nx_page_fetch_raw.nx source
↩ module page · 39 lines · 1854 B
1// nx_page_fetch_raw.nx -- fetch a page and write RAW HTML (no text-strip) to a
2// file, so we can inspect JS-rendered pages' embedded data (Next.js __NEXT_DATA__,
3// ld+json, etc.). Transient research input. expect_exit: 0.
4
5import "nx_syscalls.nx"
6import "nx_x509_trust_store.nx"
7import "nx_trust_store_load_from_certdata.nx"
8import "nx_https_fetch_follow.nx"
9const K_MAGIC_4194304: i64 = 4194304
10const K_MAGIC_8388608: i64 = 8388608
11
12func w(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(1, s, n); return 0 }
13func wn(v: i64) -> i64 {
14 let b: *u8 = sys_mmap(28); var m: i64 = v
15 if m < 0 { m = 0 - m; sys_write(1, "-" as *u8, 1) }
16 let t: *u8 = sys_mmap(28); var k: i64 = 0
17 if m == 0 { t[0] = 48 as u8; k = 1 }
18 while m > 0 { t[k] = (48 + (m % 10)) as u8; m = m / 10; k = k + 1 }
19 var i: i64 = 0; while i < k { b[i] = t[k - 1 - i]; i = i + 1 }
20 sys_write(1, b, k); return 0
21}
22
23func main() -> i64 {
24 let r: i64 = nx_trust_store_load_from_certdata("/tmp/mozilla_certdata.txt" as *u8, 300, K_MAGIC_4194304)
25 if r <= 0 { w("STORE-FAIL\n" as *u8); return 1 }
26 let store: *TrustStore = r as *TrustStore
27 let raw: *u8 = sys_mmap(K_MAGIC_8388608)
28 let st: *i64 = (sys_mmap(8)) as *i64
29 let n: i64 = nx_https_fetch_follow("https://thecadhub.com/blog/best-3d-printing-software/" as *u8, store, raw, K_MAGIC_8388608, 6, st)
30 w("status=" as *u8); wn(st[0]); w(" raw_bytes=" as *u8); wn(n); w("\n" as *u8)
31 if n <= 0 { return 1 }
32 let fd: i64 = sys_openat_wr("/tmp/cadhub_raw.html" as *u8, 420)
33 if fd < 0 { w("OPEN-FAIL\n" as *u8); return 1 }
34 var wr: i64 = 0
35 while wr < n { let c: i64 = sys_write(fd, raw + wr, n - wr); if c <= 0 { break } wr = wr + c }
36 sys_close(fd)
37 w("wrote " as *u8); wn(wr); w(" raw bytes -> /tmp/cadhub_raw.html\n" as *u8)
38 return 0
39}