* report each hardlink as its own file when scanning images
image scans previously collapsed a set of hardlinks onto a single file, so only
one path per inode showed up in results. dir scans report every hardlink path,
which made image vs dir SBOMs of the same filesystem diverge (and produce
different SPDX `packageVerificationCode` values for packages that own hardlinked
files).
now both image resolvers (squash and all-layers) surface each hardlink at its
own path as a regular file bound to the target's content, matching dir scans.
user-facing impact:
- SBOMs for images containing hardlinks will list more `file` entries
- SPDX `packageVerificationCode` values change for affected packages, now
matching the equivalent `dir:` scan
- adds `file.NewVirtualLocationFromImage` to the public API
fixes#5019
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
* fix busybox test assertion
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
---------
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
parseErlangString advances past a backslash escape and then checks
len(data) >= *i before reading the escaped byte. That condition is
almost always true (it only turns false once *i runs off the end),
so the intended out-of-range guard fires on the very first escape
character it sees instead of only at EOF. Any rebar.lock or OTP
resource file containing a backslash in a quoted string (a Windows
git path, an escaped quote, anything) fails to parse and the whole
file, and every package in it, gets dropped.
Flip the comparison to *i >= len(data) so the guard only trips when
the escape is genuinely truncated, and add a regression test for a
string with an escaped quote.
Signed-off-by: Arpit Jain <arpitjain099@gmail.com>
* Prevent duplicate Go packages after source replacement resolution
Source analysis already applies module replacements through go/packages. Avoid synthesizing the same replacement again from go.mod while retaining fallback synthesis for modules that source analysis did not resolve.
Constraint: Preserve unimported and local-path replacement cataloging.
Rejected: Deduplicate only during final assembly | That retains redundant license lookup and ambiguous metadata ownership.
Confidence: high
Scope-risk: narrow
Directive: Keep go.mod fallback packages limited to modules absent from source analysis.
Tested: Focused replacement regression, related Go module parser tests, go vet, gofmt, and diff checks.
Not-tested: Docker-backed full cataloger fixtures; local root storage was exhausted by image generation.
Signed-off-by: ychampion <ychampion@users.noreply.github.com>
* Keep replacement fixtures with the Go module test data
Constraint: The maintainer reserves internal/gotestdata for fixtures that need special Go tooling discovery.
Rejected: Leave this fixture in gotestdata | The regression opens its module explicitly and does not need the special location.
Confidence: high
Scope-risk: narrow
Directive: Use internal/gotestdata only when a fixture must avoid Go testdata discovery rules.
Tested: replacement regression repeated 10 times; Go module parser table; go vet for the Go cataloger; gofmt; diff checks.
Not-tested: Full cataloger package; three unrelated parser fixtures fail identically on exact prior head in this environment.
Signed-off-by: ychampion <ychampion@users.noreply.github.com>
---------
Signed-off-by: ychampion <ychampion@users.noreply.github.com>
Co-authored-by: ychampion <ychampion@users.noreply.github.com>
Update the binary classifier cataloger developer README to match the
current workflow:
- Testing section used the non-existent flag -must-use-full-binaries
against .../testdata/...; the real flag is -must-use-original-binaries
and Test_Cataloger_PositiveCases lives in the binary package. Fix the
command and add a single-fixture variant.
- Point add-snippet references at the make target and note the (Y/n/q)
prompt requires y to write the snippet.
- Fix the test file name (classifier_cataloger_test.go) and document the
go generate ./internal/capabilities step for new classifiers.
Fixes#4510
Signed-off-by: alliasgher <alliasgher123@gmail.com>
GHC boot-lib constraints (any.base installed) have no ' ==', so fields[1] panicked with index out of range and the recover dropped all Haskell packages. Skip constraint lines without a version.
Signed-off-by: Synvoya <16019863+Synvoya@users.noreply.github.com>
Co-authored-by: Synvoya <16019863+Synvoya@users.noreply.github.com>
The hand-written rockspec parser reads a byte past the end of the buffer
in two spots when a comment runs right up to the end of the file.
In parseRockspecBlock, when a block starts with a leading comment that
consumes the rest of the file, the SkipWhitespace afterward leaves the
index at len(data) and the following `c = data[*i]` reads out of range.
In parseComment, `data[*i]` is read after the index is advanced to check
for a CR/LF pair, so a bare carriage return as the last byte reads past
the end.
Both cases show up with a rockspec whose final line is a comment ending
in a lone \r with no trailing newline. That is malformed but harmless
input, and the panic aborts the whole Lua cataloger, so every valid Lua
package in the same scan gets dropped. Guard both reads with a length
check and return cleanly at EOF. Added table cases covering a
comment-only file and a trailing comment, both ending in a bare CR.
Signed-off-by: arpitjain099 <arpitjain099@gmail.com>
deno.lock npm keys append resolved peer dependencies after the version,
separated by "_" (e.g. "typedoc@0.28.19_typescript@6.0.3" or the scoped
"@scope/name@1.2.3_@scope+peer@4.5.6"). parseDenoNpmNameVersion split on the
last "@", capturing a peer's "@" and producing a garbage package name plus
the peer's version (and a corrupt PURL that breaks vulnerability matching).
Strip the peer suffix (the first "_" after the version "@") before splitting
name@version; a semver version never contains "_".
Signed-off-by: Synvoya <16019863+Synvoya@users.noreply.github.com>
Co-authored-by: Synvoya <16019863+Synvoya@users.noreply.github.com>
* swift: emit canonical purls without .git suffix or repeated name
Reported in anchore/syft#3961: the purl syft emits for Swift packages
parsed out of Package.resolved keeps the repository URLs .git suffix
and then re-appends the package name, producing
pkg:swift/github.com/apple/swift-nio-ssl.git/swift-nio-ssl@2.0.0
NVD / Grype cannot match that purl against the known
swift-nio-ssl@2.0.0 CVE (GHSA-frg3-gpcx-968f), so every Swift SBOM
produced by syft silently loses vulnerability coverage. cdxgen and
the wider purl ecosystem use the shorter form
pkg:swift/github.com/apple/swift-nio-ssl@2.0.0
which Grype does match.
Replace the ad-hoc strings.Replace with swiftNamespaceFromSourceURL,
which trims the common URL schemes, strips the ".git" suffix, and
drops a trailing /<name> segment so the namespace is only the
organisation path (e.g. github.com/apple). Existing test expectations
are updated to the new purl shape.
Fixes#3961
Signed-off-by: Sai Asish Y <say.apm35@gmail.com>
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
* Revert unrelated consul binary classifier changes
The swift purl commit (ff5ffc37) accidentally bundled removal of the
consul GitDescribe and NUL-wrapped version matchers plus the 1.12.9 and
1.7.14 fixtures/test cases. That drops version detection for consul
binaries not carrying the CONSUL_VERSION string. Restore the binary
cataloger to its pre-commit state so this branch is swift-only.
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
* swift: strip repeated name segment case-insensitively
Package.resolved identities are lowercased while the repo URL path may be
mixed-case (e.g. github.com/Apple/Swift-NIO vs identity swift-nio). The
case-sensitive TrimSuffix left the repo segment in the namespace for such
repos, reintroducing the duplicated-name purl. Compare the trailing segment
with EqualFold and add unit coverage for swiftPackageManagerPackageURL.
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
---------
Signed-off-by: Sai Asish Y <say.apm35@gmail.com>
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
Co-authored-by: Alex Goodman <wagoodman@users.noreply.github.com>
The [requires] section parser set inRequirements=false when it
encountered a comment line (#) inside the section, because the comment
check was OR'd with the section-change check. Any dependency listed
after a comment was silently skipped.
Skip blank lines and comments before evaluating section state so they
never toggle inRequirements. Add a regression test fixture with a
comment inside [requires].
Fixes#5017
Signed-off-by: jeff <jfjrh2014@gmail.com>
parseMixLock read the version from a hex-only token position and hardcoded a pkg:hex PURL, so git/path deps got a URL-fragment version and a false hex.pm PURL (false CVE matches). Branch on the source atom: git uses the commit SHA, path an empty version, and pkg:hex PURLs are emitted only for hex deps.
Signed-off-by: Synvoya <16019863+Synvoya@users.noreply.github.com>
Co-authored-by: Synvoya <16019863+Synvoya@users.noreply.github.com>
the plain version pin drifted out from under us: debian mirrors only keep
the latest point release, so 8.2.31 became uninstallable once 8.2.32 shipped
and apt pulled the unpinned transitive php8.2-cli at 8.2.32. pointing apt at
snapshot.debian.org keeps the exact point release installable so the fixture
(and its test expectations) stay reproducible.
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
* ruby/gemspec: resolve simple #{s.name}/#{s.version} interpolation
Reported in anchore/syft#4720: scanning projects that depend on gems
like formatador leaks literal Ruby interpolation into the emitted
SBOM, e.g.
"externalReferences": [
{ "url": "https://github.com/geemus/#{s.name}", "type": "website" }
]
because formatador.gemspec uses
s.homepage = "https://github.com/geemus/#{s.name}"
and parseGemSpecEntries reads the file as plain text instead of
evaluating it. The interpolation leaks through the captured homepage
field and on into any externalReferences entry the cataloger produces.
Dependency Track then rejects the whole BOM because '{' and '}' are
not valid IRI-reference characters (RFC 3987).
Add a post-parse pass that substitutes the common interpolation forms
(#{s.name}, #{gem.name}, #{name}, and the matching #{*.version}
variants) in captured string fields using values already parsed from
the same gemspec. Anything still containing '#{' after best-effort
substitution is an unresolvable Ruby expression, and for URL-like
fields (currently just homepage) we drop the field entirely so the
SBOM is always schema-valid; callers would rather miss a homepage URL
than emit one that breaks downstream tools.
Adds testdata/formatador.gemspec, a minimal real-world gemspec using
the #{s.name} pattern, plus a new parser test asserting that the
homepage field comes out fully resolved.
Fixes#4720
Signed-off-by: Sai Asish Y <say.apm35@gmail.com>
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
* improve test cases
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
---------
Signed-off-by: Sai Asish Y <say.apm35@gmail.com>
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
Co-authored-by: Alex Goodman <wagoodman@users.noreply.github.com>
adds dependency-of relationships between elixir locked packages, matching how other
ecosystem catalogers (alpine, arch, debian, redhat, python) express the
dependency graph via the shared dependency.Processor/Specifier mechanism.
Signed-off-by: Chris Greeno <cgreeno@gmail.com>
Borland Package Library (.bpl) files are standard Windows PE/DLL files
used in Delphi and C++Builder ecosystems. This adds the .bpl glob
pattern to the PE file discovery so these files are cataloged alongside
.dll and .exe files.
* fix(dpkg): extract License field for opkg/ipkg entries
opkg and ipkg use the dpkg cataloger but declare the package License
inline in the status DB (unlike Debian dpkg, where licenses live in
copyright files). The cataloger silently dropped the License field at
mapstructure decode time, so all opkg-managed packages reported empty
licenses.
This adds the field to the intermediate decode struct and the public
DpkgDBEntry, and populates licenses in newDpkgPackage using the alpine
cataloger's pattern: try license.ParseExpression first to keep valid
SPDX expressions whole, fall back to whitespace splitting for
space-separated lists.
Standard Debian dpkg status files never carry a License field per
Debian policy, so the new path is a no-op for them; the existing
copyright-file lookup in addLicenses is unaffected.
Closes#4940
Signed-off-by: David Dashti <47575784+Dashtid@users.noreply.github.com>
* remove license from dpkg metadata struct
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
* restore format snapshot files
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
* add additional tests
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
---------
Signed-off-by: David Dashti <47575784+Dashtid@users.noreply.github.com>
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
Co-authored-by: Alex Goodman <wagoodman@users.noreply.github.com>
Alex Goodman
Yashas Gunderiaandychampion
Keith Zantow
Eljees
Enes Deniz
Rayan Salhabandcyphercodes
Arpit Jain
Rez Moss
anchore-oss-update-botandanchore-oss-update-bot
Christopher Angelo PhillipsandAlex Goodman
Monibee-FudgekinsandAlex Goodman
Mike ZolotarovandAlex Goodman
Ali Asghar
SynvoyaandSynvoya
Sai Asish YandAlex Goodman
Marcus
gabandAlex Goodman
Archy
Chris Greeno
sputnik-mac
David DashtiandAlex Goodman
Kursat TopcuogluandKursat Topcuoglu