Fix lock-only PEP 440 pin spellings missing patches (#604) (#1334)

* Start v5 blocker fix (pypi-pep440-lockonly)

Empty commit to open the draft PR.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* Find lock-only patches for PEP 440 pin spellings

On a fresh checkout (no venv), requirements.txt discovery sent a pin
as written: `six==1.16` became `pkg:pypi/six@1.16`, `==1.16.0.0`
became `@1.16.0.0`, `==01.16.0` became `@01.16.0`. pip installs the
registry release 1.16.0 for all three, but the patch API keys it
`@1.16.0`, so the scan said "No patches available" and the unpatched
release was installed. The same file with a venv holding six was
patched.

Scan now also asks the API for the other PEP 440 spellings of a
lockfile-only PyPI pure-release pin (leading zeros dropped, release
padded or trimmed to at least three segments). A patch returned under
another spelling is counted as that lockfile-only package, so
`notInstalled` and the vendored baseline pre-check see it. The
rewriters already match `==` pins under PEP 440 (#478).

Fixes #604

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

Author: mikolalysenko

crates/socket-patch-cli/src/commands/scan/mod.rs

@@ -104,6 +104,53 @@ const BATCH_BODY_BYTE_CAP: usize = 256 * 1024;
 /// authenticated API and [`DEFAULT_PROXY_BATCH_SIZE`] on the public proxy.
 /// Floored at 1: `--batch-size 0` is otherwise unvalidated and would make
 /// the chunking below panic, so it degrades to one-package batches.
+/// `purls` plus, for each lockfile-only PyPI purl among them, its other
+/// PEP 440 spellings of the same release (#604), deduplicated in order.
+fn with_pypi_equivalents(purls: &[String], lockfile_only: &HashSet<PurlKey>) -> Vec<String> {
+    let mut out = purls.to_vec();
+    let mut seen: HashSet<PurlKey> = purls.iter().map(|p| PurlKey::new(p)).collect();
+    for purl in purls {
+        if !lockfile_only_contains(lockfile_only, purl) {
+            continue;
+        }
+        for spelling in socket_patch_core::utils::purl_key::pypi_equivalent_purls(purl) {
+            if seen.insert(PurlKey::new(&spelling)) {
+                out.push(spelling);
+            }
+        }
+    }
+    out
+}
+
+/// Mark each API purl that names a lockfile-only PyPI pin under another
+/// PEP 440 spelling (`@1.16.0` for a lock's `@1.16`, #604) as lockfile-only
+/// too, so the `notInstalled` flag, the `[NOT INSTALLED]` marker and the
+/// vendored baseline pre-check treat it as the package it is. A spelling an
+/// installed copy already carries is left alone.
+fn adopt_pypi_equivalents(
+    packages: &[BatchPackagePatches],
+    scanned: &[String],
+    lockfile_only: &mut HashSet<PurlKey>,
+) {
+    let scanned_keys: HashSet<PurlKey> = scanned.iter().map(|p| PurlKey::new(p)).collect();
+    let lock_only_pypi: Vec<&String> = scanned
+        .iter()
+        .filter(|p| p.starts_with("pkg:pypi/") && lockfile_only_contains(lockfile_only, p))
+        .collect();
+    for pkg in packages {
+        let key = PurlKey::new(&pkg.purl);
+        if scanned_keys.contains(&key) {
+            continue;
+        }
+        if lock_only_pypi
+            .iter()
+            .any(|lock| socket_patch_core::utils::purl_key::pypi_same_release(lock, &pkg.purl))
+        {
+            lockfile_only.insert(key);
+        }
+    }
+}
+
 fn effective_batch_size(requested: Option<usize>, use_public_proxy: bool) -> usize {
     requested
         .unwrap_or(if use_public_proxy {
@@ -1858,7 +1905,10 @@ async fn run_scan(
     // that have NO installed copy (fresh clone, partial install). They join
     // discovery and are flagged "not yet installed". Scoped to the crawled
     // ecosystems.
-    let lockfile_only = lockfile_supplement(&ctx, &all_crawled, crawl_scope).await;
+    let mut lockfile_only = lockfile_supplement(&ctx, &all_crawled, crawl_scope).await;
+    // Counted once: #604 adds the API's spellings of lockfile-only PyPI pins
+    // to `lockfile_only.purls` after the batch query.
+    let lockfile_only_count = lockfile_only.purls.len();
     // Unsupported layouts and malformed binary Bun locks, kept on empty
     // scans too: an unreadable graph is not evidence of no dependencies.
     let mut layout_refusals = unsupported_layout_warnings(&lockfile_only.unsupported);
@@ -2305,8 +2355,8 @@ async fn run_scan(
         plural(package_count, "package", "packages")
     ));
     if human {
-        if !lockfile_only.purls.is_empty() {
-            eprintln!("{}", render::lockfile_only_note(lockfile_only.purls.len()));
+        if lockfile_only_count > 0 {
+            eprintln!("{}", render::lockfile_only_note(lockfile_only_count));
         }
         print_layout_refusals(&layout_refusals, args.common.silent);
         policy.print_warnings(args.common.silent);
@@ -2315,7 +2365,12 @@ async fn run_scan(
     // Query API in batches
     let mut all_packages_with_patches: Vec<BatchPackagePatches> = Vec::new();
     let mut can_access_paid_patches = false;
-    let chunks: Vec<&[String]> = batch_chunks(&all_purls, batch_size, BATCH_BODY_BYTE_CAP);
+    // #604: a lockfile-only PyPI pin is spelled as the user wrote it
+    // (`six==1.16` → `@1.16`) while the API keys the release as the
+    // registry published it (`@1.16.0`); pip treats both as one release
+    // (PEP 440). Ask for its equivalent spellings too.
+    let query_purls = with_pypi_equivalents(&all_purls, &lockfile_only.purls);
+    let chunks: Vec<&[String]> = batch_chunks(&query_purls, batch_size, BATCH_BODY_BYTE_CAP);
     let total_batches = chunks.len();
     let mut batch_error_count = 0usize;
     let mut last_batch_error: Option<String> = None;
@@ -2426,6 +2481,13 @@ async fn run_scan(
     // drives the table, the `--json` `packages` array and the apply order,
     // which operators diff across runs.
     all_packages_with_patches.sort_by(|a, b| a.purl.cmp(&b.purl));
+    // #604: a patch the API returned under an equivalent spelling of a
+    // lockfile-only PyPI pin is that lockfile-only package.
+    adopt_pypi_equivalents(
+        &all_packages_with_patches,
+        &all_purls,
+        &mut lockfile_only.purls,
+    );
 
     // If every batch errored, surface a full scan failure rather than
     // silently reporting zero patches.
@@ -2440,7 +2502,7 @@ async fn run_scan(
                 "status": "error",
                 "error": { "code": API_BATCH_FAILED, "message": err },
                 "scannedPackages": package_count,
-                "lockfileOnlyPackages": lockfile_only.purls.len(),
+                "lockfileOnlyPackages": lockfile_only_count,
                 "packagesWithPatches": 0,
                 "totalPatches": 0,
                 "freePatches": 0,
@@ -2520,7 +2582,7 @@ async fn run_scan(
         let mut result = serde_json::json!({
             "status": "success",
             "scannedPackages": package_count,
-            "lockfileOnlyPackages": lockfile_only.purls.len(),
+            "lockfileOnlyPackages": lockfile_only_count,
             "packagesWithPatches": all_packages_with_patches.len(),
             "totalPatches": total_patches,
             "freePatches": free_patches,

crates/socket-patch-cli/tests/mode_migration_pypi.rs

@@ -2071,6 +2071,59 @@ async fn pipenv_hosted_to_vendored_names_the_unpatched_requirements() {
     );
 }
 
+// ── #604: PEP 440-equivalent lock-only pins ──────────────────────────────
+
+/// #604: on a fresh checkout (no venv), `six==1.16` (or `==1.16.0.0`,
+/// `==01.16.0`) is the release the patch API keys as `six@1.16.0`. Hosted
+/// mode must find the patch and rewrite the pin, as it does when a venv
+/// holds six 1.16.0, and report the package as not installed.
+#[tokio::test]
+async fn lock_only_pep440_equivalent_pin_is_patched() {
+    use wiremock::matchers::body_string_contains;
+    for pin in ["1.16", "1.16.0.0", "01.16.0"] {
+        let server = MockServer::start().await;
+        // The API matches purls exactly: only `@1.16.0` has the patch.
+        Mock::given(method("POST"))
+            .and(path(format!("/v0/orgs/{ORG}/patches/batch")))
+            .and(body_string_contains(PURL))
+            .respond_with(ResponseTemplate::new(200).set_body_json(json!({
+                "packages": [{ "purl": PURL, "patches": [{
+                    "uuid": UUID, "purl": PURL, "tier": "free", "cveIds": [], "ghsaIds": [],
+                    "severity": "high", "title": "pep440 fixture"
+                }]}],
+                "canAccessPaidPatches": false,
+            })))
+            .with_priority(1)
+            .mount(&server)
+            .await;
+        Mock::given(method("POST"))
+            .and(path(format!("/v0/orgs/{ORG}/patches/batch")))
+            .respond_with(ResponseTemplate::new(200).set_body_json(json!({
+                "packages": [], "canAccessPaidPatches": false,
+            })))
+            .with_priority(2)
+            .mount(&server)
+            .await;
+        let hosted_url = mount_hosted_api(&server, true).await;
+        let (_tmp, root) = project();
+        std::fs::write(
+            root.join("requirements.txt"),
+            format!("idna==3.7\nsix=={pin}\n"),
+        )
+        .unwrap();
+        let (code, env) = hosted_scan(&root, &server);
+        assert_eq!(code, 0, "{pin}: {env:#}");
+        assert_eq!(env["redirect"]["redirected"], 1, "{pin}: {env:#}");
+        assert_eq!(env["packages"][0]["purl"], PURL, "{pin}: {env:#}");
+        assert_eq!(env["packages"][0]["notInstalled"], true, "{pin}: {env:#}");
+        let requirements = std::fs::read_to_string(root.join("requirements.txt")).unwrap();
+        assert!(
+            requirements.contains(&format!("six @ {hosted_url}")),
+            "{pin}: the pin is rewritten:\n{requirements}"
+        );
+    }
+}
+
 // ── #1138: a uv workspace member ─────────────────────────────────────────
 
 /// A uv workspace (root `pyproject.toml` with `[tool.uv.workspace]` and its

crates/socket-patch-cli/tests/scan_requirements_lock_only.rs

@@ -376,3 +376,40 @@ async fn lock_only_scan_discovers_pin_with_dangling_eof_continuation() {
     )
     .await;
 }
+
+/// #604: pip resolves `six==1.16`, `six==1.16.0.0` and `six==01.16.0` to
+/// the registry release `1.16.0` (PEP 440), but lock-only discovery queried
+/// only the spelled version, so the patch keyed `@1.16.0` was never found.
+/// Each spelling must also ask for the release's other spellings.
+#[tokio::test]
+async fn lock_only_scan_queries_pep440_equivalent_spellings() {
+    for (pin, spelled) in [
+        ("1.16", "1.16"),
+        ("1.16.0.0", "1.16.0.0"),
+        ("01.16.0", "01.16.0"),
+    ] {
+        for mode in [&[][..], &["--mode", "vendored"][..]] {
+            let mock = MockServer::start().await;
+            mount_empty_batch(&mock).await;
+            let tmp = tempfile::tempdir().unwrap();
+            std::fs::write(
+                tmp.path().join("requirements.txt"),
+                format!("sp-fixture-six=={pin}\n"),
+            )
+            .unwrap();
+            let (code, v) = run_scan(tmp.path(), &mock.uri(), mode, &[]);
+            assert_eq!(code, 0, "{pin} {mode:?}: {v}");
+            assert_eq!(v["lockfileOnlyPackages"].as_u64(), Some(1), "{v}");
+            let purls = batch_purls(&mock).await;
+            for want in [
+                format!("pkg:pypi/sp-fixture-six@{spelled}"),
+                "pkg:pypi/sp-fixture-six@1.16.0".to_string(),
+            ] {
+                assert!(
+                    purls.contains(&want),
+                    "{pin} {mode:?}: {want} must reach the patch API; sent {purls:?}"
+                );
+            }
+        }
+    }
+}

crates/socket-patch-core/src/utils/pep440.rs

@@ -130,6 +130,37 @@ pub(crate) fn is_exact_pin(specifier: &str) -> bool {
         .is_some_and(|pinned| parse(pinned).is_some())
 }
 
+/// The other spellings a registry may have published the pure release
+/// `version` (digits and dots only) under, as PEP 440 equality reads them:
+/// leading zeros dropped and the release padded with `.0` from its
+/// trailing-zero-trimmed form up to at least three segments (`1.16` →
+/// `1.16.0`; `1.16.0` → `1.16`; `01.16.0.0` → `1.16`, `1.16.0`). `version`
+/// itself is not included; anything that is not a pure release has none.
+pub(crate) fn equivalent_release_spellings(version: &str) -> Vec<String> {
+    let valid = !version.is_empty()
+        && version.split('.').all(|segment| {
+            !segment.is_empty() && segment.bytes().all(|byte| byte.is_ascii_digit())
+        });
+    if !valid {
+        return Vec::new();
+    }
+    let segments: Vec<String> = version.split('.').map(number).collect();
+    let mut trimmed = segments.clone();
+    while trimmed.len() > 1 && trimmed.last().is_some_and(|segment| segment == "0") {
+        trimmed.pop();
+    }
+    let mut out = Vec::new();
+    for len in trimmed.len()..=segments.len().max(3) {
+        let mut release = trimmed.clone();
+        release.resize(len, "0".to_string());
+        let spelling = release.join(".");
+        if spelling != version && !out.contains(&spelling) {
+            out.push(spelling);
+        }
+    }
+    out
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;
@@ -182,6 +213,32 @@ mod tests {
         }
     }
 
+    #[test]
+    fn equivalent_release_spellings_pad_and_trim() {
+        assert_eq!(equivalent_release_spellings("1.16"), ["1.16.0"]);
+        assert_eq!(equivalent_release_spellings("1.16.0"), ["1.16"]);
+        assert_eq!(equivalent_release_spellings("1.16.0.0"), ["1.16", "1.16.0"]);
+        assert_eq!(equivalent_release_spellings("01.16.0"), ["1.16", "1.16.0"]);
+        assert_eq!(equivalent_release_spellings("2"), ["2.0", "2.0.0"]);
+        assert!(equivalent_release_spellings("2.8.2").is_empty());
+        for other in [
+            "1.0rc1",
+            "2.9.0.post0",
+            "1.0+local",
+            "",
+            "1..2",
+            "v1.0",
+            "1.*",
+        ] {
+            assert!(equivalent_release_spellings(other).is_empty(), "{other}");
+        }
+        for version in ["1.16", "1.16.0.0", "01.16.0"] {
+            for spelling in equivalent_release_spellings(version) {
+                assert!(versions_equal(version, &spelling), "{version} {spelling}");
+            }
+        }
+    }
+
     #[test]
     fn exact_pin_spellings() {
         assert!(is_exact_pin_of("==1.16", "1.16.0"));

crates/socket-patch-core/src/utils/purl_key.rs

@@ -173,6 +173,43 @@ fn composer_identity(canonical: &str) -> Option<String> {
     ))
 }
 
+/// `pkg:pypi/<name>@<v>` purls for every other spelling of `purl`'s pure
+/// release (`@1.16` → `@1.16.0`, see
+/// [`crate::utils::pep440::equivalent_release_spellings`]): the lockfile
+/// spells a pin as the user wrote it (`six==1.16`), while the patch API
+/// keys the release as the registry published it (#604). Empty for any
+/// other purl.
+pub fn pypi_equivalent_purls(purl: &str) -> Vec<String> {
+    let canonical = canonical_base_purl(purl);
+    let Some((name, version)) = canonical
+        .strip_prefix("pkg:pypi/")
+        .and_then(|rest| rest.rsplit_once('@'))
+    else {
+        return Vec::new();
+    };
+    crate::utils::pep440::equivalent_release_spellings(version)
+        .into_iter()
+        .map(|spelling| format!("pkg:pypi/{name}@{spelling}"))
+        .collect()
+}
+
+/// Whether two PyPI purls name the same release under PEP 440 equality
+/// (`@1.16` and `@1.16.0`), the way pip resolves an `==` pin. `false` for
+/// anything that is not two PyPI purls of one (PEP 503) name.
+pub fn pypi_same_release(a: &str, b: &str) -> bool {
+    let (a, b) = (canonical_base_purl(a), canonical_base_purl(b));
+    let split = |purl: &str| -> Option<(String, String)> {
+        let (name, version) = purl.strip_prefix("pkg:pypi/")?.rsplit_once('@')?;
+        Some((name.to_string(), version.to_string()))
+    };
+    match (split(&a), split(&b)) {
+        (Some((name_a, version_a)), Some((name_b, version_b))) => {
+            name_a == name_b && crate::utils::pep440::versions_equal(&version_a, &version_b)
+        }
+        _ => false,
+    }
+}
+
 #[cfg(test)]
 #[path = "purl_key_nuget_vendor_tests.rs"]
 mod nuget_vendor_tests;

docs/testing/uv-compatibility.md

@@ -165,6 +165,11 @@ frozen, locked, and ordinary installation outcomes separately where supported.
   those cases requires marker-specific source mappings. Standalone PEP 751
   rewriting selects the exact package version; duplicate entries for the same
   name and version are refused when source selection is ambiguous.
+- Lock-only discovery (a fresh checkout, no venv) queries the patch API with
+  every PEP 440 spelling of an exact pure-release pin as well as the one
+  written (`six==1.16` asks for `@1.16` and `@1.16.0`; `==1.16.0` also asks for
+  `@1.16`), so it finds the patch the registry keys under its own spelling, as a
+  venv-backed run does, and reports the package as not installed (#604).
 - Hosted requirements select exact `==`/`===` pins or socket-patch's own
   hosted archive URLs. A user-authored direct reference to the patched release
   (`name @ <url>` on any other origin, `files.pythonhosted.org` and `file://`