nx_onsite_research_fetch.nx source
↩ module page · 139 lines · 6352 B
1// nx_onsite_research_fetch.nx -- SOVEREIGN onsite-search WORLD-CLASS research fetch.
2//
3// Rung S-RESEARCH-1 of the reusable onsite-search arc: ground the S-class exceed
4// census on REAL sourced facts about the world-class open-source onsite search
5// engines (global Rule 4 -- no assumptions; every census cell must cite a real
6// fetch, not memory). Reuses the team's OWN proven sovereign HTTPS stack
7// (nx_tls13_client_session_run + nx_https_get_complete; no browser/node/curl),
8// validated against the real Mozilla CA store -- the IDENTICAL path proven live
9// in nx_library_fetch (the lib_* corpus fetched 2026-06-16).
10//
11// Fetches the fact-dense OPEN pages (en.wikipedia.org, static, NishiBot allowed,
12// canonical/parens-free titles -> no redirect) for the engines we measure against:
13// /wiki/Apache_Lucene -> srch_lucene.raw (Apache: the OSS IR library)
14// /wiki/Apache_Solr -> srch_solr.raw (Apache: Lucene-based search server)
15// /wiki/Elasticsearch -> srch_elastic.raw (the dominant OSS/commercial engine)
16// /wiki/Learning_to_rank -> srch_ltr.raw (the relevance frontier beyond BM25)
17// /wiki/Full-text_search -> srch_fulltext.raw (the onsite-search capability frame)
18//
19// expect_exit: 0
20// license_tier: ORIGINAL (clone of nx_library_fetch; only the URLs differ)
21
22import "nx_syscalls.nx"
23import "nx_csprng.nx"
24import "nx_x509_trust_store.nx"
25import "nx_trust_store_load_from_certdata.nx"
26import "nx_tls13_client_validate_certificate.nx"
27import "nx_tls13_client_session_run.nx"
28import "nx_https_url_for_fetch.nx"
29import "nx_https_url_connect.nx"
30import "nx_https_get.nx"
31import "nx_https_get_complete.nx"
32import "nx_http_response_parse.nx"
33const K_MAGIC_4194304: i64 = 4194304
34
35func sf_puts(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(1, s, n); return 0 }
36func sf_putn(v: i64) -> i64 {
37 let bb: *u8 = sys_mmap(28); var m: i64 = v
38 if m < 0 { m = 0 - m; sys_write(1, "-" as *u8, 1) }
39 let t: *u8 = sys_mmap(28); var k: i64 = 0
40 if m == 0 { t[0] = 48 as u8; k = 1 }
41 while m > 0 { t[k] = (48 + (m % 10)) as u8; m = m / 10; k = k + 1 }
42 var i: i64 = 0; while i < k { bb[i] = t[k - 1 - i]; i = i + 1 } sys_write(1, bb, k); return 0
43}
44func sf_strlen(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } return n }
45
46// Fetch one page over the sovereign TLS1.3 client, save the raw response.
47// Returns the HTTP status (200 ok), or a negative error code.
48func fetch_page(store: *TrustStore, full_url: *u8, path: *u8, path_len: i64, out_path: *u8) -> i64 {
49 sf_puts("--- fetch "); sf_puts(full_url); sf_puts("\n")
50
51 let cr: *u8 = sys_mmap(32)
52 var i: i64 = 0
53 nx_csprng_fill(cr, 32) // CWE-330 (debt 1785970852): was the constant 0xC0..0xDF
54 let priv: *u8 = sys_mmap(32)
55 i = 0
56 nx_csprng_fill(priv, 32) // CWE-330: the X25519 scalar was the constant 0xA0..0xBF on EVERY session
57
58 let url_p: *NxUrl = nx_url_new()
59 let target_raw: *u8 = sys_mmap(32)
60 let target: *NxHttpsTarget = target_raw as *NxHttpsTarget
61 target.url = url_p
62 target.port = 0
63 if nx_https_url_for_fetch(full_url, target) != NX_HTTPS_URL_OK { return 0 - 41 }
64
65 let fd_p: *i64 = sys_mmap(16) as *i64
66 if nx_https_url_connect(target, full_url, sys_now_realtime_sec(), fd_p) != NX_HTTPS_CONNECT_OK { return 0 - 42 }
67 let fd: i64 = *fd_p
68
69 let val_ctx_raw: *u8 = sys_mmap(64)
70 let val_ctx: *TlsValidationContext = val_ctx_raw as *TlsValidationContext
71 val_ctx.store = store
72 val_ctx.sni_host = full_url + target.url.host_off
73 val_ctx.sni_host_len = target.url.host_len
74 val_ctx.now_epoch = sys_now_realtime_sec()
75
76 let sr: i64 = nx_tls13_client_session_run(
77 fd, full_url + target.url.host_off, target.url.host_len,
78 cr, priv, val_ctx
79 )
80 if sr <= 0 { sys_close(fd); return 0 - (200 + (0 - sr)) }
81
82 let session: *Tls13ClientSession = sr as *Tls13ClientSession
83 let buf: *u8 = sys_mmap(K_MAGIC_4194304)
84 let gc: i64 = nx_https_get_complete(
85 session, fd, path, path_len,
86 full_url + target.url.host_off, target.url.host_len,
87 buf, K_MAGIC_4194304
88 )
89 sys_close(fd)
90 if gc < 0 { return 0 - (100 + (0 - gc)) }
91
92 let rs: *i64 = sys_mmap(128) as *i64
93 nx_http_response_parse(buf, gc, rs)
94 let status: i64 = rs[1]
95
96 let ofd: i64 = sys_openat_wr(out_path, 0x1A4)
97 if ofd <= 0 { return 0 - 70 }
98 sys_write(ofd, buf, gc)
99 sys_close(ofd)
100
101 sf_puts(" ST="); sf_putn(status); sf_puts(" GC="); sf_putn(gc); sf_puts(" -> "); sf_puts(out_path); sf_puts("\n")
102 return status
103}
104
105func main() -> i64 {
106 let cpath: *u8 = "/tmp/mozilla_certdata.txt\x00"
107 let r: i64 = nx_trust_store_load_from_certdata(cpath, 512, K_MAGIC_4194304)
108 if r <= 0 { sf_puts("SRCH-FETCH: certdata load failed\n"); return 1 }
109 let store: *TrustStore = r as *TrustStore
110 let n: i64 = trust_store_count(store)
111 if n < 50 { sf_puts("SRCH-FETCH: too few CAs\n"); return 3 }
112 sf_puts("CA="); sf_putn(n); sf_puts("\n")
113
114 var ok: i64 = 0
115
116 let u1: *u8 = "https://en.wikipedia.org/wiki/Apache_Lucene\x00"
117 let p1: *u8 = "/wiki/Apache_Lucene\x00"
118 if fetch_page(store, u1, p1, sf_strlen(p1), "knowledge/fetched/srch_lucene.raw\x00" as *u8) == 200 { ok = ok + 1 }
119
120 let u2: *u8 = "https://en.wikipedia.org/wiki/Apache_Solr\x00"
121 let p2: *u8 = "/wiki/Apache_Solr\x00"
122 if fetch_page(store, u2, p2, sf_strlen(p2), "knowledge/fetched/srch_solr.raw\x00" as *u8) == 200 { ok = ok + 1 }
123
124 let u3: *u8 = "https://en.wikipedia.org/wiki/Elasticsearch\x00"
125 let p3: *u8 = "/wiki/Elasticsearch\x00"
126 if fetch_page(store, u3, p3, sf_strlen(p3), "knowledge/fetched/srch_elastic.raw\x00" as *u8) == 200 { ok = ok + 1 }
127
128 let u4: *u8 = "https://en.wikipedia.org/wiki/Learning_to_rank\x00"
129 let p4: *u8 = "/wiki/Learning_to_rank\x00"
130 if fetch_page(store, u4, p4, sf_strlen(p4), "knowledge/fetched/srch_ltr.raw\x00" as *u8) == 200 { ok = ok + 1 }
131
132 let u5: *u8 = "https://en.wikipedia.org/wiki/Full-text_search\x00"
133 let p5: *u8 = "/wiki/Full-text_search\x00"
134 if fetch_page(store, u5, p5, sf_strlen(p5), "knowledge/fetched/srch_fulltext.raw\x00" as *u8) == 200 { ok = ok + 1 }
135
136 sf_puts("SRCH-SOVEREIGN-FETCH-OK pages_200="); sf_putn(ok); sf_puts("/5\n")
137 if ok < 1 { return 51 }
138 return 0
139}