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 π
Go aicertificationcloud-engineeringcybercybersecuritycybersecurity-analystcybersecurity-certificationcybersecurity-certificationscybersecurity-educationcybersecurity-jobscybersecurity-portfoliocybersecurity-projectscybersecurity-toolsgrchackinglearningpentestingproject-repositoryprojetcspython
feat(lisdex): B66 - the defect was already fixed and the clause guarding it could not fail
CarterPerez-deve7562f613e21
diff --git a/PROJECTS/advanced/zero-day-vulnerability-scanner/scripts/m3-gate.sh b/PROJECTS/advanced/zero-day-vulnerability-scanner/scripts/m3-gate.shindex d737405f..90ed7b5e 100755--- a/PROJECTS/advanced/zero-day-vulnerability-scanner/scripts/m3-gate.sh+++ b/PROJECTS/advanced/zero-day-vulnerability-scanner/scripts/m3-gate.sh@@ -351,11 +351,31 @@ fi# Isolates the binding rule from the run rule: `archive_read_free` is plumbing-shaped and# sits inside the opening run of several targets, and `archive_read_set_options` takes a# string. Neither may be replayed, and replaying the free would tear the handle down first.-if echo "$REPLAY" | grep -qE "archive_read_free|archive_read_set_options"; then+#+# Both names are DOUBLY unreachable on this target and this assertion was measured vacuous:+# the seek fuzzer's run breaks at `archive_read_set_read_callback` long before its+# `archive_read_free` at position 13, and `emitted` would skip that free anyway because the+# plan schedules it. Gutting `is_teardown` left the whole gate green - the unit test+# `a_destructor_in_the_opening_run_is_never_replayed_as_setup` is what actually covers that+# guard. Kept as the regression statement it always was, now with the liveness guard that+# stops it passing on the empty replay a broken run rule produces.+if [ -z "$REPLAY" ]; then+ bad "no replay line to assert against, so the teardown assertion would pass vacuously"+elif echo "$REPLAY" | grep -qE "archive_read_free|archive_read_set_options"; thenbad "libarchive replayed a teardown or a string-taking call as setup: $REPLAY"elseok "only handle-taking registrations are replayed, no teardown and no string-taking call"fi+# The live negative control for the run rule, and the one the clause above cannot be. With+# `is_plumbing` gutted to always return true the run walks past the callback setters it+# cannot bind and reaches the author's OWN opener: the replay becomes the six registrations+# plus `archive_read_open1`, which would open the archive a second time before ours ever+# runs. Measured, so this reddens alone when the break rule is loosened.+if echo "$REPLAY" | grep -q "archive_read_open1"; then+ bad "the run walked past the callback setters and replayed a second opener: $REPLAY"+else+ ok "the replay stops at the first call it cannot bind, so no second opener is replayed"+fi# A harness that never passes the corpus to the library is a constant program: every seed# drives the same code and any crash it reports is one we manufactured.