mirror of
https://github.com/LLVMParty/striga
synced 2026-06-21 13:43:42 +00:00
Duncan Ogilvie
111 lines
3.2 KiB
Python
111 lines
3.2 KiB
Python
from __future__ import annotations
|
|
|
|
import re
|
|
from html.parser import HTMLParser
|
|
from pathlib import Path
|
|
|
|
ROOT = Path(__file__).resolve().parent
|
|
HTML = ROOT / "binaryshield-devirt-presentation.html"
|
|
ARTIFACTS = ROOT / "devirt-presentation-artifacts"
|
|
|
|
REQUIRED_SECTIONS = {
|
|
"problem",
|
|
"pipeline",
|
|
"state",
|
|
"resolver",
|
|
"memory",
|
|
"graph",
|
|
"recovery",
|
|
"intermediate",
|
|
"optimization",
|
|
"result",
|
|
"verification",
|
|
"specifics",
|
|
"limits",
|
|
}
|
|
|
|
REQUIRED_SNIPPETS = {
|
|
"resolver-built-call.ll",
|
|
"resolver-inlined-hook-call.ll",
|
|
"resolver-hook-rewritten-return.ll",
|
|
"resolver-optimized-target.ll",
|
|
"jcc-resolver-select.ll",
|
|
"recovered-skeleton-switch.ll",
|
|
"recovered-round8-first-compare.ll",
|
|
"recovered-residual-membership.ll",
|
|
"recovered-final.ll",
|
|
}
|
|
|
|
|
|
class StructureParser(HTMLParser):
|
|
def __init__(self) -> None:
|
|
super().__init__()
|
|
self.sections: set[str] = set()
|
|
self.links: list[str] = []
|
|
self.code_blocks = 0
|
|
self.tables = 0
|
|
|
|
def handle_starttag(self, tag: str, attrs: list[tuple[str, str | None]]) -> None:
|
|
attr = dict(attrs)
|
|
if tag == "section" and attr.get("id"):
|
|
self.sections.add(attr["id"] or "")
|
|
if tag == "a" and attr.get("href"):
|
|
self.links.append(attr["href"] or "")
|
|
if tag == "pre":
|
|
self.code_blocks += 1
|
|
if tag == "table":
|
|
self.tables += 1
|
|
|
|
|
|
def main() -> None:
|
|
html = HTML.read_text(encoding="utf-8")
|
|
parser = StructureParser()
|
|
parser.feed(html)
|
|
|
|
missing_sections = sorted(REQUIRED_SECTIONS - parser.sections)
|
|
assert not missing_sections, f"missing sections: {missing_sections}"
|
|
|
|
missing_links = sorted(
|
|
f"#{section}"
|
|
for section in REQUIRED_SECTIONS
|
|
if f"#{section}" not in parser.links
|
|
)
|
|
assert not missing_links, f"missing toc links: {missing_links}"
|
|
|
|
snippet_dir = ARTIFACTS / "snippets"
|
|
missing_snippets = sorted(
|
|
name for name in REQUIRED_SNIPPETS if not (snippet_dir / name).exists()
|
|
)
|
|
assert not missing_snippets, f"missing snippets: {missing_snippets}"
|
|
|
|
required_terms = [
|
|
"Intermediate IR snapshots",
|
|
"VM-specific knowledge encoded",
|
|
"Final membership cleanup",
|
|
"STRIGA_DEVIRT_DUMP_DIR",
|
|
"recovered-round8-first-compare.ll",
|
|
"recovered-residual-membership.ll",
|
|
"_try_clean_binaryshield_membership_ir",
|
|
"llvm-nanobind stale wrappers",
|
|
]
|
|
for term in required_terms:
|
|
assert term in html, f"missing term: {term}"
|
|
|
|
final_ir = (ARTIFACTS / "recovered" / "06-final-clean.ll").read_text(
|
|
encoding="utf-8"
|
|
)
|
|
constants = {int(x) for x in re.findall(r"icmp eq i32 %x, (\d+)", final_ir)}
|
|
assert constants == {1859, 2418, 1638, 299902, 29763}, constants
|
|
for banned in ["@RAM", "__striga_", "switch ", "unresolved"]:
|
|
assert banned not in final_ir, banned
|
|
|
|
print(f"sections={len(parser.sections)}")
|
|
print(f"toc_links={len(parser.links)}")
|
|
print(f"code_blocks={parser.code_blocks}")
|
|
print(f"tables={parser.tables}")
|
|
print("presentation validation passed")
|
|
|
|
|
|
if __name__ == "__main__":
|
|
main()
|