Closed wtwhite closed 1 year ago
vul4j
does state that the failing module is jspwiki-main
.
The commit, https://github.com/apache/jspwiki/commit/88d89d6523802c044cfcb7930cba40d8eeb21da2, that vul4j
claims contains the human patch also contains a new test, which is what we currently use (modulo some other changes described in the README) -- but that test has a bug, which was fixed a few minutes later in https://github.com/apache/jspwiki/commit/3ad9e5e5d9748750d49cc17ade48637dc0effbc6.
With that fix to the test, it should pass in all versions >= 2.11.0.M3
, and fail in all earlier versions.
Solution: Make the same change to our copy of the test (and notify vul4j
as well, to be nice).
The
pom.xml
forCVE-2019-0225
specifiesorg.apache.jspwiki:jspwiki-main:2.11.0.M6
, but there are 2 unusual things about this:/tools/create-pov-project.js
auto-assigned thearticleId
jspwiki-war
(instead ofjspwiki-main
) based on the GHSA2.11.0.M2
are vulnerable, and the GHSA even states that2.11.0.M3
is already patched -- so how could the later2.11.0.M6
still be vulnerable?The initial issue is that this discrepancy broke
shadedetector
's metadata-reading codepath. Beyond that, the underlying test might need fixing anyway.