code wiki / (root) / nx_paper_native.nx
`nx_paper_native.nx` source

↩ module page · 855 lines · 40567 B
1// nx_paper_native.nx -- PAPERS AS NATIVE NISHI DOCS, NOT PDF (operator 2026-08-13: "i dont want
2// the papers in 3rd party formats like pdf ... ingested into html or whatever from the first bit
3// up we can do in nishi that makes it native to nishi os and browser").
4//
5// NXPAP2 (2026-08-14) -- v1 was a readable DIGEST, measured NOT sota. Against the real LaTeXML
6// source of one paper (arXiv 2411.15205, 242,460 B) v1 preserved: 0 of 82 <math>, 0 of 8 <figure>,
7// 0 of 84 bibitems, and collapsed all 10 <h3> subsections into their parent section. It also
8// DECODED source entities (&lt; -> <) and then copied section text into HTML unescaped, so decoded
9// markup reached the page raw. NXPAP2 fixes all of it, and REFUSES rather than drop silently.
10//
11// FORMAT: line-oriented, greppable, sovereign. Header `KEY\tVALUE` lines, then typed blocks:
12//   .SEC <level>\t<id>\t<title>     sectioning, level 1..3, id = the source anchor
13//   .PARA <inline>                  one paragraph of SAFE inline html (see below)
14//   .MATH <id>\t<mathml>            display equation, MathML subtree VERBATIM
15//   .FIG <id>                       figure open   | .TBL <id>  table open
16//   .IMG <url>                      one image of the current figure (absolute url)
17//   .CAP <inline>                   caption of the current figure/table
18//   .ROW <cell>\t<cell>...          one table row, cells are inline
19//   .REF <key>\t<label>\t<text>     one parsed reference
20//
21// SAFE INLINE is the load-bearing invariant: source entities are kept VERBATIM (never decoded),
22// every tag is dropped except (a) <math>..</math> copied byte-for-byte so MathML survives to the
23// browser natively, and (b) fragment links rebuilt as anchors so citations and cross-references
24// stay real links. Nothing else reaches the output, so the emitted page cannot carry injected
25// markup -- escaping is by construction, not by a filter that must be remembered.
26//
27// FIDELITY IS ENFORCED, NOT HOPED FOR: the ingest counts <math>, <figure> and bibitems in the
28// SOURCE and in what it EMITTED, prints both, and REFUSES (exit 3) if it emitted fewer. A silent
29// drop -- the v1 defect -- is impossible by construction.
30//
31//   nx_paper_native abs <cvf-abstract.raw> <out.nxpap> [arxiv-html.raw]
32//   nx_paper_native view <in.nxpap> <out.html>
33// exit: 0 ok | 1 io | 2 usage | 3 refuse (named). license_tier: ORIGINAL. No hw writes (Rule 26).
34import "nx_syscalls.nx"
35const PN_MAGIC_1048576: i64 = 1048576
36
37// NO READ CAP EXISTS HERE BY DESIGN. Reads compose sys_read_file, which sizes its buffer from the
38// file itself (lseek END) and therefore cannot short-read; the estate fixed that defect class on
39// 2026-07-15 and this organ had been hand-rolling a capped reader beside it. A buffer ceiling that
40// has to be guessed is a defect generator either way -- too small truncates in silence, too large
41// wastes memory -- so the constant is removed, not raised. Output buffers are DERIVED from input
42// size (below), and mmap faults pages in on demand, so headroom costs address space, not memory.
43const PN_ARXBASE: *u8 = "https://arxiv.org/html/"
44
45func pn_len(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } return n }
46func pn_puts(s: *u8) -> i64 { sys_write(1, s, pn_len(s)); return 0 }
47func pn_num(v: i64) -> i64 {
  if v == 0 { sys_write(1, "0" as *u8, 1); return 0 }
  var m: i64 = v
  let t: *u8 = sys_mmap(32)
  var k: i64 = 0
  while m > 0 { t[k] = (48 + (m % 10)) as u8; m = m / 10; k = k + 1 }
  let o: *u8 = sys_mmap(32)
  var i: i64 = 0
  while i < k { o[i] = t[k - 1 - i]; i = i + 1 }
  sys_write(1, o, k)
  return 0
58}
59func pn_app(d: *u8, off: i64, s: *u8) -> i64 {
  var i: i64 = 0
  var o: i64 = off
  while s[i] != (0 as u8) { d[o] = s[i]; o = o + 1; i = i + 1 }
  return o
64}
65func pn_appn(d: *u8, off: i64, src: *u8, a: i64, b: i64) -> i64 {
  var o: i64 = off
  var i: i64 = a
  while i < b { d[o] = src[i]; o = o + 1; i = i + 1 }
  return o
70}
71func pn_find(src: *u8, a: i64, b: i64, pat: *u8) -> i64 {
  let pl: i64 = pn_len(pat)
  if pl == 0 { return 0 - 1 }
  var i: i64 = a
  var hit: i64 = 0 - 1
  var run: i64 = 1
  while run == 1 {
      if i + pl > b { run = 0 } else {
          var j: i64 = 0
          var ok: i64 = 1
          var scan: i64 = 1
          while scan == 1 {
              if j >= pl { scan = 0 } else {
                  if src[i + j] != pat[j] { ok = 0; scan = 0 } else { j = j + 1 }
              }
          }
          if ok == 1 { hit = i; run = 0 } else { i = i + 1 }
      }
  }
  return hit
91}
92func pn_count(src: *u8, a: i64, b: i64, pat: *u8) -> i64 {
  var c: i64 = 0
  var p: i64 = a
  var run: i64 = 1
  while run == 1 {
      let h: i64 = pn_find(src, p, b, pat)
      if h < 0 { run = 0 } else { c = c + 1; p = h + 1 }
  }
  return c
101}
102func pn_num_to(d: *u8, off: i64, v: i64) -> i64 {
  var o: i64 = off
  if v == 0 { d[o] = 48 as u8; return o + 1 }
  var m: i64 = v
  let t: *u8 = sys_mmap(32)
  var k: i64 = 0
  while m > 0 { t[k] = (48 + (m % 10)) as u8; m = m / 10; k = k + 1 }
  while k > 0 { k = k - 1; d[o] = t[k]; o = o + 1 }
  return o
111}
112// start of the value of `attr` (which must include the opening quote) inside src[a..b); -1 if none
113func pn_attr_s(src: *u8, a: i64, b: i64, attr: *u8) -> i64 {
  let p: i64 = pn_find(src, a, b, attr)
  if p < 0 { return 0 - 1 }
  return p + pn_len(attr)
117}
118func pn_write(path: *u8, buf: *u8, n: i64) -> i64 {
  let fd: i64 = sys_openat_wr(path, 0x1a4)
  if fd < 0 { return 0 - 1 }
  sys_write(fd, buf, n)
  sys_close(fd)
  return 0
124}
125
126// ---- text: drop tags, KEEP entities verbatim, collapse whitespace ----
127func pn_txt(d: *u8, off: i64, src: *u8, a: i64, b: i64) -> i64 {
  var o: i64 = off
  var i: i64 = a
  var sp: i64 = 1
  while i < b {
      let c: i64 = src[i] as i64
      if c == 60 {
          let te: i64 = pn_find(src, i, b, ">" as *u8)
          if te < 0 { i = b } else { i = te + 1 }
      } else {
          var ch: i64 = c
          if ch == 9 { ch = 32 }
          if ch == 10 { ch = 32 }
          if ch == 13 { ch = 32 }
          if ch == 32 {
              if sp == 0 { d[o] = 32 as u8; o = o + 1; sp = 1 }
          } else {
              d[o] = ch as u8
              o = o + 1
              sp = 0
          }
          i = i + 1
      }
  }
  return o
152}
153// ---- inline: safe html. entities verbatim, <math> verbatim, fragment links rebuilt,
154// every other tag dropped. cnt[0] += maths copied. THIS is the escaping invariant. ----
155func pn_inline(d: *u8, off: i64, src: *u8, a: i64, b: i64, cnt: *i64) -> i64 {
  var o: i64 = off
  var i: i64 = a
  var sp: i64 = 1
  while i < b {
      let c: i64 = src[i] as i64
      if c == 60 {
          var done: i64 = 0
          if pn_find(src, i, b, "<math" as *u8) == i {
              let me: i64 = pn_find(src, i, b, "</math>" as *u8)
              if me >= 0 {
                  let end: i64 = me + 7
                  var k: i64 = i
                  while k < end {
                      var ch2: i64 = src[k] as i64
                      if ch2 == 10 { ch2 = 32 }
                      if ch2 == 13 { ch2 = 32 }
                      if ch2 == 9 { ch2 = 32 }
                      d[o] = ch2 as u8
                      o = o + 1
                      k = k + 1
                  }
                  cnt[0] = cnt[0] + 1
                  sp = 0
                  i = end
                  done = 1
              }
          }
          if done == 0 {
              if pn_find(src, i, b, "<a " as *u8) == i {
                  let te: i64 = pn_find(src, i, b, ">" as *u8)
                  if te > 0 {
                      let hs: i64 = pn_attr_s(src, i, te, "href=\"" as *u8)
                      if hs > 0 {
                          let he: i64 = pn_find(src, hs, te, "\"" as *u8)
                          if he > hs {
                              if src[hs] == (35 as u8) {
                                  let ae: i64 = pn_find(src, te, b, "</a>" as *u8)
                                  if ae > te {
                                      o = pn_app(d, o, "<a href=\"" as *u8)
                                      o = pn_appn(d, o, src, hs, he)
                                      o = pn_app(d, o, "\">" as *u8)
                                      o = pn_txt(d, o, src, te + 1, ae)
                                      o = pn_app(d, o, "</a>" as *u8)
                                      sp = 0
                                      i = ae + 4
                                      done = 1
                                  }
                              }
                          }
                      }
                  }
              }
          }
          if done == 0 {
              let te2: i64 = pn_find(src, i, b, ">" as *u8)
              if te2 < 0 { i = b } else { i = te2 + 1 }
          }
      } else {
          var ch: i64 = c
          if ch == 9 { ch = 32 }
          if ch == 10 { ch = 32 }
          if ch == 13 { ch = 32 }
          if ch == 32 {
              if sp == 0 { d[o] = 32 as u8; o = o + 1; sp = 1 }
          } else {
              d[o] = ch as u8
              o = o + 1
              sp = 0
          }
          i = i + 1
      }
  }
  return o
229}
230
231// ---- one figure/table region [a..b) -> .FIG/.TBL + .IMG* + .CAP (+ .ROW* for tables) ----
232func pn_figure(d: *u8, off: i64, src: *u8, a: i64, b: i64, istbl: i64, cnt: *i64) -> i64 {
  var o: i64 = off
  let ids: i64 = pn_attr_s(src, a, b, "id=\"" as *u8)
  if istbl == 1 { o = pn_app(d, o, ".TBL " as *u8) } else { o = pn_app(d, o, ".FIG " as *u8) }
  if ids > 0 {
      let ide: i64 = pn_find(src, ids, b, "\"" as *u8)
      if ide > ids { o = pn_appn(d, o, src, ids, ide) }
  }
  o = pn_app(d, o, "\n" as *u8)
  // images -- absolute so the document is self-describing
  var p: i64 = a
  var run: i64 = 1
  while run == 1 {
      let im: i64 = pn_find(src, p, b, "<img " as *u8)
      if im < 0 { run = 0 } else {
          let ss: i64 = pn_attr_s(src, im, b, "src=\"" as *u8)
          if ss > 0 {
              let se: i64 = pn_find(src, ss, b, "\"" as *u8)
              if se > ss {
                  o = pn_app(d, o, ".IMG " as *u8)
                  if pn_find(src, ss, se, "http" as *u8) != ss { o = pn_app(d, o, PN_ARXBASE) }
                  o = pn_appn(d, o, src, ss, se)
                  o = pn_app(d, o, "\n" as *u8)
              }
          }
          p = im + 5
      }
  }
  // caption
  let cs: i64 = pn_find(src, a, b, "<figcaption" as *u8)
  if cs >= 0 {
      let ce: i64 = pn_find(src, cs, b, "</figcaption>" as *u8)
      if ce > cs {
          let cb: i64 = pn_find(src, cs, ce, ">" as *u8)
          if cb > cs {
              o = pn_app(d, o, ".CAP " as *u8)
              o = pn_inline(d, o, src, cb + 1, ce, cnt)
              o = pn_app(d, o, "\n" as *u8)
          }
      }
  }
  // table rows -- cells carry math, so they go through pn_inline or the count is short
  if istbl == 1 {
      var rp: i64 = a
      var rr: i64 = 1
      while rr == 1 {
          let tr: i64 = pn_find(src, rp, b, "<tr " as *u8)
          if tr < 0 { rr = 0 } else {
              var tre: i64 = pn_find(src, tr, b, "</tr>" as *u8)
              if tre < 0 { tre = b }
              o = pn_app(d, o, ".ROW " as *u8)
              var cp: i64 = tr
              var nc: i64 = 0
              var cr: i64 = 1
              while cr == 1 {
                  var cell: i64 = pn_find(src, cp, tre, "<td " as *u8)
                  let cellh: i64 = pn_find(src, cp, tre, "<th " as *u8)
                  if cell < 0 { cell = cellh } else {
                      if cellh >= 0 { if cellh < cell { cell = cellh } }
                  }
                  if cell < 0 { cr = 0 } else {
                      let cb2: i64 = pn_find(src, cell, tre, ">" as *u8)
                      var cend: i64 = pn_find(src, cell, tre, "</t" as *u8)
                      if cend < 0 { cend = tre }
                      if cb2 > cell {
                          if nc > 0 { o = pn_app(d, o, "\t" as *u8) }
                          o = pn_inline(d, o, src, cb2 + 1, cend, cnt)
                          nc = nc + 1
                      }
                      cp = cend + 3
                  }
              }
              o = pn_app(d, o, "\n" as *u8)
              rp = tre + 5
          }
      }
  }
  return o
310}
311
312// ---- the bibliography region -> one .REF per bibitem ----
313// the refnum span repeats the authors that the bibblocks also carry -- taking ONLY the bibblocks
314// is what kills the v1 duplication ("Achlioptas et al. 2018; Achlioptas et al. 2018").
315func pn_refs(d: *u8, off: i64, src: *u8, a: i64, b: i64, nref: *i64) -> i64 {
  var o: i64 = off
  var p: i64 = a
  var run: i64 = 1
  while run == 1 {
      // ANCHOR ON THE ELEMENT, THEN TEST IT -- never scan backwards a guessed distance.
      // This used to find class="ltx_bibitem" and then look BACK a fixed 40 bytes for the id,
      // which is a magic number wearing arithmetic: one bibitem whose opening tag is longer
      // than the window and the reference key comes out wrong or empty, silently. Finding the
      // <li first means the id is read FORWARD inside the tag and no window exists to guess.
      let li: i64 = pn_find(src, p, b, "<li " as *u8)
      if li < 0 { run = 0 } else {
          let lt: i64 = pn_find(src, li, b, ">" as *u8)
          var le: i64 = pn_find(src, li, b, "</li>" as *u8)
          if le < 0 { le = b }
          var isbib: i64 = 0
          if lt > li { if pn_find(src, li, lt, "ltx_bibitem" as *u8) > 0 { isbib = 1 } }
          if isbib == 1 {
              o = pn_app(d, o, ".REF " as *u8)
              let ks: i64 = pn_attr_s(src, li, lt, "id=\"" as *u8)
              if ks > 0 {
                  let ke: i64 = pn_find(src, ks, lt, "\"" as *u8)
                  if ke > ks { o = pn_appn(d, o, src, ks, ke) }
              }
              o = pn_app(d, o, "\t" as *u8)
              // label: the refnum span
              let ts: i64 = pn_find(src, li, le, "ltx_tag_bibitem" as *u8)
              var afterlabel: i64 = li
              if ts >= 0 {
                  let t1: i64 = pn_find(src, ts, le, ">" as *u8)
                  if t1 > 0 {
                      let t2: i64 = pn_find(src, t1, le, "</span></span>" as *u8)
                      if t2 > t1 {
                          o = pn_txt(d, o, src, t1 + 1, t2)
                          afterlabel = t2 + 14
                      }
                  }
              }
              o = pn_app(d, o, "\t" as *u8)
              // text: the bibblocks only -- the refnum span above repeats the authors these
              // carry, and emitting both is exactly the v1 duplication defect
              var q: i64 = afterlabel
              var qr: i64 = 1
              while qr == 1 {
                  let bb: i64 = pn_find(src, q, le, "class=\"ltx_bibblock\"" as *u8)
                  if bb < 0 { qr = 0 } else {
                      let b1: i64 = pn_find(src, bb, le, ">" as *u8)
                      var b2: i64 = pn_find(src, bb, le, "</span>\n</span>" as *u8)
                      if b2 < 0 { b2 = pn_find(src, bb, le, "</span></span>" as *u8) }
                      if b2 < 0 { b2 = le }
                      if b1 > 0 { o = pn_txt(d, o, src, b1 + 1, b2) }
                      q = b2 + 7
                  }
              }
              o = pn_app(d, o, "\n" as *u8)
              nref[0] = nref[0] + 1
          }
          p = le + 5
      }
  }
  return o
376}
377
378// the first structural block. Skips the abstract (already taken from the CVF metadata) while
379// keeping the teaser figure that precedes section 1. The fidelity counts use the SAME start, so
380// the denominator is exactly the region that was walked -- an aggregate assertion bound to a
381// wider denominator than its subject is how a conversion "passes" while dropping content.
382func pn_bodystart(src: *u8, n: i64) -> i64 {
  var start: i64 = pn_find(src, 0, n, "<section id=\"" as *u8)
  let f0: i64 = pn_find(src, 0, n, "<figure id=\"" as *u8)
  if f0 >= 0 { if start < 0 { start = f0 } else { if f0 < start { start = f0 } } }
  return start
387}
388// ---- the LaTeXML body walk: emit typed blocks in document order ----
389func pn_body(d: *u8, off: i64, src: *u8, n: i64, cnt: *i64, nfig: *i64, nref: *i64, nsec: *i64) -> i64 {
  var o: i64 = off
  let start: i64 = pn_bodystart(src, n)
  if start < 0 { return o }
  var p: i64 = start
  var run: i64 = 1
  while run == 1 {
      let sec: i64 = pn_find(src, p, n, "<section id=\"" as *u8)
      let fig: i64 = pn_find(src, p, n, "<figure id=\"" as *u8)
      let par: i64 = pn_find(src, p, n, "<p id=\"" as *u8)
      let eqt: i64 = pn_find(src, p, n, "<table id=\"" as *u8)
      // earliest marker wins
      var best: i64 = 0 - 1
      var kind: i64 = 0
      if sec >= 0 { best = sec; kind = 1 }
      if fig >= 0 { if best < 0 { best = fig; kind = 2 } else { if fig < best { best = fig; kind = 2 } } }
      if par >= 0 { if best < 0 { best = par; kind = 3 } else { if par < best { best = par; kind = 3 } } }
      if eqt >= 0 { if best < 0 { best = eqt; kind = 4 } else { if eqt < best { best = eqt; kind = 4 } } }
      if best < 0 { run = 0 } else {
          let te: i64 = pn_find(src, best, n, ">" as *u8)
          if te < 0 { run = 0 } else {
              if kind == 1 {
                  // bibliography section -> references, not prose
                  if pn_find(src, best, te, "ltx_bibliography" as *u8) > 0 {
                      o = pn_refs(d, o, src, best, n, nref)
                      run = 0
                  } else {
                      var lvl: i64 = 1
                      if pn_find(src, best, te, "ltx_subsubsection" as *u8) > 0 { lvl = 3 } else {
                          if pn_find(src, best, te, "ltx_subsection" as *u8) > 0 { lvl = 2 }
                      }
                      let ids: i64 = pn_attr_s(src, best, te, "id=\"" as *u8)
                      // the heading that opens this section
                      var hs: i64 = pn_find(src, te, n, "<h" as *u8)
                      if hs >= 0 {
                          let h1: i64 = pn_find(src, hs, n, ">" as *u8)
                          var h2e: i64 = pn_find(src, hs, n, "</h" as *u8)
                          if h2e < 0 { h2e = n }
                          if h1 > 0 {
                              o = pn_app(d, o, ".SEC " as *u8)
                              o = pn_num_to(d, o, lvl)
                              o = pn_app(d, o, "\t" as *u8)
                              if ids > 0 {
                                  let ide: i64 = pn_find(src, ids, te, "\"" as *u8)
                                  if ide > ids { o = pn_appn(d, o, src, ids, ide) }
                              }
                              o = pn_app(d, o, "\t" as *u8)
                              o = pn_txt(d, o, src, h1 + 1, h2e)
                              o = pn_app(d, o, "\n" as *u8)
                              nsec[0] = nsec[0] + 1
                              p = h2e
                          } else { p = te + 1 }
                      } else { p = te + 1 }
                  }
              }
              if kind == 2 {
                  var fe: i64 = pn_find(src, best, n, "</figure>" as *u8)
                  if fe < 0 { fe = n }
                  var istbl: i64 = 0
                  if pn_find(src, best, te, "ltx_table" as *u8) > 0 { istbl = 1 }
                  o = pn_figure(d, o, src, best, fe, istbl, cnt)
                  nfig[0] = nfig[0] + 1
                  p = fe + 9
              }
              if kind == 3 {
                  var pe: i64 = pn_find(src, best, n, "</p>" as *u8)
                  if pe < 0 { pe = n }
                  o = pn_app(d, o, ".PARA " as *u8)
                  o = pn_inline(d, o, src, te + 1, pe, cnt)
                  o = pn_app(d, o, "\n" as *u8)
                  p = pe + 4
              }
              if kind == 4 {
                  var qe: i64 = pn_find(src, best, n, "</table>" as *u8)
                  if qe < 0 { qe = n }
                  if pn_find(src, best, te, "ltx_eqn_table" as *u8) > 0 {
                      let ids2: i64 = pn_attr_s(src, best, te, "id=\"" as *u8)
                      var mp: i64 = best
                      var mr: i64 = 1
                      while mr == 1 {
                          let m0: i64 = pn_find(src, mp, qe, "<math" as *u8)
                          if m0 < 0 { mr = 0 } else {
                              let m1: i64 = pn_find(src, m0, qe, "</math>" as *u8)
                              if m1 < 0 { mr = 0 } else {
                                  o = pn_app(d, o, ".MATH " as *u8)
                                  if ids2 > 0 {
                                      let ide2: i64 = pn_find(src, ids2, te, "\"" as *u8)
                                      if ide2 > ids2 { o = pn_appn(d, o, src, ids2, ide2) }
                                  }
                                  o = pn_app(d, o, "\t" as *u8)
                                  var k2: i64 = m0
                                  let end2: i64 = m1 + 7
                                  while k2 < end2 {
                                      var c3: i64 = src[k2] as i64
                                      if c3 == 10 { c3 = 32 }
                                      if c3 == 13 { c3 = 32 }
                                      if c3 == 9 { c3 = 32 }
                                      d[o] = c3 as u8
                                      o = o + 1
                                      k2 = k2 + 1
                                  }
                                  o = pn_app(d, o, "\n" as *u8)
                                  cnt[0] = cnt[0] + 1
                                  mp = end2
                              }
                          }
                      }
                  }
                  p = qe + 8
              }
              if p <= best { p = best + 1 }
              if p >= n { run = 0 }
          }
      }
  }
  return o
505}
506
507// ---- CVF abstract-page metadata ----
508func pn_meta(src: *u8, a: i64, b: i64, key: *u8, r: *i64) -> i64 {
  let kp: i64 = pn_find(src, a, b, key)
  if kp < 0 { return 0 - 1 }
  let cp: i64 = pn_find(src, kp, b, "content=" as *u8)
  if cp < 0 { return 0 - 1 }
  let vs: i64 = cp + 9
  let ve: i64 = pn_find(src, vs, b, "\"" as *u8)
  if ve < 0 { return 0 - 1 }
  r[0] = vs
  r[1] = ve
  r[2] = ve
  return 0
520}
521func pn_hdr(out: *u8, off: i64, fb: *u8, n: i64, key: *u8, label: *u8, r: *i64) -> i64 {
  var o: i64 = off
  if pn_meta(fb, 0, n, key, r) == 0 {
      o = pn_app(out, o, label)
      o = pn_txt(out, o, fb, r[0], r[1])
      o = pn_app(out, o, "\n" as *u8)
  }
  return o
529}
530
531// ---- INGEST: CVF abstract page (+ optional arXiv LaTeXML full text) -> NXPAP2 ----
532func pn_ingest(inp: *u8, outp: *u8, axp: *u8) -> i64 {
  let ln: *i64 = sys_mmap(16) as *i64
  let fb: *u8 = sys_read_file(inp, ln)
  if (fb as i64) == 0 { pn_puts("PAPER-NATIVE REFUSE: cannot read source\n" as *u8); return 3 }
  let n: i64 = ln[0]
  if n <= 0 { pn_puts("PAPER-NATIVE REFUSE: cannot read source\n" as *u8); return 3 }
  // the full text is read UP FRONT so the output buffer can be sized from the REAL inputs
  var ab: *u8 = 0 as *u8
  var an: i64 = 0
  if (axp as i64) != 0 {
      let l2: *i64 = sys_mmap(16) as *i64
      ab = sys_read_file(axp, l2)
      if (ab as i64) != 0 { an = l2[0] }
  }
  // DERIVED from the real inputs, never a fixed ceiling. Every emit path copies a SUBSET of the
  // source plus bounded per-block markup, so output cannot approach 8x input; the reservation is
  // deliberately generous because mmap faults pages in on demand, so headroom costs address
  // space and not resident memory. outcap is kept because an unchecked bound is worse than a
  // magic number -- the tripwire below turns a violated derivation into a loud refusal.
  let outcap: i64 = (n + an) * 8 + PN_MAGIC_1048576
  let out: *u8 = sys_mmap(outcap)
  let r: *i64 = sys_mmap(32) as *i64
  var o: i64 = 0
  o = pn_app(out, o, "NXPAP2\n" as *u8)
  if pn_meta(fb, 0, n, "citation_title" as *u8, r) != 0 { pn_puts("PAPER-NATIVE REFUSE: no citation_title (broken parse, not an empty paper)\n" as *u8); return 3 }
  o = pn_app(out, o, "TITLE\t" as *u8)
  o = pn_txt(out, o, fb, r[0], r[1])
  o = pn_app(out, o, "\n" as *u8)
  var scan: i64 = 0
  var nau: i64 = 0
  var more: i64 = 1
  while more == 1 {
      if pn_meta(fb, scan, n, "citation_author" as *u8, r) != 0 { more = 0 } else {
          o = pn_app(out, o, "AUTHOR\t" as *u8)
          o = pn_txt(out, o, fb, r[0], r[1])
          o = pn_app(out, o, "\n" as *u8)
          scan = r[2]
          nau = nau + 1
      }
  }
  o = pn_hdr(out, o, fb, n, "citation_publication_date" as *u8, "YEAR\t" as *u8, r)
  o = pn_hdr(out, o, fb, n, "citation_conference_title" as *u8, "VENUE\t" as *u8, r)
  o = pn_hdr(out, o, fb, n, "citation_pdf_url" as *u8, "SOURCE\t" as *u8, r)
  let ax: i64 = pn_find(fb, 0, n, "arxiv.org/abs/" as *u8)
  if ax >= 0 {
      let ae: i64 = pn_find(fb, ax, n, "\"" as *u8)
      if ae > ax {
          o = pn_app(out, o, "ARXIV\t" as *u8)
          o = pn_appn(out, o, fb, ax + 14, ae)
          o = pn_app(out, o, "\n" as *u8)
      }
  }
  let cnt: *i64 = sys_mmap(16) as *i64
  let nfig: *i64 = sys_mmap(16) as *i64
  let nref: *i64 = sys_mmap(16) as *i64
  let nsec: *i64 = sys_mmap(16) as *i64
  cnt[0] = 0
  nfig[0] = 0
  nref[0] = 0
  nsec[0] = 0
  let abs0: i64 = pn_find(fb, 0, n, "id=\"abstract\"" as *u8)
  if abs0 < 0 { pn_puts("PAPER-NATIVE REFUSE: no abstract block\n" as *u8); return 3 }
  let abs1: i64 = pn_find(fb, abs0, n, ">" as *u8) + 1
  let abs2: i64 = pn_find(fb, abs1, n, "</div>" as *u8)
  if abs2 < 0 { pn_puts("PAPER-NATIVE REFUSE: unterminated abstract block\n" as *u8); return 3 }
  o = pn_app(out, o, ".SEC 1\tabstract\tAbstract\n" as *u8)
  nsec[0] = 1
  o = pn_app(out, o, ".PARA " as *u8)
  o = pn_inline(out, o, fb, abs1, abs2, cnt)
  o = pn_app(out, o, "\n" as *u8)
  var smath: i64 = 0
  var sfig: i64 = 0
  var sref: i64 = 0
  var bodyb: i64 = 0
  if an > 0 {
      bodyb = an
      let st: i64 = pn_bodystart(ab, an)
      if st >= 0 {
          smath = pn_count(ab, st, an, "<math" as *u8)
          sfig = pn_count(ab, st, an, "<figure" as *u8)
          sref = pn_count(ab, st, an, "class=\"ltx_bibitem\"" as *u8)
      }
      o = pn_body(out, o, ab, an, cnt, nfig, nref, nsec)
  }
  if o >= outcap { pn_puts("PAPER-NATIVE REFUSE: emitted output reached its derived reservation, so the size derivation is WRONG and this document may be corrupt. Do not widen the reservation without first finding which emit path expands its source.\n" as *u8); return 3 }
  if pn_write(outp, out, o) != 0 { pn_puts("PAPER-NATIVE REFUSE: cannot write\n" as *u8); return 1 }
  pn_puts("PAPER-NATIVE-OK authors=" as *u8); pn_num(nau)
  pn_puts(" sections=" as *u8); pn_num(nsec[0])
  pn_puts(" math=" as *u8); pn_num(cnt[0]); pn_puts("/" as *u8); pn_num(smath)
  pn_puts(" figs=" as *u8); pn_num(nfig[0]); pn_puts("/" as *u8); pn_num(sfig)
  pn_puts(" refs=" as *u8); pn_num(nref[0]); pn_puts("/" as *u8); pn_num(sref)
  pn_puts(" fulltext_src_bytes=" as *u8); pn_num(bodyb)
  pn_puts(" nxpap_bytes=" as *u8); pn_num(o); pn_puts("\n" as *u8)
  var lost: i64 = 0
  if cnt[0] < smath { lost = 1 }
  if nfig[0] < sfig { lost = 1 }
  if nref[0] < sref { lost = 1 }
  if lost == 1 {
      pn_puts("PAPER-NATIVE REFUSE: FIDELITY LOSS -- the source carries more math, figures or references than were emitted (see the emitted/source pairs above). Silently dropping them is exactly the v1 defect, so this run FAILS. The document was still written so the gap can be inspected; it is not fit to publish.\n" as *u8)
      return 3
  }
  return 0
634}
635
636// ---- VIEW: NXPAP1 or NXPAP2 -> the reader page (Nishi bytes rendering Nishi bytes) ----
637func pn_field(fb: *u8, p: i64, e: i64, idx: i64, r: *i64) -> i64 {
  var s: i64 = p
  var k: i64 = 0
  var run: i64 = 1
  while run == 1 {
      if k == idx {
          var t: i64 = pn_find(fb, s, e, "\t" as *u8)
          if t < 0 { t = e }
          r[0] = s
          r[1] = t
          return 0
      }
      let t2: i64 = pn_find(fb, s, e, "\t" as *u8)
      if t2 < 0 { run = 0 } else { s = t2 + 1; k = k + 1 }
  }
  return 0 - 1
653}
654func pn_view(inp: *u8, outp: *u8) -> i64 {
  let ln: *i64 = sys_mmap(16) as *i64
  let fb: *u8 = sys_read_file(inp, ln)
  if (fb as i64) == 0 { pn_puts("PAPER-VIEW REFUSE: cannot read nxpap\n" as *u8); return 3 }
  let n: i64 = ln[0]
  if n <= 0 { pn_puts("PAPER-VIEW REFUSE: cannot read nxpap\n" as *u8); return 3 }
  var v2: i64 = 0
  if pn_find(fb, 0, n, "NXPAP2" as *u8) == 0 { v2 = 1 }
  if v2 == 0 { if pn_find(fb, 0, n, "NXPAP1" as *u8) != 0 { pn_puts("PAPER-VIEW REFUSE: not an NXPAP document\n" as *u8); return 3 } }
  let outcap: i64 = n * 8 + PN_MAGIC_1048576
  let out: *u8 = sys_mmap(outcap)
  let r: *i64 = sys_mmap(32) as *i64
  var o: i64 = 0
  var ti: i64 = pn_find(fb, 0, n, "TITLE\t" as *u8)
  var te: i64 = n
  if ti >= 0 { te = pn_find(fb, ti, n, "\n" as *u8) }
  o = pn_app(out, o, "<!DOCTYPE html><html lang=\"en\"><head><meta charset=\"utf-8\"><meta name=\"viewport\" content=\"width=device-width,initial-scale=1\"><title>" as *u8)
  if ti >= 0 { o = pn_appn(out, o, fb, ti + 6, te) }
  o = pn_app(out, o, " - Nishi Papers</title><style>body{margin:0;background:#0b0e16;color:#e8ecf6;font:16px/1.7 system-ui,sans-serif}main{max-width:78ch;margin:0 auto;padding:26px 18px 90px}a{color:#c9a0ff}h1{font-size:25px;line-height:1.25;margin:0 0 8px}h2{font-size:12px;letter-spacing:.13em;text-transform:uppercase;color:#c9a0ff;margin:32px 0 8px;border-bottom:1px solid #222c44;padding-bottom:6px}h3{font-size:15px;color:#e8ecf6;margin:22px 0 6px}h4{font-size:13px;color:#b9c4d8;margin:18px 0 6px}p{margin:0 0 14px}.meta{color:#8a97ad;font-size:.9rem;margin:0 0 18px}.eq{overflow-x:auto;margin:16px 0;padding:6px 0;text-align:center}figure{margin:20px 0;padding:0}figure img{max-width:100%;height:auto;border-radius:6px;background:#111827}figcaption{color:#8a97ad;font-size:.86rem;margin-top:8px}table{border-collapse:collapse;width:100%;font-size:.9rem;display:block;overflow-x:auto}td,th{border:1px solid #222c44;padding:5px 9px;text-align:left}ol.refs{font-size:.88rem;color:#b9c4d8;padding-left:22px}ol.refs li{margin:0 0 8px}.k{color:#c9a0ff}.note{color:#8a97ad;font-size:.82rem;border-top:1px solid #222c44;margin-top:34px;padding-top:12px}</style></head><body><main>\n" as *u8)
  o = pn_app(out, o, "<p class=\"meta\"><a href=\"/papers\">Nishi Papers</a> &rsaquo; native document</p>\n<h1>" as *u8)
  if ti >= 0 { o = pn_appn(out, o, fb, ti + 6, te) }
  o = pn_app(out, o, "</h1>\n<p class=\"meta\">" as *u8)
  var p: i64 = 0
  var nau: i64 = 0
  var run: i64 = 1
  while run == 1 {
      let a: i64 = pn_find(fb, p, n, "AUTHOR\t" as *u8)
      if a < 0 { run = 0 } else {
          let e: i64 = pn_find(fb, a, n, "\n" as *u8)
          if nau > 0 { o = pn_app(out, o, " &middot; " as *u8) }
          o = pn_appn(out, o, fb, a + 7, e)
          nau = nau + 1
          p = e
      }
  }
  let vi: i64 = pn_find(fb, 0, n, "VENUE\t" as *u8)
  if vi >= 0 {
      let ve2: i64 = pn_find(fb, vi, n, "\n" as *u8)
      o = pn_app(out, o, "<br>" as *u8)
      o = pn_appn(out, o, fb, vi + 6, ve2)
  }
  let yi: i64 = pn_find(fb, 0, n, "YEAR\t" as *u8)
  if yi >= 0 {
      let ye: i64 = pn_find(fb, yi, n, "\n" as *u8)
      o = pn_app(out, o, ", " as *u8)
      o = pn_appn(out, o, fb, yi + 5, ye)
  }
  o = pn_app(out, o, "</p>\n" as *u8)
  var nsec: i64 = 0
  var nmath: i64 = 0
  var nfig: i64 = 0
  var nref: i64 = 0
  var figopen: i64 = 0
  var refopen: i64 = 0
  var tblopen: i64 = 0
  var lp: i64 = 0
  var lr: i64 = 1
  while lr == 1 {
      if lp >= n { lr = 0 } else {
          var e: i64 = pn_find(fb, lp, n, "\n" as *u8)
          if e < 0 { e = n }
          if pn_find(fb, lp, e, ".SEC " as *u8) == lp {
              if tblopen == 1 { o = pn_app(out, o, "</table>" as *u8); tblopen = 0 }
              if figopen == 1 { o = pn_app(out, o, "</figure>\n" as *u8); figopen = 0 }
              if refopen == 1 { o = pn_app(out, o, "</ol>\n" as *u8); refopen = 0 }
              var lvl: i64 = 1
              if fb[lp + 5] == (50 as u8) { lvl = 2 }
              if fb[lp + 5] == (51 as u8) { lvl = 3 }
              if pn_field(fb, lp + 5, e, 1, r) == 0 {
                  let ids: i64 = r[0]
                  let ide: i64 = r[1]
                  if pn_field(fb, lp + 5, e, 2, r) == 0 {
                      if lvl == 1 { o = pn_app(out, o, "<h2 id=\"" as *u8) }
                      if lvl == 2 { o = pn_app(out, o, "<h3 id=\"" as *u8) }
                      if lvl == 3 { o = pn_app(out, o, "<h4 id=\"" as *u8) }
                      o = pn_appn(out, o, fb, ids, ide)
                      o = pn_app(out, o, "\">" as *u8)
                      o = pn_appn(out, o, fb, r[0], r[1])
                      if lvl == 1 { o = pn_app(out, o, "</h2>\n" as *u8) }
                      if lvl == 2 { o = pn_app(out, o, "</h3>\n" as *u8) }
                      if lvl == 3 { o = pn_app(out, o, "</h4>\n" as *u8) }
                      nsec = nsec + 1
                  }
              }
          } else {
          if pn_find(fb, lp, e, ".PARA " as *u8) == lp {
              o = pn_app(out, o, "<p>" as *u8)
              o = pn_appn(out, o, fb, lp + 6, e)
              o = pn_app(out, o, "</p>\n" as *u8)
          } else {
          if pn_find(fb, lp, e, ".MATH " as *u8) == lp {
              if pn_field(fb, lp + 6, e, 0, r) == 0 {
                  o = pn_app(out, o, "<div class=\"eq\" id=\"" as *u8)
                  o = pn_appn(out, o, fb, r[0], r[1])
                  o = pn_app(out, o, "\">" as *u8)
                  if pn_field(fb, lp + 6, e, 1, r) == 0 { o = pn_appn(out, o, fb, r[0], r[1]) }
                  o = pn_app(out, o, "</div>\n" as *u8)
                  nmath = nmath + 1
              }
          } else {
          if pn_find(fb, lp, e, ".FIG " as *u8) == lp {
              if figopen == 1 { o = pn_app(out, o, "</figure>\n" as *u8) }
              o = pn_app(out, o, "<figure id=\"" as *u8)
              o = pn_appn(out, o, fb, lp + 5, e)
              o = pn_app(out, o, "\">" as *u8)
              figopen = 1
              nfig = nfig + 1
          } else {
          if pn_find(fb, lp, e, ".TBL " as *u8) == lp {
              if figopen == 1 { o = pn_app(out, o, "</figure>\n" as *u8) }
              o = pn_app(out, o, "<figure id=\"" as *u8)
              o = pn_appn(out, o, fb, lp + 5, e)
              o = pn_app(out, o, "\"><table>" as *u8)
              figopen = 1
              tblopen = 1
              nfig = nfig + 1
          } else {
          if pn_find(fb, lp, e, ".IMG " as *u8) == lp {
              o = pn_app(out, o, "<img loading=\"lazy\" alt=\"\" src=\"" as *u8)
              o = pn_appn(out, o, fb, lp + 5, e)
              o = pn_app(out, o, "\">" as *u8)
          } else {
          if pn_find(fb, lp, e, ".CAP " as *u8) == lp {
              if tblopen == 1 { o = pn_app(out, o, "</table>" as *u8); tblopen = 0 }
              o = pn_app(out, o, "<figcaption>" as *u8)
              o = pn_appn(out, o, fb, lp + 5, e)
              o = pn_app(out, o, "</figcaption>" as *u8)
          } else {
          if pn_find(fb, lp, e, ".ROW " as *u8) == lp {
              o = pn_app(out, o, "<tr>" as *u8)
              var ci: i64 = 0
              var cr: i64 = 1
              while cr == 1 {
                  if pn_field(fb, lp + 5, e, ci, r) == 0 {
                      o = pn_app(out, o, "<td>" as *u8)
                      o = pn_appn(out, o, fb, r[0], r[1])
                      o = pn_app(out, o, "</td>" as *u8)
                      if r[1] >= e { cr = 0 } else { ci = ci + 1 }
                  } else { cr = 0 }
              }
              o = pn_app(out, o, "</tr>" as *u8)
          } else {
          if pn_find(fb, lp, e, ".REF " as *u8) == lp {
              if tblopen == 1 { o = pn_app(out, o, "</table>" as *u8); tblopen = 0 }
              if figopen == 1 { o = pn_app(out, o, "</figure>\n" as *u8); figopen = 0 }
              if refopen == 0 { o = pn_app(out, o, "<h2 id=\"bib\">References</h2>\n<ol class=\"refs\">\n" as *u8); refopen = 1 }
              o = pn_app(out, o, "<li id=\"" as *u8)
              if pn_field(fb, lp + 5, e, 0, r) == 0 { o = pn_appn(out, o, fb, r[0], r[1]) }
              o = pn_app(out, o, "\"><span class=\"k\">" as *u8)
              if pn_field(fb, lp + 5, e, 1, r) == 0 { o = pn_appn(out, o, fb, r[0], r[1]) }
              o = pn_app(out, o, "</span> " as *u8)
              if pn_field(fb, lp + 5, e, 2, r) == 0 { o = pn_appn(out, o, fb, r[0], r[1]) }
              o = pn_app(out, o, "</li>\n" as *u8)
              nref = nref + 1
          } else {
          if pn_find(fb, lp, e, ".SECTION " as *u8) == lp {
              // NXPAP1 compatibility: the legacy corpus is still on disk and must stay readable
              var nx: i64 = pn_find(fb, e, n, ".SECTION " as *u8)
              if nx < 0 { nx = n }
              o = pn_app(out, o, "<h2>" as *u8)
              o = pn_appn(out, o, fb, lp + 9, e)
              o = pn_app(out, o, "</h2>\n<p>" as *u8)
              o = pn_appn(out, o, fb, e + 1, nx)
              o = pn_app(out, o, "</p>\n" as *u8)
              nsec = nsec + 1
              e = nx - 1
          } } } } } } } } } }
          lp = e + 1
      }
  }
  if tblopen == 1 { o = pn_app(out, o, "</table>" as *u8) }
  if figopen == 1 { o = pn_app(out, o, "</figure>\n" as *u8) }
  if refopen == 1 { o = pn_app(out, o, "</ol>\n" as *u8) }
  o = pn_app(out, o, "<p class=\"note\">Rendered by <code>nx_paper_native</code> from an NXPAP sovereign document held on the Nishi estate &mdash; no PDF, no third-party reader, and the equations are browser-native MathML. Text, figures and copyright belong to the authors; the source link is preserved in the document header.</p>\n</main></body></html>\n" as *u8)
  if o >= outcap { pn_puts("PAPER-VIEW REFUSE: emitted page reached its derived reservation, so the size derivation is WRONG and this page may be corrupt. Find the expanding emit path; do not just widen the reservation.\n" as *u8); return 3 }
  if pn_write(outp, out, o) != 0 { pn_puts("PAPER-VIEW REFUSE: cannot write\n" as *u8); return 1 }
  pn_puts("PAPER-VIEW-OK sections=" as *u8); pn_num(nsec)
  pn_puts(" authors=" as *u8); pn_num(nau)
  pn_puts(" math=" as *u8); pn_num(nmath)
  pn_puts(" figs=" as *u8); pn_num(nfig)
  pn_puts(" refs=" as *u8); pn_num(nref)
  pn_puts(" html_bytes=" as *u8); pn_num(o); pn_puts("\n" as *u8)
  return 0
837}
838
839func main(argc: i64, argv: *i64) -> i64 {
  if argc < 4 {
      pn_puts("usage: nx_paper_native abs <cvf-abstract.raw> <out.nxpap> [arxiv-html.raw] | nx_paper_native view <in.nxpap> <out.html>\n" as *u8)
      sys_exit(2)
      return 2
  }
  let v: *u8 = argv[1] as *u8
  var rc: i64 = 2
  if v[0] == (97 as u8) {
      var ax: *u8 = 0 as *u8
      if argc > 4 { ax = argv[4] as *u8 }
      rc = pn_ingest(argv[2] as *u8, argv[3] as *u8, ax)
  }
  if v[0] == (118 as u8) { rc = pn_view(argv[2] as *u8, argv[3] as *u8) }
  sys_exit(rc)
  return rc
855}
nx_paper_native.nx source

`nx_paper_native.nx` source
