就是它）、neili→force、max_neili→max_force、jiali→force_factor 改名（scripts/lpc_port_es_props.py --any-file）。

内力/上限当作伤害种类（补）（2026-10-15）

victim->receive_wound("force", n, me)（兵器吸内力）、receive_damage("max_qi", random(100))（走火入魔扣上限）等本库不接受的种类，1 处改为 add(type, -N)。

技能名拼错（2026-10-15，KB 06 §7.249）

query_skill() 里的技能名拼错了，或用的是本库改名前的旧名；本库没有这个技能，调用永远得 0：门槛（< N 就拒绝）永远过不去，相关功夫或绝招学不了、用不了；内功互斥（> N 就拒绝）从不生效；公式里这一项一直是 0。经汉字名（提示文字或中文名词典）或纯拼写差异（_/-、少一个连字符、相邻字母颠倒）核对后，改为本库的技能名（scripts/lpc_fix_skill_typo.py）：yihua-jimu→yihua-jiemu（/cmds/daemon/skill/yihua-jiemu、qishier-jueji、yihua-jiemu）；putfiforce→putiforce（/cmds/daemon/skill/qiankunwuxing）；xizia-shengong→zixia-shengong（hunyuan-zhang）。

NPC 带着本库没有的物品，载入失败（2026-10-15，KB 06 §7.73）

178 个文件、249 处在 create() 里 carry_object("<本库没有的文件>")->wear()（或 wield()、set_amount()）：carry_object() 得 0，接着的调用报 Bad argument 1 to EFUN call_other()，NPC 根本载入不了，该有它的房间里就少了这个人。多是从别的系统搬来的 NPC 用了原系统的物品路径（最常见：/clone/misc/cloth、/d/obj/cloth/linen、/d/obj/cloth/choupao）。每处改为 { object carried_ob = carry_object(...); if (carried_ob) carried_ob->wear(); }（scripts/lpc_guard_carry_object.py）：NPC 照常出现，只是不带这件东西；文件补回来（或是虚拟对象）照样会带上。载入检查：改动的文件改前能载入 1 个，改后 85 个（84 个原来载入不了），没有退步；其余 93 个另有原因（多为 set_skill() 了本库没有的武功）。

物品清单把名字的第一个字吃掉（2026-10-18，KB 07 §8.1）

inventory、房间的物品列表（feature/move）和部分 look 把同名物品合并计数，键是 query("unit") + short()，只有一件时按 key[2..] 去掉量词。量词是一个汉字，在原来的 GBK 里占 2 字节，[2..] 正好去掉量词；转成 UTF-8 以后一个汉字只占一个下标，这一刀把名字的第一个字也切掉了（「把长剑」只显示「剑」，带颜色的名字切掉的是转义序列开头）。改为 [1..]（scripts/lpc_fix_unit_cut.py）：cmds/usr/inventory.lpc。

set_skill() 遇到本库没有文件的武功不再抛错（2026-10-18，KB 06 §7.256）

feature/skill.lpc 里的 set_skill()、map_skill()、prepare_skill() 碰到没有武功文件的名字就 error("F_SKILL: No such skill (…)")，调用它们的 NPC create() 整个中断：NPC 载入不了，刷出它的房间也少一个人。本库有 94 个对象设了本库没有文件的武功（161 个名字，最多的是 claw、yunlong-shengong、yunlong-shenfa）。三处都改成 return：这一门武功不给、不映射，对象照常载入，只是少了它（缺的武功本身仍是内容缺口）。抽样载入这些对象：30 个里 18 个通过，其余 12 个停在别的错误上（d/city/npc/xiaoxiao：*No program in object '/d/city/npc/xiaoxiao'!；d/city/npc/yg/yg-dragon：*Inherited file '/d/dntg/yunlou/npc/yaoguai' does not exist!……）。（scripts/lpc_guard_set_skill.py）

通向不存在房间的出口，按对面房间的回头出口改正（2026-10-18，KB 06 §7.257）

房间出口 "west": __DIR__"xxx" 指向的文件本库没有，玩家走过去就被卡住（笔误、改过名的房间，或者没带进档案的编号房间）。对面的房间存在、并且它相反方向的出口正好指回这个房间，说明出口原本就是通向它：/d/menggu/ta2 的 down：liuhe1 → ta1；/d/menggu/tadi3 的 northeast：tadi4 → yadi4。改动后每个出口都能解析到本库的文件（scripts/lpc_fix_dead_exits.py；目标在运行时本来就能载入的库——虚拟房间——不动）。

补回本库没有的基类文件（2026-10-18，KB 04 §6.15）

下面这些基类本库没有，继承它们的对象（房间、NPC、物品）整批载入不了（inherit XXX; 的宏没定义，或者 inherit "/std/…"; 的文件不在）。同系的库里有同名文件，两库有大量逐字节相同的文件（亲缘关系），所以从那边原样拷来，缺的宏补进全局头文件 include/globals.h：/feature/vendor_sale（宏 F_VENDOR_SALE）←xianlvqiyuan，15 个对象继承它，抽样 10/10 载入；/feature/unique（宏 F_UNIQUE）←haiyang2，14 个对象继承它，抽样 10/10 载入；/std/weapon/stick（宏 STICK）←xianlvqiyuan，5 个对象继承它，抽样 4/5 载入；/std/char/npcsave（宏 NPC_SAVE）←xianlvqiyuan，3 个对象继承它，抽样 3/3 载入；/std/weapon/unarmed（宏 UNARMED）←sjsh，1 个对象继承它，抽样 1/1 载入。每个基类补回后单独载入、再抽样载入使用它的对象，基类载不进来或者一个用户也载不进来的就撤掉（本库撤掉了它们，不再补）。（scripts/lpc_restore_bases.py）

武功名拼错的 NPC（2026-10-19，KB 06 §7.258）

set_skill("kunlun-shnfa", …)、map_skill("dodge", "shaolin-shnfa") 这样的调用把武功名拼错了：本库没有这个武功文件，有的是拼对的那个。以前这种 NPC 在 create() 里报 No such skill 载入不了；武功缺失时跳过之后它们能载入，可是少了那门武功（set_skill 被跳过，map_skill 映射不到，门派弟子用普通轻功/拳脚）。按核对过的对照表改正，共 27 处、9 个文件：taiji-force→taijiforce（16 处）、liumai-shenjian→liumaishenjian（4 处）、lotus-force→lotusforce（4 处）、tang-balde→tang-blade（2 处）、duan-yirangzhi→duan-yiyangzhi（1 处）。（scripts/lpc_fix_skill_names.py）

武功基类缺少 NewRandom()（2026-10-19，KB 06 §7.261）

基本武功（boxing、houquan、westsword、taiji-quan 等）用 action[NewRandom(i, 20, level/5)] 挑招式；NewRandom() 和它调用的 sum() 在别的库的武功基类里有，本库的 /std/skill.lpc 是更精简的版本，没有，于是这些武功全都报 Undefined function NewRandom 编译不过，有这门武功的人物只能空手。把这一对函数（115 个库共用的同一份文字）补进武功基类，受益武功 4 个：/cmds/daemon/skill/jueqing-zhang.lpc、/cmds/daemon/skill/yifeng-jian.lpc、/daemon/skill/fanliangyi-dao.lpc、/daemon/skill/poyu-quan.lpc。（scripts/lpc_restore_skill_helpers.py）

技能基类丢了“预备武功”那一段（2026-10-19，KB 06 §7.262）

feature/skill.lpc 有 skill_map/map_skill()，却没有预备武功（prepare_skill）这一块：变量 skill_prepare、prepare_skill()、query_skill_prepared()、query_skill_prepare()。来自 xkx 系列的 NPC 在 create() 里调用 prepare_skill("cuff", ...)，于是 Undefined function prepare_skill 整个 NPC 载入不了。从同系的 xkx100 把这一块原样补回（声明加在 skill_map 后面，函数加在文件末尾），受益 NPC 27 个：d/kaifeng/npc/tufei2、d/kaifeng/npc/tufei3、d/kaifeng/npc/tufei4、d/kaifeng/npc/tufei5。（scripts/lpc_transplant_function.py）

laozhe 的 command("give yancao to", me)（2026-10-19，KB 06 §7.266）

u/wiz/u/xxy/room/heimuya/npc/laozhe.lpc（巫师目录里的副本）的 init() 里写着 command("give yancao to", me);，command() 只收一个字符串，驱动报 Too many arguments to 'command'，整个 NPC 载入不了。照 cctx、fys 等库已经改好的写法，改成 command("give yancao to " + ob->query("id"));（ob 是刚进房间的玩家）。

标准整理轮（2026-10-19，scripts/toolkit_fix.sh）

本库在新驱动下重扫，机械修复：35 个未使用的局部变量，1 处函数返回类型改成与基类一致。修后与 HEAD 逐个物件对比载入，没有退步。

找不到父类函数的 ::fn() 调用（2026-10-19，KB 06 §7.259）

/adm/etc/ranputthing.lpc、/d/city/npc/duguoyin.lpc、/questnew/ranputthing.lpc共 3 个物件里写着 ::create();、::init(); 这类调用，而它们继承的基类并没有这个函数。老驱动把这种调用当成什么都不做、返回 0 的空调用，本驱动在编译时报 Unable to find the inherited function 'init'.，整个物件载入不了。这些物件的错误只有这一种（其他错误清掉之后才露出来的），逐个删掉这些调用（return ::fn(...) 改成 return 0;）。（scripts/lpc_fix_dangling_super.py）

区域物件的零碎修正（2026-10-19，KB 06 §7.267–§7.271）

• adm/simul_efun/file.lpc 补上同系列（fy3dz、fy3xd）的 write_file2()：adm/daemons/brd.lpc 和订做衣服、打铁的 NPC 调用它，之前编不过
• d/wudujiao/npc/baichan.lpc 删掉 map_skill("baichan-hammer", 100);（第二个参数应是字符串；上面已有 set_skill，下面有 map_skill("hammer", "baichan-hammer")）

逐个物件与 HEAD 对比载入，没有退步。

RANK_D 对 0 的称呼（2026-10-19，KB 06 §7.273）

adm/daemons/rankd.lpc 的 query_respect()、query_rude()、query_self()、query_self_rude() 里 4 个函数开头补上 if (!objectp(ob)) ob = this_object();。NPC 在 create() 里写 "…" + RANK_D->query_respect(this_player()) + "…"（inquiry 对话、chat_msg），驱动自己跑 reset() 时没有人在场，this_player() 是 0。旧驱动里 0->query("age") 返回 0，称呼走默认分支；新驱动直接报 Bad argument 1 to EFUN call_other()，房间的 make_inventory() 吞掉错误，被打死的 NPC 再也刷新不回来。用 scripts/lpc_reach_census.py --walk 找到（xkx100 的武当弟子、一灯、汴梁书生…）；修在守护程序里，不动 1,800 个 NPC 文件。

技能书的基类（2026-10-19，KB 04 §6.15、KB 06 §7.262）

obj/book/*.lpc 的 40 本技能书（与 wqfy、xkyx3b 逐字节相同）都是 inherit BOOK;，本库没有这个宏，也没有它的基类，整批载入不了。补回：/std/item/book.lpc（宏 BOOK，xkyx3b 的同一份，经 wqfy 拷来），adm/simul_efun/atoi.lpc 补上基类要用的 itoa()、space()，include/ansi.h 里被注释掉的背景色宏 BWHT 恢复。41 个对象（40 本书和基类）全部载入，以前一个都不行，逐个与 HEAD 对比没有退步。（scripts/lpc_restore_bases.py --evidence --retry、scripts/lpc_add_sefun.py space）

补回丢失的标点（2026-10-19，KB 03 §4.6）

转换时丢掉的分号、逗号、引号或文件末尾的 }，共 3 个文件，修前 HEAD 载入失败。别的库里同路径的完好副本只比这里少几个标点，按它补回（括号里是那个库）：/d/menggu/tadi1.lpc（wqfy）。文件末尾少了 }（最后一个或两个字节丢了），补上：/u/wiz/u/legato/obj/skeleton.lpc（2 个）、/u/wiz/u/xxy/city2/npc/zhengmen.lpc（2 个）。修后与 HEAD 逐个对比载入：这 3 个物件全部载入，没有退步（scripts/lpc_restore_from_twin.py、lpc_fix_eaten_quote.py --comma、lpc_fix_eof_braces.py）。

旧驱动的类型宽松：缺省参数、无类型形参、数组初始化、class 变量（2026-10-19，KB 04 §6.19、§6.20）

原来的 MudOS 把调用少传的参数补 0、把没写类型的形参当 mixed、不检查变量的声明类型与初始值，也允许变量叫 class；这个驱动在编译时全部拒绝，整个物件载入失败。修了 2 个文件：/u/wiz/u/mool/kid.lpc、/u/wiz/u/xxy/room/heimuya/ren.lpc（scripts/lpc_round_files.py 的修复轮：调用处补写 0、无类型形参补 mixed、string X = ({...}) 补星号、void 函数的返回值，以及 class 变量改名 pick_class）。逐个文件在 HEAD 和改后载入对比：这些文件里 1 个原来载入失败、现在载入，没有退步。

x != "a" || x != "b" 和 x == "a" && x == "b"：同一个值的比较链永远成立/永远不成立（2026-10-20，KB 06 §7.282）

if (arg != "wan" || arg != "丸") return notify_fail(...) 对任何 arg 都为真，dir == "south" && dir == "north" 永远为假：条件下面的守卫要么永远拦、要么永远不拦。作者要的是“既不是 A 也不是 B”（&&）或“是 A 或 B”（||）。!= 链的 || 改成 &&（仅当条件下面是 return 的拒绝/守卫），== 链的 && 改成 ||（scripts/lpc_fix_cmp_chain.py）。改动：/d/xj/obj/lotus.lpc（2 处）。

member_array('/', str) 改为 strsrch(str, "/")：字符字面量的 member_array 按字节搜索（2026-10-20，KB 06 §7.283）

member_array('/', str) 里的 '/' 是整数，驱动按 UTF-8 字节搜索，返回字节偏移（带第三个参数时起点也是字节）；最新的 FluffOS（923cdbc0）在字符串里有非 ASCII 文字时，对每个调用点警告一次（Warning: member_array(int, string) searches BYTES）。而 str[0..idx-1] 按字符切，中文字符串会在错误的位置截断。改为 strsrch(str, "/")（找不到同样返回 -1，按字符计数；scripts/lpc_fix_member_array_char.py，JSON 编码器等少数几处手改）。改动：/adm/daemons/vrm_server.lpc（1 处）。

if (x = 常量)：条件里写成赋值的比较（2026-10-20，KB 06 §7.290）

if (arg = "button" || arg = "按钮")、if (!arg || arg = "") return 0;、else if (check = 0 && tempB != "")：把 == 写成了 =，条件的值就是那个常量，分支对所有非零数和任何字符串（连 "" 都算）永远成立、对 0 永远不成立，变量还被覆盖；驱动对此没有任何警告。每一处 = 改成 ==（scripts/lpc_fix_cond_assign.py，共 6 处）。改动：/d/city/mjlady3.lpc（3 处）、/d/wiz/sku/obj/mj.lpc（3 处）。

socket / dumpsocket 指令：dump_socket_status() 驱动里没有（2026-10-20，KB 06 §7.291）

这些巫师指令 write(dump_socket_status());，MudOS 的 efun 在 FluffOS 里不存在，指令编译不过。在模拟 efun 文件里补了 dump_socket_status()：用 socket_status() 排出同样的表（没有 sockets 包的构建，比如 WASM，只有表头；scripts/lpc_add_sefun.py dump_socket_status）。

登录密码超过 255 字节：crypt() 抛错（2026-10-20，KB 06 §7.294）

驱动的 crypt() 对超过 255 字节的密钥直接报错（Error in crypt(), check your salt），老驱动则按任意长度散列。在登录提示符下输入 256 字节以上的密码，密码检查那一行就抛错，连线停在一个裸的 > 提示符上，登录走不完（password 指令的旧密码检查同样）；用 300 个 x 当参数的扫描发现的。/adm/obj/simul_efun.lpc 里补了一个同名的 simul_efun crypt(key, salt)：只有超长的密钥才被截短（先截到 255 个字符，再逐字符截到 sizeof(to_buffer(key)) <= 255，限制按字节算，一个汉字占三个字节），再交给 efun::crypt()；普通密码原样散列，已存的密码散列不受影响（scripts/lpc_add_sefun.py crypt）。

get_config(__MUD_PORT__) 经旧版 runtime_config.h 得到字符串（2026-10-20，KB 06 §7.302）

本库的 include/runtime_config.h 是老 MudOS 的编号（整数设置从 15 起，本驱动从 256 起），get_config(__MUD_PORT__) 读到的是另一项设置（字符串），(int) 又只是编译期标注，于是 socket_bind(fd, LOCAL_PORT()) 报 Bad argument 2 to socket_bind()，版本守护进程 port = get_config(__MUD_PORT__) + VERSION_PORT 是字符串相加。头文件自己的注释写明 get_config(__MUD_PORT__) == __PORT__，所以代码里的这个调用（连同外面的 (int) / to_int()）改成了驱动的预定义 __PORT__（scripts/lpc_fix_mud_port.py）。改动：/adm/daemons/network/getconfig.h（1 处）、/include/getconfig.h（1 处）。

pettrainer 的 say_news() 读不到 /log/news 就抛错（2026-10-20，KB 06 §7.326）

str = read_file("/log/news"); flag = sizeof(explode(str, "\n"));：/log/news 是运行中写出来的新闻频道记录，全新的树里没有，explode(0) 报错；宠物训练师的 chat_msg 调它，每次都出错。读不到时 flag = 0，读出的 news 不是字符串就直接返回。三份相同。加载检查各 1/1。

register 指令编译不过：main() 里粘进了登录流程的代码（2026-10-20，KB 04 §6.25）

main() 开头有两段 if (yn == "") { ... input_to("confirm_id", ob); } 和 if (yn[0] != 'y' ...) { ... input_to("get_id", ob); }：yn 和 ob 是登录程序里 confirm_id() 的变量，这里没有声明（Undefined variable 'yn'），登记表指令从来没有加载过。删掉这两段，留下 display() 和对代号的提问（后面的 a()……g() 逐项提问本来就是这条流程）。加载检查：HEAD FAIL → tree PASS。

top 指令编译不过：o->query("spi")（2026-10-20，KB 04 §6.25）

score += o->query("spi") / 2 + ob->query("cps") / 2 + ... 里的 o 是 ob 少打了一个字母（Undefined variable 'o'）；wqfy 和 xhcii 的同一行是 ob->query("spi")。加载检查：HEAD FAIL → tree PASS。

NPC 的 chat() 在 chat_msg 为空数组时下标越界（2026-10-21，KB 06 §7.334）

NPC 基类的 chat() 是 if (arrayp(msg = query(...))) { if (random(100) < chance) { rnd = random(sizeof(msg)); if (stringp(msg[rnd])) ...，不检查列表是否为空。本库有 4 个 NPC 文件写成 set("chat_msg", ({ }))（或 chat_msg_combat）并带正的 chat_chance（例如 /d/fy/npc/oldgun.lpc、/d/xgchen/shounao.lpc），每次心跳有 chat_chance% 的概率 random(0) 为 0、msg[0] 报 Array index out of bounds（驱动捕获，心跳继续，但每次一条运行时错误），这个 NPC 在内存里就一直发生。在 random(100) < chance 前面加 sizeof(msg) &&（scripts/lpc_fix_chat_empty.py；if (random(100) > chance) return 0; 的写法是 !sizeof(msg) ||）。改动：/std/char/npc.lpc（共 1 个文件）。

存档里的 .c 文件名改成 .lpc（2026-10-21，KB 03 §4.2 第 9 项）

存档里写着 "/d/city/npc/wei.c" 这样的文件名；extract 时源文件已经全部改成 .lpc，而驱动对带扩展名的路径只认字面，new() / load_object() / move() 都找不到它，调用方多半把失败当成「没有这个东西」悄悄跳过。本库 38 个存档共 38 处按 scripts/lpc_fix_save_dotc.py 改成 .lpc（只改 .lpc 存在而 .c 不存在的字符串；玩家存档、备份、留言板、巫师目录和文件索引缓存不动）。改动：/p/residence/baiyun982545370.o（1 处）、/p/residence/bibi950532639.o（1 处）、/p/residence/bll951500300.o（1 处）、/p/residence/bll951517371.o（1 处）（等共 38 个存档）。

巫师命令不带参数时对 0 调用 efun（2026-10-21，KB 06 §7.336）

巫师（管理员）把命令敲成不带参数时，驱动把 0 传给字符串参数；下面的命令第一句就把这个 0 交给 efun，运行时报 Bad argument 1 to ...，这条指令从来没有用过。用 scripts/lpcc_bare.py 不登录、在一次 lpcc 求值里以 0 调用 main() 逐个确认后，由 scripts/lpc_fix_bare_arg.py 在第一处使用参数的语句前加守卫：文件自带 int help(object me) 的，守卫是 if (!arg) return help(me);（help() 定义在 main() 之后的 1 个文件补了原型 int help(object me);，否则编译报 Undefined function help）。改动（1 个文件，括号里是接住 0 的 efun）：/cmds/adm/sleepwith.lpc（present）。加载检查 HEAD 与工作树对照，没有退步。

doc：整份指令被编辑器弄坏，编译不了（2026-10-21，KB 04 §6.25）

管理员扫描里 doc（cmds/adm/doc.lpc，整理 NPC / 房间 / 物品档案的指令）报 Illegal character 0xef：源文件被中文编辑器改过又被排版器整理过，每个 -> 成了全角的 -＞（40 处），>、< 成了 ＞、＜，每个 || 成了 // 并吞掉了该行的其余部分：if (!arg) { // sscanf(arg, "%s %s", path, var) != 2){（用法检查丢了 sscanf，path 和 var 从来没有被赋值），if (prop[i] == "armor" //prop[i]=="damage")（continue 成了语法错误）。按原意恢复（||、被注释掉的 sscanf），MONEY_D->price_str()（同系没有这个守护进程）改成直接输出 value，删去 9 个没用的局部变量。编译没有警告，不带参数时打印用法。wqfy、zzfy、zzfy3 三个库的这个文件是同一份。

观礼台 d/city/under.lpc：wait_and_start() 里的 fan 没赋值，比武开不了（2026-10-21，KB 06 §7.346）

接受挑战（defend）两秒后 wait_and_start() 调用 fan->command("chat ...比武开始")，object fan; 声明了却没有 fan = load_object(__DIR__ "npc/fan");（dtxywzxzb 的同一份文件有），抛错之后两条 kill 和 call_out("finish_match", 300) 都不执行，比赛不开始。补上这一句（npc/fan.lpc、ground.lpc 本库都有）。

注册到不存在函数的 add_action（2026-10-21，KB 06 §7.349）

add_action("函数", "动词") 绑定的函数，在对象本身、它包含的头文件、它继承的父类和继承它的子类里都没有定义（scripts/lpc_scan_callout_targets.py 静态解析，子类用传递闭包复核，scripts/lpc_verify_handlers.py 运行时 function_exists() 复核）。驱动从玩家最新的一条动作往回找，遇到这样的句子就报 *Function for verb 'drop' not found.，同一个动词的其他处理函数不再尝试（不是最新的一条就悄悄跳过）。删除了 24 条注册（23 个文件，scripts/lpc_fix_dead_add_action.py），动词留给别的处理函数：/d/city/npc/duguoyin.lpc（do_vendor_list）、/d/dali/npc/pker.lpc（give_quest）、/d/death/npc/kkking.lpc（give_quest）、/d/fy/npc/kijy.lpc（give_quest）、/d/fy/npc/oldgun.lpc（give_quest）、/d/fy/npc/renkuang.lpc（give_quest）、/d/fy/npc/smith.lpc（do_buy、do_list）、/d/fy/npc/sqx.lpc（give_quest）、/d/fy/npc/tiandi.lpc（give_quest）、/d/huangshan/npc/xiangke.lpc（do_vendor_list） 等共 23 个文件。最多的函数名：give_quest 16 处、do_vendor_list 3 处、do_buy 3 处、do_list 1 处、do_jump 1 处。

调用了、但这个库自己的守护进程里没有的函数（2026-10-22，KB 06 §7.354）

MACRO->函数() 调用的函数在这个库的守护进程里不存在：call_other() 对不存在的函数不报错、只返回 0，所以这些功能（时间显示、昼夜事件、名字检查、价格文字……）一直没起作用。同一血统另一个库的同名守护进程里有这些函数，原文照抄到本库守护进程末尾（scripts/lpc_port_function.py；调用处传的参数个数、函数体用到的函数 / 全局变量 / 宏 / 字符串键都在本库里有，已核对；守护进程在 HEAD 和工作区都能加载，每个函数用固定参数各调用一次：HEAD 返回 0，现在返回正常的值）：/adm/daemons/chinesed.lpc ← aoxiangtianji：chinese_month()（1 处调用）。

(: 函数 :) 指针指向对象自己没有定义的函数（2026-10-22，KB 06 §7.355）

item_desc / inquiry / chat_msg_combat 里的 (: 函数名 :) 指向本对象（含包含文件、继承的父类）没有函数体的函数：编译不报错，用到时驱动报 *Undefined lfun pointer called: 函数名（look 某物、ask NPC about 话题 直接出错，战斗 / 闲聊里的指针在 NPC 的 chat() 里出错，这一拍心跳的后半段被跳过）。scripts/lpc_scan_callout_targets.py --kinds=lfun。 还剩 5 个函数没有补（chat_msg 3, chat_msg_combat 1, item_desc 1）：原型还在函数体丢了、或者根本没写过，没有可以直接补回的捐赠（别的库同一路径的文件里没有这个函数，或者它还要用本库没有的辅助函数、签名不同），不编造：/d/fy/npc/biaotou.lpc:be_protect、/d/fy/npc/no3.b1k.lpc:brag、/d/fy/npc/no3.lpc:brag、/d/meiyuan/npc/zhuge.lpc:random_action、/d/yihua/lianhuachi.lpc:look_CAI。

读不到的留言板："/clone/board/x"->foo(); 在 create() 里报错（2026-10-22，KB 06 §7.357）

xkx 系的房间在 setup() 之后写 "/clone/board/gaibang_r"->foo(); 来加载留言板（foo() 本来就没有定义，只为了触发加载）。这个库没有这个文件，call_other() 报 *call_other() couldn't find object '/clone/board/gaibang_r'，对象留在内存里、load_object() 却抛错：玩家第一次走进房间就看到错误，写在 create() 中间的还会切掉后半段。scripts/lpc_fix_missing_foo.py 把 2 条单独成行的语句（/clone/board/honghua_b、/clone/board/xiake_b；/d/menggu/ta13.lpc、/d/xiakedao/dating.lpc）注释掉并写明原因（别的房间里同样的语句本来就是作者自己注释掉的）。证明和 45 个库的整体加载检查见 KB 06 §7.357。

log/ 下的日志子目录没有人建：写日志的 write_file() / log_file() 报错（2026-10-22，KB 06 §7.358）

这个库的代码往 log/ 的 1 个子目录写日志：log/doc/register（cmds/usr/help.lpc）。压缩包、git clone 和站点包里都没有这些目录（本地启动没有写过它们，gen_keep_dirs.py 没记），write_file() 对不存在的目录抛 *Wrong permissions for opening file /log/dlog/money for append. "No such file or directory"，调用它的函数在写日志那一行中断，后面的奖励、标志、步骤都没有执行。本库的 log_file() 会自己建目录（实测，或读源码：nitan 系的 assure_file()），所以这里只记了直接用 write_file() / save_object() 写的目录。把 log/doc/register 记进 scripts/wasm_keep_dirs.txt（scripts/lpc_keep_log_dirs.py；站点和 pristine 树按它建空目录）。证明（scripts/lpcc_eval.py，在 /adm/daemons 这类目录下的求值对象里写 /log/zzdirprobe/...，新进程）见 KB 06 §7.358 的实测表。

(: 函数名 :) 指针指向本库没有的函数：只含这个指针的行加了注释（2026-10-22，KB 06 §7.355 补充）

inquiry / item_desc / chat_msg* 里的 (: 函数名 :) 指向本对象（含包含文件、继承的父类）没有函数体的函数：用到时（ask、look、NPC 闲聊）驱动抛 *Undefined lfun pointer called: 函数名，这一拍心跳的后半段被跳过。这些函数没有可以补回的来源（别的库同一路径的文件里没有，或者它还要用本库没有的东西），所以 5 个文件里 5 行只含这个指针的行（(: be_protect :), (: brag :) (: random_action :)）加了注释，原文保留并写明原因，这一项当作没有这个话题 / 物品 / 闲聊条目（scripts/lpc_fix_dangling_lfun.py；#include 的文件在库里缺失的对象、指针不在整行上的写法没动）。 前提（function_exists()）和加载检查（HEAD 与工作区对比，本库改动的 .lpc 文件）的结果见 KB 06 §7.355：没有一个回归。

汉字按字节切片的老写法没有换成按字符：改法名 "澄" + name[2..3]、辈分判断 name[0..1] == "澄" 等 2 处（2026-10-22，KB 06 §7.360）

原档按字节算字符串区间（一个汉字 2 字节）：name[2..3] 是名字的第二个字，name[0..1] == "澄" 问名字是不是以「澄」开头。转成 UTF-8 之后，这个驱动的区间和 strlen() 都按字符算（strlen("澄观") 是 2，"澄观"[0..1] 是两个字），老算术原样留着、意思全变：拜入少林 / 武当 / 峨嵋等派的玩家被改成错误的法名（二字名 张三 变成一个字的 澄，更长的名字取到第三个字）；各派掌门的辈分判断（name[0..1] == "澄"）永远不成立。

本库 2 处（2 个文件）：拼名字的赋值 1 处、与汉字常量比较 1 处。例：name[0..1] → name[0..0]（/d/qianfo/npc/master2.lpc:101）；things[0..13] == "天南地北双飞客" → things[0..6] == "天南地北双飞客"（/d/tieflag/magroom.lpc:42）。scripts/lpc_fix_gbk_slice.py 把下标按字符重算（社区的 UTF-8 版 xkx100utf8 里就是同样的写法，作对照）；没有改的：截断（name = name[0..7]，现在最多 8 个字，只是放宽）、ASCII 的区间（fname[0..3] == "room"、ctime(time())[4..19]）和上下文看不出是汉字的区间。证明（scripts/lpcc_eval.py，bixiecanyang；其余各库是同一份文本）：让澄字辈掌门 attempt_apprentice() 一个名叫「张三」的访客，HEAD 的法名是 慧，改后是 慧三；抽样 28 个库的 HEAD / 工作树加载对比 0 退步。工具不幂等（[0..3] 折半得 [0..1]，又是对齐区间），改过的文件记在 scripts/gbk_slice_converted.txt，不要重跑。

长度检查按字节写，旁边的提示按汉字写：18 处（2026-10-22，KB 06 §7.362）

老档的 strlen() 数字节（一个汉字 2 字节），转成 UTF-8 后它数字符：提示写「必须是 2 到 6 个中文字」的检查还是 strlen(name) < 4 || strlen(name) > 12，于是两三个字的房名、帮会名、群名、聊天室名、称号被提示自己说允许的那句话拒绝，上限却是提示的两倍；留言板的标题检查 i < 4 || i > 24 || i % 2 还会拒绝所有奇数个字的标题（post 你好吗）。本库 18 处（9 个文件）：if ((strlen(name) < 2) || (strlen(name) > 40)) {（/cmds/std/buildroom.lpc:68）；if ((strlen(name) < 2) || (strlen(name) > 12)) {（/cmds/std/editskill.lpc:385）。scripts/lpc_fix_strlen_gate.py 以提示为准：条件里每个 strlen 门槛都恰是提示所说数字的两倍才折半，存长度的变量（i = strlen(arg);）的门槛和 || i % 2 奇偶项、其后循环的 i % 2 == 0 && 一并处理；提示里有英文 / 字母 / 字节的不动。证明（scripts/lpcc_eval.py，bxsj 的留言板，求值对象继承 NPC 并重写 edit() 记录调用，command("post 标题")）：HEAD 拒绝「你好」和「你好吗」、放行 23 个字的标题；改后接受前两个、拒绝第三个；shenmo 的 named.lpc invalid_new_name("张三")：HEAD 回 名字只能是2-4个中文字。，改后回 0。抽样 24 + 15 个库的 HEAD / 工作树加载对比 0 退步。

按字节写的汉字循环、茶馆猜诗和同盟频道标题（2026-10-23，KB 06 §7.363）

老档按字节做「一个汉字」：while (i--) if (i % 2 == 0 && !is_chinese(name[i..<0])) 只检查偶数下标，奇数位上的拉丁字母（张a）会过；茶馆猜诗 mixup() 用 strlen(str)/2 再 j=j+j、str[k..k+1] 交换两字节；泥潭系 channeld 的 switch (strlen(fname)) case 4/6/8/10 和 strlen(family)==4/6 按字节长度给同盟 / 门派名加空格。本驱动的 strlen() 和区间都按字符算。scripts/lpc_fix_byte_char.py：去掉奇偶项、把区间改成 name[i..i]，按字符重写 mixup() 并把 strlen(first)>=14 折成 7，频道标题的 case / 门派长度改成字符数（对照 aoxiangtianji）。本库 汉字循环 / is_chinese 区间 17 个文件；茶馆猜诗 mixup 1 个文件。证明（scripts/lpcc_eval.py）：sjplii 的 check_legal_name 在 HEAD 接受「张a」、改后拒绝，两边都接受「张三」；xkx100 的 clubpoem->mixup("离离原上草") 得到「离离原草上」（真交换汉字）；nitan_san 两字同盟名得到 【 兄 弟 】。关键对象加载通过。

中文名 / 验证码仍按 GBK 取「两字节一字」（2026-10-23，KB 06 §7.373）

pname = pname[(random(sizeof(pname)) & 0xFFFE)..<1]; pname = pname[0..1]; 在按字符计下标时会取出两个汉字。scripts/lpc_fix_pname_slice.py 去掉掩码并把收尾改成 [0..0]（pname/pname2/sec_name/str/str1）。本库 4 处。证明：nt1 生成名长度 2–4；HEAD 常见 3–5。

无参数时命令对 0 的调用（2026-10-27，KB 06 §7.336 Also）

巫师/管理员命令在通过权限门之后仍会把 0 交给 sscanf / read_file / find_living 等。lpc_fix_bare_arg.py 在首次使用参数前加上 if (!arg)（或 help(me) / GetHelp()）。本库本批改动见该提交。

深度功能测试（2026-10-27，§6.25 ZHAOSHI 截断玩家技能）

玩家自创武功在 // ZHAOSHI : 0 处截断（EOF），无法加载。按同系完整 moming-blade 补一招 + valid_*/query_*/practice（2）：daemon/skill/dianwen-unarmed.lpc, daemon/skill/wujoysword-sword.lpc。

深度功能测试（2026-10-27，§6.25 tell_object 缺 me（自动放弃/愿否））

一参 tell_object("…") 在已有 me 时改为 tell_object(me, "…")（city.lpc）。证明：抽样 city / xiaolongnv 加载。

深度功能测试（2026-10-27，§6.25 one-arg message_vision）

message_vision("…") 补第二参（me / TP / this_player()）。

深度功能测试（2026-10-27，§6.10 未知 pragma save_binary/optimize）

删除本库中 FluffOS 不认的 #pragma save_binary / #pragma optimize… 行。