Skip to content

Building 70 Projects ranging from beginner to advanced so anyone can β€” learn from, build upon, use as a reference, or even copy directly. Owned by my cybersecurity company πŸ‘‡

feat(lisdex): B78 - a one-character vocabulary entry is a wildcard, and the row's floor was too high

CarterPerez-devc03f6b0894cf
diff --git a/PROJECTS/advanced/zero-day-vulnerability-scanner/crates/lisdex-harness/src/synth.rs b/PROJECTS/advanced/zero-day-vulnerability-scanner/crates/lisdex-harness/src/synth.rs
index 930f075e..7c06543a 100644
--- a/PROJECTS/advanced/zero-day-vulnerability-scanner/crates/lisdex-harness/src/synth.rs
+++ b/PROJECTS/advanced/zero-day-vulnerability-scanner/crates/lisdex-harness/src/synth.rs
@@ -4200,6 +4200,28 @@ mod tests {
assert!(!is_extent_name("color_type"));
}
+ #[test]
+ fn a_word_ending_in_n_is_not_an_extent_because_n_is_a_vocabulary_entry() {
+ for name in [
+ "align",
+ "distortion",
+ "pattern",
+ "column",
+ "origin",
+ "version",
+ ] {
+ assert!(
+ !is_extent_name(name),
+ "{name} is not an extent; it merely ends in the one-character entry n"
+ );
+ }
+ assert!(is_extent_name("n"), "n itself is still an extent name");
+ assert!(
+ is_extent_name("inlen") && is_extent_name("keylen") && is_extent_name("isize"),
+ "the suffix arm still carries the glued lowercase extents it was added for"
+ );
+ }
+
#[test]
fn a_config_that_is_not_an_extent_still_binds() {
let expat = candidate(
diff --git a/PROJECTS/advanced/zero-day-vulnerability-scanner/crates/lisdex-surface/src/extract.rs b/PROJECTS/advanced/zero-day-vulnerability-scanner/crates/lisdex-surface/src/extract.rs
index 91b1b2bd..a6b7ce71 100644
--- a/PROJECTS/advanced/zero-day-vulnerability-scanner/crates/lisdex-surface/src/extract.rs
+++ b/PROJECTS/advanced/zero-day-vulnerability-scanner/crates/lisdex-surface/src/extract.rs
@@ -24,6 +24,7 @@ impl Drop for ReentryGuard {
}
pub const MIN_PROJECT_STEM: usize = 3;
+pub const MIN_SUFFIX_ENTRY: usize = 2;
pub const C_ARGS: [&str; 3] = ["-x", "c", "-std=c11"];
pub const CPP_ARGS: [&str; 3] = ["-x", "c++", "-std=c++17"];
pub const TEARDOWN_VERBS: [&str; 15] = [
@@ -414,9 +415,11 @@ pub fn carries_word(name: &str, vocabulary: &[&str]) -> bool {
// read as one word. Fall back to a suffix match, which is where the verb or the
// extent lands in every case measured.
match words.as_slice() {
- [only] => vocabulary
- .iter()
- .any(|entry| only.len() > entry.len() && only.to_lowercase().ends_with(entry)),
+ [only] => vocabulary.iter().any(|entry| {
+ entry.len() >= MIN_SUFFIX_ENTRY
+ && only.len() > entry.len()
+ && only.to_lowercase().ends_with(entry)
+ }),
_ => false,
}
}
@@ -932,6 +935,40 @@ void WebPFreeDecBuffer(WebPDecBuffer *buffer);
assert!(carries_word("uncompress2", &READ_WORDS));
}
+ #[test]
+ fn a_one_character_entry_is_a_wildcard_on_the_suffix_arm_and_never_matches_there() {
+ let vocabulary = ["n", "len"];
+ for name in ["align", "token", "column", "pattern", "origin", "main"] {
+ assert!(
+ !carries_word(name, &vocabulary),
+ "{name} ends in n and would match a single-character entry as a suffix"
+ );
+ }
+ }
+
+ #[test]
+ fn the_floor_applies_to_the_suffix_arm_alone_and_not_to_token_equality() {
+ let vocabulary = ["n", "len"];
+ assert!(
+ carries_word("n", &vocabulary),
+ "a parameter literally named n is an extent by token equality"
+ );
+ assert!(
+ carries_word("num_n", &vocabulary),
+ "n as a whole word must still match once there is a boundary to split on"
+ );
+ }
+
+ #[test]
+ fn a_two_character_entry_still_matches_as_a_suffix() {
+ let vocabulary = ["sz", "len"];
+ assert!(carries_word("bufsz", &vocabulary));
+ assert!(
+ carries_word("inlen", &vocabulary),
+ "the suffix arm is load-bearing for all-lowercase extents and must survive the floor"
+ );
+ }
+
// The isolating control for stripping per WORD rather than per name: the ladder's old
// local workaround trimmed the whole identifier, which cannot reach a digit that is not
// the last character. Both must pass, and only the per-word rule satisfies the second.