wandering1/STM32F4-Base
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects 1 Releases Wiki Activity

test(host): 断言字面量统一纯 ASCII，30 文件 1535 行

Browse Source 
用户裁定「统一」。口径就是 AGENTS 铁律：字符串字面量必须纯 ASCII
（无 BOM 文件里奇数个汉字会吞掉结束引号，偶数才侥幸通过），中文只写注释。
四名队友按目录分四组并行、写范围互斥：net_socket 418 行（6 文件）、
console_log/sys_log_mcu/soe_store/cfg_file_store_mcu 268 行（4 文件）、
11 个存量套件 404 行 + adc_input 89 行、iec104 与 lftpd 与 NET 四套等 356 行
（9 文件）。runner 的 .py 文档串不在口径内（Python 源不经 ARMCC 代码页）。

两套口径必须分清：粗口径（按行切引号）普查 = 34 文件 / 1,552 行；严格口径
（状态机剥掉块/行注释与字符常量后再取字符串字面量）= 30 文件 / 1,535 行。
差值那 4 个文件的命中全在注释里（ns_peers.c、net_socket/stub/ns_rtos.h、
common/ch395f_fake.h、common/fatfs_fake.c）——按口径一处都不该改，改了等于
把中文从注释里搬走、正好违背口径。粗口径绝不能当交付判据，否则要么白改
注释、要么留下「这行为什么没改」的悬案。

Lead 独立终检（不采信自报）五条硬判据，结果 30 文件漂移 0：① 每文件 TC 号
集合与 HEAD 逐字相同 ② 断言宏调用点数不变 ③ 行数与 HEAD 完全相等（专抓
整文件回写把尾部含 main() 吃掉的截断，第 1 轮真发生过两次）④ 剥注释后代码区
非 ASCII 字符 = 0 ⑤ 字节卫生（BOM 与 HEAD 一致、无 CR、无 U+FFFD、单尾换行）。
另用「逐行把字面量整体替换成空串后与 HEAD 必须逐行相等」机械证明只改了字面量；
各套 --module 覆盖率与断言数逐字不变（net_socket 887/889·53/53、console_log
199/200·16/16、sys_log_mcu 324/326·13/13、adc_input 103/103、sys_clock 100/100）
→ 生产码零改动，基线不需重登。

闸拦下一处真实回归（我的验收器没抓到）：ns_s6_edges.c 翻译时把一个上一轮
已整改成平实描述的编号又写回了 CHECK 标签。原因很清楚：「只改字面量」并不
等于「只改无害内容」，标签里完全可以带进未登记族的编号形态；而 Lead 的自动
验收器只比对了 TC 形 id 集合，未覆盖 L/E/N/S/TRAP 等族。教训：批量改标签的
守卫必须加一条「新串不得出现任何已登记前缀族之外的编号形态」，且结构不变式
只覆盖自己想到的维度时不能替代真闸。
Lead 自查另清掉 126 处写坏的反斜杠（JS→Python 转义层数翻倍，把 markdown
反引号写成反斜杠+反引号），恰好发生在记录这条教训的那一节里。

文档：docs/测试说明/主机单测覆盖率口径.md §9 新增「第 2 轮」小节（两套口径对照
表、五条终检判据、六条手法含「转换说明符序列必须逐位相同」取代「禁止裸 %」、
三条协作纪律含先改板再派工与临时文件带组前缀、后续项 3 条）+ 变更记录一行。

后续项（本轮不做）：把剥注释词法扫描器收进 scripts 当门禁（需用户点头并先定
test/ 是否纳入 coding-standard）；iec104 的简写编号归族（牵 id_registry，
不能顺手做）；替换器加「编号形态白名单」校验。

gate: pre-commit run --all-files -> 39 hooks, 37 Passed; failed = mem-numbers +
mem-budget（受管快照与本树 .map 的漂移，属第 1 轮生产码改动的遗留，非本批
增量）。本批只碰 test/host/**，不进 Keil 编译面也不进模拟器 CMake，**RAM/Flash
增量 = 0**；实测水位 RW_IRAM1 57.4%、RW_CCM 87.9%，均低于 92% 红线。
id-check 0 台账外新账（可判定 84.63% > 基线 82.6%）；test-coverage 33/33 全绿；
cppcheck / coding-standard / bsp-err / layer-deps / doc-links 全 Passed。
按仓内分支约定 --no-verify 提交，台账 --write 回填留待主工作树合并后。
This commit is contained in:
Wanderingss
2026-10-24 20:57:46 +08:00
parent 5a806d8a26
commit 924ac3d428
31 changed files with 1626 additions and 1542 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/host/net_events/net_events_test.c
+34 -34
View File
@@ -217,16 +217,16 @@ static void tc_connect(void)
CHECK((p->remote_ip_arr[0] == 192u) && (p->remote_ip_arr[3] == 50u) &&
(p->remote_ip == ((192u) | (168u << 8) | (1u << 16) | (50u << 24))) &&
(p->remote_port == 51234u),
"TC-NETEV-001 对端六元组解析：ip 低字节在前拼 32 位、端口取第 5/6 字节");
"TC-NETEV-001 peer tuple parsing: ip assembled little-endian into 32 bits, port from bytes 5/6");
CHECK((p->state == NET_SOCK_STATE_ESTABLISHED) && (p->session_dead == 0u) &&
(p->lost_conn_seen == 0u) && (p->send_ready == 1u),
"TC-NETEV-002 监听口自接后转已连接并立刻可写");
"TC-NETEV-002 a listened socket becomes established after self-accept and is writable at once");
CHECK((g_chip.keepalive_calls == 1) && (g_chip.last_keepalive == 1u) &&
(g_chip.int_status_calls == 1),
"TC-NETEV-003 建链即开 keepalive 并清一次挂起位（读即清）");
"TC-NETEV-003 on connect, enable keepalive and clear the pending bit once (read clears)");
CHECK((g_ev_n == 1) && (g_ev_kind[0] == (int)NET_EVENT_CONNECTED) &&
(g_ev_fd[0] == 0) && (g_ev_arg[0] == (void *)&g_ev_token),
"TC-NETEV-004 事件带 fd 与调用方 arg 回调给应用");
"TC-NETEV-004 the event calls back into the app with fd and the caller arg");
reset_all();
p = mk(0, NET_SOCK_STATE_CONNECTING, 0u);
@@ -234,13 +234,13 @@ static void tc_connect(void)
handle_connect_event(p);
CHECK((p->state == NET_SOCK_STATE_ESTABLISHED) && (p->send_ready == 0u) &&
(g_chip.keepalive_calls == 1) && (g_ev_n == 1),
"TC-NETEV-005 主动连接分支同样转已连接，但**不动 send_ready**（与自接有别）");
"TC-NETEV-005 the active-connect branch also becomes established but does NOT touch send_ready (differs from self-accept)");
reset_all();
p = mk(0, NET_SOCK_STATE_CLOSED, 2404u);
handle_connect_event(p);
CHECK((g_ev_n == 0) && (p->state == NET_SOCK_STATE_CLOSED),
"TC-NETEV-006 其它状态收到 CONNECT 中断一律忽略（不许凭空造会话）");
"TC-NETEV-006 a CONNECT interrupt in any other state is ignored (never invent a session)");
}
/* ===================== DISCONNECT ===================== */
@@ -257,7 +257,7 @@ static void tc_disconnect(void)
handle_disconnect_event(p);
CHECK((p->state == NET_SOCK_STATE_ESTABLISHED) && (g_ev_n == 0) &&
(g_nev_error_count == 1),
"TC-NETEV-010 芯片说还在连接：忽略这个陈旧 DISCONNECT（不许拆活会话）");
"TC-NETEV-010 the chip still says connecting: ignore this stale DISCONNECT (never tear down a live session)");
reset_all();
p = mk(0, NET_SOCK_STATE_ESTABLISHED, 2404u);
@@ -267,7 +267,7 @@ static void tc_disconnect(void)
handle_disconnect_event(p);
CHECK((p->state == NET_SOCK_STATE_ESTABLISHED) && (p->lost_conn_seen == 1u) &&
(g_ev_n == 0) && (g_settle_n == 0),
"TC-NETEV-011 疑似误报（一条数据都没收就断）：延后交给对账器，不立即拆");
"TC-NETEV-011 suspected false alarm (dropped before any data): defer to the reconciler instead of tearing down");
reset_all();
p = mk(0, NET_SOCK_STATE_ESTABLISHED, 2404u);
@@ -280,7 +280,7 @@ static void tc_disconnect(void)
(p->last_peer_port == 51234u) &&
(p->state == NET_SOCK_STATE_CLOSED) &&
(g_ev_kind[0] == (int)NET_EVENT_DISCONNECTED),
"TC-NETEV-012 正常断开：优雅沉降 + 会话标死 + 记旧端口 + 发事件");
"TC-NETEV-012 normal disconnect: graceful settle + mark session dead + remember old port + post event");
reset_all();
p = mk(0, NET_SOCK_STATE_LISTENING, 2404u);
@@ -289,20 +289,20 @@ static void tc_disconnect(void)
handle_disconnect_event(p);
CHECK((p->state == NET_SOCK_STATE_LISTENING) && (g_ev_n == 0) &&
(g_nev_info_count == 1) && (g_nev_error_count == 0),
"TC-NETEV-013 监听还在：陈旧监听 DISCONNECT 只留痕，不报 ERROR");
"TC-NETEV-013 listen still alive: a stale listen DISCONNECT is only logged, not reported as ERROR");
reset_all();
p = mk(0, NET_SOCK_STATE_LISTENING, 2404u);
g_fake.status_byte1[0] = CH395F_TCP_CLOSED; /* 芯片说不在 LISTEN */
handle_disconnect_event(p);
CHECK((p->state == NET_SOCK_STATE_LISTENING) && (g_ev_n == 0) && (g_req_n == 0),
"TC-NETEV-013b 监听丢了也只留痕：重开交给对账器，事件层不越权");
"TC-NETEV-013b even a lost listener is only logged: reopening belongs to the reconciler, the event layer must not overreach");
reset_all();
p = mk(0, NET_SOCK_STATE_CONNECTING, 0u);
handle_disconnect_event(p);
CHECK((g_nev_error_count == 1) && (g_ev_kind[0] == (int)NET_EVENT_DISCONNECTED),
"TC-NETEV-014 非常规状态收到 DISCONNECT：报 ERROR 但仍要通知应用");
"TC-NETEV-014 DISCONNECT in an unusual state: report ERROR but still notify the application");
}
/* ===================== RECV / SEND / TIMEOUT ===================== */
@@ -316,21 +316,21 @@ static void tc_recv_send_timeout(void)
p = mk(0, NET_SOCK_STATE_ESTABLISHED, 2404u);
handle_recv_event(p);
CHECK((p->readable == 1u) && (g_ev_kind[0] == (int)NET_EVENT_DATA_RECEIVED),
"TC-NETEV-020 RECV 事件置可读并通知应用");
"TC-NETEV-020 a RECV event sets readable and notifies the application");
reset_all();
p = mk(0, NET_SOCK_STATE_ESTABLISHED, 2404u);
p->send_ready = 1u;
handle_send_event(p);
CHECK((g_ev_kind[0] == (int)NET_EVENT_SEND_COMPLETE) && (p->send_ready == 1u),
"TC-NETEV-021 SEND 完成只发事件，字段留给写路径自己管");
"TC-NETEV-021 SEND completion only posts the event; the fields stay owned by the write path");
reset_all();
p = mk(0, NET_SOCK_STATE_CONNECTING, 0u);
handle_timeout_event(p);
CHECK((p->state == NET_SOCK_STATE_CREATED) && (g_nev_error_count == 1) &&
(g_ev_kind[0] == (int)NET_EVENT_TIMEOUT),
"TC-NETEV-030 连接超时：退回 CREATED 并通知应用");
"TC-NETEV-030 connect timeout: fall back to CREATED and notify the application");
reset_all();
p = mk(0, NET_SOCK_STATE_ESTABLISHED, 2404u);
@@ -339,14 +339,14 @@ static void tc_recv_send_timeout(void)
CHECK((g_settle_level == (int)NET_LISTENER_SETTLE_QUARANTINE) &&
(p->released_abnormal == 1u) && (p->session_dead == 1u) &&
(p->last_peer_port == 777u) && (p->state == NET_SOCK_STATE_CLOSED),
"TC-NETEV-031 会话超时：隔离级沉降 + 标异常释放（net_pending 收尾据此报 CONNRESET）");
"TC-NETEV-031 session timeout: quarantine-level settle + mark released abnormally (net_pending reports CONNRESET from it)");
reset_all();
p = mk(0, NET_SOCK_STATE_LISTENING, 2404u);
handle_timeout_event(p);
CHECK((p->state == NET_SOCK_STATE_LISTENING) && (g_nev_error_count == 1) &&
(g_ev_kind[0] == (int)NET_EVENT_TIMEOUT),
"TC-NETEV-032 监听超时只留痕：不改状态，但仍要通知等待方");
"TC-NETEV-032 listen timeout is only logged: state unchanged but waiters are still notified");
}
/* ===================== 掉线自愈 ===================== */
@@ -362,7 +362,7 @@ static void tc_lost_probe(void)
g_fake.cached_tcp[0] = CH395F_TCP_CLOSED;
net_lost_connect_probe(p);
CHECK((p->lost_conn_seen == 0u) && (g_ev_n == 0),
"TC-NETEV-040 芯片不在已连接态：只清标记，不自作主张建会话");
"TC-NETEV-040 the chip is not established: only clear the flag, do not create a session on our own");
reset_all();
p = mk(0, NET_SOCK_STATE_CLOSED, 2404u);
@@ -372,7 +372,7 @@ static void tc_lost_probe(void)
ch395f_fake_set_peer(0, 10u, 0u, 0u, 9u, 51234u);
net_lost_connect_probe(p);
CHECK((p->lost_conn_seen == 0u) && (g_ev_n == 0),
"TC-NETEV-041 对端口与上次相同：认定是自己没看到 CONNECT，不自接第二遍");
"TC-NETEV-041 peer port identical to last time: assume we missed CONNECT, do not accept a second time");
/* 自愈的适用面是软件还写着 LISTENING、芯片其实已经连上——accept 分支只认
* LISTENING/CONNECTING。若软件态已被别处改成 CLOSED，本函数打完 INFO 后什么都
@@ -387,7 +387,7 @@ static void tc_lost_probe(void)
CHECK((g_ev_n == 1) && (g_ev_kind[0] == (int)NET_EVENT_CONNECTED) &&
(p->state == NET_SOCK_STATE_ESTABLISHED) && (p->remote_port == 55555u) &&
(p->lost_conn_seen == 0u),
"TC-NETEV-042 丢了 CONNECT 中断时自愈：补一次 accept，会话不至于永久失联");
"TC-NETEV-042 self-heal after a lost CONNECT interrupt: make up one accept so the session is not stranded forever");
reset_all();
p = mk(0, NET_SOCK_STATE_CLOSED, 2404u);
@@ -395,7 +395,7 @@ static void tc_lost_probe(void)
ch395f_fake_set_peer(0, 10u, 0u, 0u, 9u, 55555u);
net_lost_connect_probe(p);
CHECK((g_ev_n == 0) && (p->state == NET_SOCK_STATE_CLOSED),
"TC-NETEV-042b 现状：软件已 CLOSED 时自愈接不上（只留痕，不凭空造会话）");
"TC-NETEV-042b current state: self-heal cannot accept once the software side is CLOSED (logged only, no invented session)");
}
/* ===================== 周期对账 ===================== */
@@ -419,10 +419,10 @@ static void tc_reconcile(void)
infos = g_nev_info_count;
recon_step(1000u); /* priming：链路计数记基线 */
CHECK(g_nev_info_count == infos,
"TC-NETEV-051a 上电第一趟不因链路计数打日志（0xFFFF 哨兵只记基线）");
"TC-NETEV-051a the first pass after power-up logs nothing for the link counter (the 0xFFFF sentinel only records the baseline)");
c = g_chip.cached_calls;
recon_step(1000u); /* 同一时刻再调一次 */
CHECK(g_chip.cached_calls == c, "TC-NETEV-050 100ms 周期闸门：同一时刻不重复查芯片");
CHECK(g_chip.cached_calls == c, "TC-NETEV-050 100ms period gate: do not query the chip twice at the same instant");
reset_all();
p = mk(1, NET_SOCK_STATE_ESTABLISHED, 2404u);
@@ -434,17 +434,17 @@ static void tc_reconcile(void)
recon_step(1300u);
CHECK((p->state == NET_SOCK_STATE_ESTABLISHED) && (g_ev_n == 0) &&
(g_nev_info_count == infos),
"TC-NETEV-052 会话死亡要**连观 4 次**才动手：前三次保持原状");
"TC-NETEV-052 a dead session needs 4 consecutive observations before acting: the first three change nothing");
recon_step(1400u);
CHECK((p->state == NET_SOCK_STATE_CLOSED) && (p->session_dead == 1u) &&
(g_ev_kind[0] == (int)NET_EVENT_DISCONNECTED),
"TC-NETEV-052b 第 4 次稳定后走断开收尾");
"TC-NETEV-052b once the 4th observation is stable, run the disconnect teardown");
p->state = NET_SOCK_STATE_ESTABLISHED; /* 假装又活了，验证冷却 */
recon_step(1500u);
recon_step(1600u);
recon_step(1700u);
recon_step(1800u);
CHECK(g_ev_n == 1, "TC-NETEV-052c 动作后 1000ms 冷却内不再重复折腾（哪怕稳定计数够了）");
CHECK(g_ev_n == 1, "TC-NETEV-052c no repeat action inside the 1000ms cooldown after acting (even when the stable count is met)");
reset_all();
p = mk(2, NET_SOCK_STATE_ESTABLISHED, 2404u);
@@ -452,11 +452,11 @@ static void tc_reconcile(void)
g_fake.cached_tcp[2] = CH395F_TCP_LISTEN; /* 芯片自己重开监听了，却没发 DISCONNECT */
recon_step(2900u);
CHECK((p->state == NET_SOCK_STATE_ESTABLISHED) && (g_pcan_n == 0),
"TC-NETEV-053 芯片偷偷重开监听：需要 2 次稳定，第一次只累计");
"TC-NETEV-053 the chip reopened the listener behind our back: needs 2 stable passes, the first one only accumulates");
recon_step(3000u);
CHECK((p->state == NET_SOCK_STATE_LISTENING) && (g_pcan_n == 1) &&
(g_scan_n == 1) && (g_pcan_err == NET_ERR_CONNRESET),
"TC-NETEV-053b 第 2 次稳定：重同步成监听并**取消两条挂起**（否则任务永远等不到回）");
"TC-NETEV-053b second stable pass: resync to listen and cancel both pending waits (otherwise a task would wait forever)");
reset_all();
p = mk(3, NET_SOCK_STATE_LISTENING, 2404u);
@@ -466,7 +466,7 @@ static void tc_reconcile(void)
recon_step(4100u);
CHECK((p->state == NET_SOCK_STATE_ESTABLISHED) && (p->remote_port == 40000u) &&
(g_ev_kind[0] == (int)NET_EVENT_CONNECTED),
"TC-NETEV-054 连接完成但中断丢了：1 次即补 accept");
"TC-NETEV-054 connect completed but the interrupt was lost: a single pass is enough to make up the accept");
reset_all();
p = mk(4, NET_SOCK_STATE_LISTENING, 2404u);
@@ -476,7 +476,7 @@ static void tc_reconcile(void)
recon_step(5300u);
CHECK((g_req_n == 1) && (g_req_sock == 4) &&
(g_req_reason == (int)NET_LISTENER_REQ_RECON),
"TC-NETEV-055 监听丢了：2 次稳定后请求重开监听（不自已在事件层里开）");
"TC-NETEV-055 listener lost: request a reopen after 2 stable passes (the event layer does not open it itself)");
reset_all();
p = mk(5, NET_SOCK_STATE_CLOSED, 2404u);
@@ -487,7 +487,7 @@ static void tc_reconcile(void)
recon_step(6400u);
CHECK((p->state == NET_SOCK_STATE_LISTENING) && (p->last_peer_port == 999u) &&
(p->lost_conn_seen == 0u) && (g_req_n == 0),
"TC-NETEV-056 U11 重同步：软件 CLOSED 而芯片在 LISTEN -> 直接跟芯片，不多余重开");
"TC-NETEV-056 U11 resync: software CLOSED while the chip is LISTEN -> follow the chip, no redundant reopen");
reset_all();
p = mk(6, NET_SOCK_STATE_CLOSED, 2404u);
@@ -495,13 +495,13 @@ static void tc_reconcile(void)
recon_step(7500u);
recon_step(7600u);
CHECK((g_req_n == 1) && (g_req_sock == 6),
"TC-NETEV-057 软件 CLOSED 且芯片也不在监听：请求重开（重建交回监听层）");
"TC-NETEV-057 software CLOSED and the chip is not listening either: request a reopen (the listener layer rebuilds it)");
reset_all();
(void)memset(&g_tbl[7], 0, sizeof(g_tbl[7])); /* 槽位不在用 */
infos = g_nev_info_count;
recon_step(8700u);
CHECK(g_nev_info_count == infos, "TC-NETEV-058 不在用的槽位直接跳过（不产生任何动作）");
CHECK(g_nev_info_count == infos, "TC-NETEV-058 unused slots are skipped outright (no action at all)");
/* 链路日志：s_link_last 初值 0xFFFF 是哨兵——第一次看到计数只记基线、不打日志
* （否则每次上电都会冒一条"链路 up"的假事件）；此后计数一变就必须留痕。
@@ -511,7 +511,7 @@ static void tc_reconcile(void)
infos = g_nev_info_count;
recon_step(9800u);
CHECK(g_nev_info_count == (infos + 1),
"TC-NETEV-051b 链路变更计数一变就留痕（基线那趟已由 051a 验证不打）");
"TC-NETEV-051b any change of the link-change counter is logged (the baseline pass is already covered by 051a)");
}
int main(void)