Keep the in-memory hosted engine off the disk for Go workspaces

crates/socket-patch-core/src/hosted/engine.rs

@@ -487,64 +487,63 @@ async fn read_advisory(
     }
 }
 
+/// An in-memory project's Go files as the hosted engine reads them
+/// ([`read_advisory`]: a link or an entry without content is unreadable,
+/// but still exists).
+struct AdvisoryGoFiles<'v, 'p> {
+    view: &'v ProjectView<'p>,
+    unreadable: &'v BTreeSet<String>,
+}
+
+impl crate::vendor::go_workspace::MemoryFiles for AdvisoryGoFiles<'_, '_> {
+    async fn exists(&mut self, rel: &str) -> bool {
+        self.unreadable.contains(rel) || self.view.exists_no_follow(rel).await
+    }
+
+    async fn read(&mut self, rel: &str) -> Option<String> {
+        read_advisory(self.view, self.unreadable, rel).await
+    }
+
+    async fn read_sum(&mut self, rel: &str) -> Option<String> {
+        read_advisory(self.view, self.unreadable, rel).await
+    }
+
+    fn unparseable(&mut self, _rel: &str, _problem: &str) {}
+}
+
 /// The Go workspace in effect for the project, handed to the golang
 /// rewriter under [`crate::patch::redirect::GO_WORKSPACE_KEY`]. On disk it is
 /// the one go resolves (`GOWORK`, else a go.work at or above the root, with
 /// members anywhere, plus a `-modfile` that GOFLAGS sets); in memory a root
-/// `go.work` and its in-root members. A root go.work and go.work.sum are candidate files the rewriter writes.
+/// `go.work` and its in-root members, read as discovery reads them
+/// ([`crate::vendor::go_workspace::read_memory`]), whose paths every check
+/// compares lexically, never probing the disk or the process's working
+/// directory. A root go.work and go.work.sum are candidate files the
+/// rewriter writes.
 async fn read_go_workspace(
     view: &ProjectView<'_>,
     unreadable: &BTreeSet<String>,
     out: &mut CandidateFiles,
 ) {
-    use crate::vendor::go_workspace::{self, GoWorkspace, Loaded, Member};
+    use crate::vendor::go_workspace::{self, Loaded};
     let (root, loaded, modfile, module_graph) = match view {
-        ProjectView::Memory(_) => 'memory: {
+        ProjectView::Memory(_) => {
             let Some(work) = read_advisory(view, unreadable, "go.work").await else {
                 return;
             };
             let work = crate::vendor::go_mod_edit::normalize_for_read(&work);
             // As on disk (`go_workspace::load`): a go.work go cannot parse
             // is a workspace go refuses to build, never module mode.
-            if let Some(problem) = crate::vendor::go_mod_edit::block_structure_error(&work) {
-                break 'memory (
-                    std::path::PathBuf::new(),
-                    Loaded::Invalid(format!("go.work: {problem}, so go refuses to build")),
-                    None,
-                    None,
-                );
-            }
-            let mut members = Vec::new();
-            for raw in crate::vendor::go_mod_edit::parse_use_dirs(&work) {
-                let Some(dir) = go_workspace::in_root_dir(&raw) else {
-                    members.push(Member {
-                        dir: raw.into(),
-                        go_mod: None,
-                        go_sum: None,
-                    });
-                    continue;
-                };
-                let at = |file: &str| match dir.as_str() {
-                    "" => file.to_string(),
-                    dir => format!("{dir}/{file}"),
-                };
-                members.push(Member {
-                    go_mod: read_advisory(view, unreadable, &at("go.mod"))
-                        .await
-                        .map(|t| crate::vendor::go_mod_edit::normalize_for_read(&t)),
-                    go_sum: read_advisory(view, unreadable, &at("go.sum")).await,
-                    dir: dir.into(),
-                });
-            }
-            let work_sum = read_advisory(view, unreadable, "go.work.sum").await;
-            let ws = GoWorkspace {
-                work_file: "go.work".into(),
-                work,
-                work_sum,
-                members,
-                graph: None,
+            let loaded = match crate::vendor::go_mod_edit::block_structure_error(&work) {
+                Some(problem) => {
+                    Loaded::Invalid(format!("go.work: {problem}, so go refuses to build"))
+                }
+                None => {
+                    let mut files = AdvisoryGoFiles { view, unreadable };
+                    Loaded::Workspace(go_workspace::read_memory(&work, &mut files).await.0)
+                }
             };
-            (std::path::PathBuf::new(), Loaded::Workspace(ws), None, None)
+            (std::path::PathBuf::new(), loaded, None, None)
         }
         ProjectView::Disk(_) | ProjectView::Snapshot(_) => {
             let Some(root) = view.disk_root_reading(std::iter::empty::<&std::path::Path>()) else {
@@ -3691,6 +3690,120 @@ mod tests {
         );
     }
 
+    /// The in-memory engine's workspace is judged lexically, from the
+    /// project's own files: a root module go.work doesn't `use` is "not a
+    /// member", whatever the process's working directory holds. The disk
+    /// checks probed `<cwd>/go.mod` instead, so with this crate's directory
+    /// as the cwd (no go.mod) the rewriter called the root module-less
+    /// ("nowhere to go"), and with a go.mod there it gave the member answer.
+    /// A root go.mod go cannot parse, or one that cannot be read (a link),
+    /// still exists: the same files, the same verdict, as discovery's
+    /// `in_memory_workspace_files_are_the_files_that_exist` finds.
+    #[tokio::test]
+    async fn memory_go_workspace_never_probes_the_cwd() {
+        use crate::patch::redirect::{
+            rewrite_registry_redirect, HostedGoWorkspace, Integrity, RegistryOverride,
+            RegistryOverrideIdentifiers, GO_WORKSPACE_KEY,
+        };
+        use crate::vendor::go_workspace::{Loaded, WorkspaceConflict};
+        const MODULE: &str = "github.com/foo/bar";
+        let socket = format!("patch.socket.dev/gopatch/{PATCH_UUID}");
+        let hash = format!("h1:{}=", "A".repeat(43));
+        let ovr = DepOverride {
+            ecosystem: "golang".into(),
+            name: MODULE.into(),
+            version: "v1.4.2".into(),
+            registry_override: Some(RegistryOverride {
+                kind: "goproxy".into(),
+                index_url: "https://patch.socket.dev/patch-registry/golang".into(),
+                identifiers: RegistryOverrideIdentifiers {
+                    name: MODULE.into(),
+                    version: "v1.4.2".into(),
+                    go_module_path: Some(socket.clone()),
+                    go_module_version: Some("v1.4.2-socketpatch.1".into()),
+                    ..Default::default()
+                },
+            }),
+            integrity: Integrity {
+                dirhash_h1: Some(hash.clone()),
+                go_mod_h1: Some(hash),
+                ..Default::default()
+            },
+            ..grant_dep("golang", String::new())
+        };
+        let go_mod = format!("module example.com/app\n\ngo 1.21\n\nrequire {MODULE} v1.4.2\n");
+        let b = format!("module example.com/b\n\ngo 1.21\n\nrequire {MODULE} v1.4.2\n");
+        let work = "go 1.21\n\nuse ./b\n";
+        let ours = crate::vendor::go_mod_edit::ReplaceEntry {
+            module: MODULE.into(),
+            version: Some("v1.4.2".into()),
+            path: None,
+            rhs_module: Some(socket),
+            rhs_version: Some("v1.4.2-socketpatch.1".into()),
+            owner: Some(crate::vendor::go_mod_edit::ReplaceOwner::Hosted),
+        };
+        let unparseable = MemoryEntry::Text("module example.com/app\n\nrequire (\n".into());
+        for root in [
+            MemoryEntry::Text(go_mod.as_str().into()),
+            unparseable,
+            MemoryEntry::Symlink,
+        ] {
+            let mut p = MemoryProject::new();
+            p.insert("go.mod", root.clone());
+            p.insert_text("go.work", work);
+            p.insert_text("b/go.mod", b.as_str());
+            let mut out = CandidateFiles::default();
+            read_go_workspace(&ProjectView::Memory(&p), &BTreeSet::new(), &mut out).await;
+            let hosted: HostedGoWorkspace =
+                serde_json::from_str(&out.files[GO_WORKSPACE_KEY]).unwrap();
+            let Loaded::Workspace(ws) = &hosted.loaded else {
+                panic!("{root:?}: {:?}", hosted.loaded);
+            };
+            assert_eq!(
+                ws.files,
+                Some(
+                    ["go.mod", "go.work", "b/go.mod"]
+                        .into_iter()
+                        .map(std::path::PathBuf::from)
+                        .collect()
+                ),
+                "{root:?}"
+            );
+            assert!(
+                matches!(
+                    hosted
+                        .loaded
+                        .conflict(&hosted.root, MODULE, "v1.4.2", &ours),
+                    Some(WorkspaceConflict::NotAMember { .. })
+                ),
+                "{root:?}"
+            );
+        }
+        // End to end through the rewriter.
+        let mut p = MemoryProject::new();
+        p.insert_text("go.mod", go_mod.as_str());
+        p.insert_text("go.work", work);
+        p.insert_text("b/go.mod", b.as_str());
+        let mut read = CandidateFiles::default();
+        read_go_workspace(&ProjectView::Memory(&p), &BTreeSet::new(), &mut read).await;
+        let mut files = read.files;
+        for (path, text) in [("go.mod", go_mod.as_str()), ("go.work", work)] {
+            files.insert(path.into(), text.into());
+        }
+        let out = rewrite_registry_redirect(&files, std::slice::from_ref(&ovr));
+        assert!(out.files.is_empty(), "{:?}", out.files);
+        assert!(out.confirmed_golang_uuids.is_empty());
+        assert_eq!(out.warnings.len(), 1, "{:?}", out.warnings);
+        assert_eq!(out.warnings[0].code, "redirect_golang_workspace_conflict");
+        assert!(
+            out.warnings[0]
+                .detail
+                .contains("go.work does not `use` this module"),
+            "{:?}",
+            out.warnings
+        );
+    }
+
     fn no_cargo_outer() -> CratesIoSource {
         CratesIoSource::Registry
     }

crates/socket-patch-core/src/patch/redirect/mod.rs

@@ -24245,8 +24245,13 @@ packages:
     }
 
     /// The engine's [`GO_WORKSPACE_KEY`] for an in-memory go.work whose
-    /// members are `(dir, go.mod)` pairs (`.` for the root module).
-    fn go_workspace_key(work_extra: &str, members: &[(&str, &str)]) -> (String, String) {
+    /// members are `(dir, go.mod)` pairs (`.` for the root module), in a
+    /// project with a root go.mod when `root_go_mod`.
+    fn go_workspace_key(
+        work_extra: &str,
+        members: &[(&str, &str)],
+        root_go_mod: bool,
+    ) -> (String, String) {
         use crate::vendor::go_workspace::{GoWorkspace, Loaded, Member};
         let uses: String = members.iter().map(|(d, _)| format!("\t{d}\n")).collect();
         let hosted = HostedGoWorkspace {
@@ -24278,6 +24283,10 @@ packages:
                         })
                         .collect(),
                 ),
+                files: Some(memory_go_files(
+                    members.iter().map(|(dir, _)| *dir),
+                    root_go_mod,
+                )),
             }),
             modfile: None,
             module_graph: None,
@@ -24288,14 +24297,37 @@ packages:
         )
     }
 
+    /// The files an in-memory workspace of the `use` members `dirs` holds,
+    /// as [`crate::vendor::go_workspace::read_memory`] lists them: go.work,
+    /// the root go.mod when the project has one (`root_go_mod`, whether or
+    /// not `.` is a member) and each in-root member's go.mod.
+    fn memory_go_files<'a>(
+        dirs: impl IntoIterator<Item = &'a str>,
+        root_go_mod: bool,
+    ) -> std::collections::BTreeSet<std::path::PathBuf> {
+        let mut files = std::collections::BTreeSet::from(["go.work".into()]);
+        if root_go_mod {
+            files.insert("go.mod".into());
+        }
+        for dir in dirs {
+            files.extend(crate::vendor::go_workspace::in_root_dir(dir).map(
+                |dir| match dir.as_str() {
+                    "" => "go.mod".into(),
+                    dir => format!("{dir}/go.mod").into(),
+                },
+            ));
+        }
+        files
+    }
+
     /// [`go_workspace_key`] plus the root go.work itself as a candidate
     /// file, as the engine reads it when go.work sits at the root.
     fn with_root_go_work(
         files: &mut BTreeMap<String, String>,
         work_extra: &str,
         members: &[(&str, &str)],
     ) {
-        let (k, v) = go_workspace_key(work_extra, members);
+        let (k, v) = go_workspace_key(work_extra, members, files.contains_key("go.mod"));
         let uses: String = members.iter().map(|(d, _)| format!("\t{d}\n")).collect();
         files.insert(k, v);
         files.insert(
@@ -24324,8 +24356,11 @@ packages:
             ("\nreplace github.com/foo/bar => ./fork\n", ""),
         ] {
             let mut files = files.clone();
-            let (k, v) =
-                go_workspace_key(work, &[(".", &root), ("b", &go_member("v1.4.2", other))]);
+            let (k, v) = go_workspace_key(
+                work,
+                &[(".", &root), ("b", &go_member("v1.4.2", other))],
+                true,
+            );
             files.insert(k, v);
             let out = rewrite_registry_redirect(&files, std::slice::from_ref(&ovr));
             assert!(out.files.is_empty(), "{work}{other}: {:?}", out.files);
@@ -24343,6 +24378,7 @@ packages:
         let (k, v) = go_workspace_key(
             "\nreplace github.com/foo/bar v1.0.0 => ./fork\n",
             &[(".", &root), ("b", &go_member("v1.4.2", ""))],
+            true,
         );
         files.insert(k, v);
         let out = rewrite_registry_redirect(&files, std::slice::from_ref(&ovr));
@@ -24358,7 +24394,7 @@ packages:
         let ovr = golang_override();
         let mut files = golang_files();
         let root = files["go.mod"].clone();
-        let (k, v) = go_workspace_key("", &[(".", &root), ("b", &go_member("v1.5.0", ""))]);
+        let (k, v) = go_workspace_key("", &[(".", &root), ("b", &go_member("v1.5.0", ""))], true);
         files.insert(k, v);
         let out = rewrite_registry_redirect(&files, std::slice::from_ref(&ovr));
         assert!(out.files.is_empty());
@@ -24368,7 +24404,7 @@ packages:
         let bare = "module example.com/app\n\ngo 1.21\n";
         let mut files = BTreeMap::new();
         files.insert("go.mod".to_string(), bare.to_string());
-        let (k, v) = go_workspace_key("", &[(".", bare), ("b", &go_member("v1.4.2", ""))]);
+        let (k, v) = go_workspace_key("", &[(".", bare), ("b", &go_member("v1.4.2", ""))], true);
         files.insert(k, v);
         let out = rewrite_registry_redirect(&files, std::slice::from_ref(&ovr));
         assert!(out.warnings.is_empty(), "{:?}", out.warnings);
@@ -24501,6 +24537,7 @@ packages:
                     member("../out", None),
                 ],
                 graph: None,
+                files: Some(memory_go_files([".", "../out"], true)),
             }),
             modfile: None,
             module_graph: None,