{"entries":[{"lastCommit":{"author":{"email":"matthew@cue.works","name":"Matthew Sackman","when":"2026-02-02T12:17:07.000000Z"},"committer":{"email":"matthew@cue.works","name":"Matthew Sackman","when":"2026-02-11T15:18:55.000000Z"},"hash":"6825394279bd8546fbdde42383fb82c4ad97a321","message":"lsp/fscache: read json and yaml files as CUE\n\n[module.ReadCUEFile] takes an arbitrary file path to read as a CUE AST:\nthere's no requirement that the file path be to a cue file. However its\none call site in modimports called it with only file paths with a \".cue\"\nsuffix.\n\nBy relaxing this (giving the implementation of [module.ReadCUEFile] the\noption of converting non-.cue files to a CUE AST), and by making some\nminor changes to fscache, the fscache can read json and yaml files, and\nconvert them to CUE ASTs.\n\nThis means that the LSP, with suitable editor configuration, can work\nwith json and yaml files too: for these files, they are treated as\nstandalone packages. But the existing machinery managing their\nlifecycle, reloading and so forth, all works without further changes.\n\nThere is one other part of the modpkgload package that tests for \".cue\"\nsuffix, and that's in isDirWithCUEFiles. We extend the dirfs.ReadCUEFS\ninterface, adding a IsDirWithCUEFiles method, and so allowing the\nimplementation to decide whether the directory in question contains any\nfiles which it would attempt to convert to CUE ASTs.\n\nSigned-off-by: Matthew Sackman \nChange-Id: Ie71b2b3d4f661f6e5ad82e0bb46bf82535e84b0f\nReviewed-on: https://cue.gerrithub.io/c/cue-lang/cue/+/1230265\nUnity-Result: CUE porcuepine \nTryBot-Result: CUEcueckoo \nReviewed-by: Daniel Martí \n","tree":"ef52af287808f824dcb486d5d68d2cd723be495b"},"mode":"0100644","name":"dirfs.go","size":2160},{"lastCommit":{"author":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2024-03-02T20:19:40.000000Z"},"committer":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2024-03-04T11:00:36.000000Z"},"hash":"6cba4c0b2dee8009f3c2578980312506f805af73","message":"mod/...: remove three somewhat unnecessary APIs\n\nUsing golang.org/x/tools/cmd/deadcode via the command\n\n deadcode ./... | grep '^mod/'\n\nI spotted a number of APIs which weren't reachable by cmd/cue,\nwhich is a signal that perhaps they're not very useful or necessary.\nGoing one by one:\n\n mod/modregistry/client.go:86:6: unreachable func: NewClient\n mod/modregistry/client.go:471:25: unreachable func: SingleResolver.ResolveToRegistry\n\nNewClient uses SingleResolver to interact with a registry interface:\n\n modregistry.NewClient(reg)\n\nOne could create the resolver if we properly exposed SingleResolver:\n\n modregistry.NewClientWithResolver(modregistry.SingleResolver(reg))\n\nHowever, SingleResolver is an opaque struct that can't be used directly,\nand it is only useful as part of NewClient anyway, so NewClient is best.\nKeep NewClient as-is for the common use case, and hide SingleResolver,\nsince no third party user should have a direct need for it.\n\n mod/module/error.go:17:6: unreachable func: VersionError\n\nThe returned ModuleError type is used, but this func is never used.\n\n mod/module/module.go:165:6: unreachable func: MustParseVersion\n mod/module/module.go:177:6: unreachable func: ParseVersion\n\nWhile ParseVersion is only used in the tests, it's the only way to\nconstruct a module.Version from a string, so it's good for consistency.\nThe \"must\" helpers like MustParseVersion and MustNewVersion\nseem to exist mainly for the sake of writing tests,\nso I would personally remove them, but Roger prefers to leave them.\n\n mod/module/path.go:483:6: unreachable func: MatchPathMajor\n\nThis is a shortcut for `CheckPathMajor(v, pathMajor) == nil`\nand is never used, not even in the tests. Remove it.\n\n mod/modzip/zip.go:201:6: unreachable func: CheckFiles\n mod/modzip/zip.go:369:6: unreachable func: CheckDir\n mod/modzip/zip.go:396:6: unreachable func: CheckZipFile\n\nThese aren't used by our implementation, as it only needs CheckZip,\nbut they make the API a bit friendlier to use.\nCheckZip takes a ReaderAt plus a size, which is a bit awkward,\nwhereas the others take more straightforward values such a a filename.\n\nFor instance, trying to rewrite the tests to use CheckZip\nmade me realise that most users will not want to use CheckZip directly.\n\n mod/modzip/zip.go:886:22: unreachable func: ZipFileIO.Path\n mod/modzip/zip.go:889:18: unreachable func: ZipFileIO.Lstat\n mod/modzip/zip.go:892:18: unreachable func: ZipFileIO.Open\n\nThis implementation of FileIO is never used; while it seems nice to have\ngiven that the package is already geared towards zip archives,\nthe FileIO abstraction is for the slice taken by CheckFiles,\nand CheckZip works entirely differently and does not use FileIO at all.\n\nWhile the FileIO abstraction seems nice as we use it elsewhere\nfor CheckFiles on txtar archives, ZipFileIO seems unnecessary for now.\n\nFor #2896.\n\nSigned-off-by: Daniel Martí \nChange-Id: Ib57e4b224ba3ee7c2d713395b50ac9942c56034c\nReviewed-on: https://review.gerrithub.io/c/cue-lang/cue/+/1177674\nReviewed-by: Roger Peppe \nTryBot-Result: CUEcueckoo \nUnity-Result: CUE porcuepine \n","tree":"b9a61e6a11b2aa8eba70e34d3f8733341a419503"},"mode":"0100644","name":"error.go","size":1482},{"lastCommit":{"author":{"email":"rogpeppe@gmail.com","name":"Roger Peppe","when":"2024-02-26T11:24:40.000000Z"},"committer":{"email":"rogpeppe@gmail.com","name":"Roger Peppe","when":"2024-02-29T10:50:23.000000Z"},"hash":"272776663738b3175c3a7eb8b901bf15d0f10ba5","message":"mod/module: move from internal/mod/module\n\nThis is one stage towards making the modules API public.\n\nFor #2874.\n\nSigned-off-by: Roger Peppe \nChange-Id: I96cca53ecc714f4fd92ee298a9934e43be4994a1\nReviewed-on: https://review.gerrithub.io/c/cue-lang/cue/+/1177410\nTryBot-Result: CUEcueckoo \nUnity-Result: CUE porcuepine \nReviewed-by: Daniel Martí \n","tree":"35b502da59e50cba521e9047dfd5816b7d3e2eb9"},"mode":"0100644","name":"escape.go","size":1770},{"lastCommit":{"author":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2025-10-21T22:17:58.000000+01:00"},"committer":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2025-10-22T16:33:01.000000Z"},"hash":"70d6a9c84be275aad1ae6a297455e6f88b0c1965","message":"all: start leveraging `go fix -inline`\n\nAdd `//go:fix inline` to deprecated APIs which can be inlined directly\nto resolve the deprecation warning:\n* cue/errors.NewMessage\n* mod/module.SplitPathVersion\n* mod/module.ImportPath\n* mod/module.ParseImportPath\n\nAnd run `go fix -inline ./...` to apply these rewrites.\n\nSigned-off-by: Daniel Martí \nChange-Id: Iea15035e204d6db99c1f2d56f680068233b6d908\nReviewed-on: https://cue.gerrithub.io/c/cue-lang/cue/+/1224692\nUnity-Result: CUE porcuepine \nTryBot-Result: CUEcueckoo \nReviewed-by: Roger Peppe \n","tree":"41ffcf73e3c8724d6eb39541e3a75a079a189d08"},"mode":"0100644","name":"module.go","size":9103},{"lastCommit":{"author":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2026-03-17T16:24:07.000000Z"},"committer":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2026-03-17T17:05:58.000000Z"},"hash":"1de56c5730b9b7c9bbab04eb3d321312777f4814","message":"mod/module,cue/load,internal/filetypes: reject version in import qualifier\n\nGive a clear error when a version appears inside the package qualifier\nof an import path, such as \"foo.com/bar:baz@v0\" or \"foo.com/bar@v0:baz@v1\",\nrather than confusing messages like \"no files in package directory with\npackage name other@v0\" or \"cannot combine scope with file\".\n\nThe validation is in CheckImportPath, with an early check in the loader\nfor CLI arguments that may not reach CheckImportPath (e.g. no module).\nIsPackage uses the presence of \"/\" in the path to distinguish package\npaths from filetype scopes when the qualifier isn't a valid identifier.\n\nFixes #2859.\n\nSigned-off-by: Daniel Martí \nChange-Id: I92e4e9e50e6f56dc35bf51c97f073ca83b1248c5\nReviewed-on: https://review.gerrithub.io/c/cue-lang/cue/+/1233779\nUnity-Result: CUE porcuepine \nReviewed-by: Roger Peppe \nTryBot-Result: CUEcueckoo \n","tree":"00ac74365eb6fb9bcf88f4d3af134333ccd4b04b"},"mode":"0100644","name":"module_test.go","size":21448},{"lastCommit":{"author":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2026-03-17T16:24:07.000000Z"},"committer":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2026-03-17T17:05:58.000000Z"},"hash":"1de56c5730b9b7c9bbab04eb3d321312777f4814","message":"mod/module,cue/load,internal/filetypes: reject version in import qualifier\n\nGive a clear error when a version appears inside the package qualifier\nof an import path, such as \"foo.com/bar:baz@v0\" or \"foo.com/bar@v0:baz@v1\",\nrather than confusing messages like \"no files in package directory with\npackage name other@v0\" or \"cannot combine scope with file\".\n\nThe validation is in CheckImportPath, with an early check in the loader\nfor CLI arguments that may not reach CheckImportPath (e.g. no module).\nIsPackage uses the presence of \"/\" in the path to distinguish package\npaths from filetype scopes when the qualifier isn't a valid identifier.\n\nFixes #2859.\n\nSigned-off-by: Daniel Martí \nChange-Id: I92e4e9e50e6f56dc35bf51c97f073ca83b1248c5\nReviewed-on: https://review.gerrithub.io/c/cue-lang/cue/+/1233779\nUnity-Result: CUE porcuepine \nReviewed-by: Roger Peppe \nTryBot-Result: CUEcueckoo \n","tree":"00ac74365eb6fb9bcf88f4d3af134333ccd4b04b"},"mode":"0100644","name":"path.go","size":13504},{"lastCommit":{"author":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2025-10-16T09:57:24.000000+01:00"},"committer":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2025-10-20T17:03:24.000000Z"},"hash":"5b42b3897ea746396967de789b6431a5f0880dfa","message":"all: fix a number of bad godocs spotted by gopls\n\nMostly typos of the declared name, sentences which mentioned the name\nlater on rather than at the beginning, as well as some comments\nwhich did not intend to be a godoc at all.\n\nWhile here, we tweak a few to use links too.\n\nSigned-off-by: Daniel Martí \nChange-Id: I0ef2f22ddfc6b1e3b67acbbaa7c74cb6187eae1b\nReviewed-on: https://review.gerrithub.io/c/cue-lang/cue/+/1224421\nReviewed-by: Roger Peppe \nUnity-Result: CUE porcuepine \nTryBot-Result: CUEcueckoo \n","tree":"21093e95d53e02b0f59677cc9297ed5c8cee173b"},"mode":"0100644","name":"versions.go","size":906},{"lastCommit":{"author":{"email":"rogpeppe@gmail.com","name":"Roger Peppe","when":"2024-02-26T11:24:40.000000Z"},"committer":{"email":"rogpeppe@gmail.com","name":"Roger Peppe","when":"2024-02-29T10:50:23.000000Z"},"hash":"272776663738b3175c3a7eb8b901bf15d0f10ba5","message":"mod/module: move from internal/mod/module\n\nThis is one stage towards making the modules API public.\n\nFor #2874.\n\nSigned-off-by: Roger Peppe \nChange-Id: I96cca53ecc714f4fd92ee298a9934e43be4994a1\nReviewed-on: https://review.gerrithub.io/c/cue-lang/cue/+/1177410\nTryBot-Result: CUEcueckoo \nUnity-Result: CUE porcuepine \nReviewed-by: Daniel Martí \n","tree":"35b502da59e50cba521e9047dfd5816b7d3e2eb9"},"mode":"0100644","name":"versions_test.go","size":152}],"lastCommit":{"author":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2026-03-17T16:24:07.000000Z"},"committer":{"email":"mvdan@mvdan.cc","name":"Daniel Martí","when":"2026-03-17T17:05:58.000000Z"},"hash":"1de56c5730b9b7c9bbab04eb3d321312777f4814","message":"mod/module,cue/load,internal/filetypes: reject version in import qualifier\n\nGive a clear error when a version appears inside the package qualifier\nof an import path, such as \"foo.com/bar:baz@v0\" or \"foo.com/bar@v0:baz@v1\",\nrather than confusing messages like \"no files in package directory with\npackage name other@v0\" or \"cannot combine scope with file\".\n\nThe validation is in CheckImportPath, with an early check in the loader\nfor CLI arguments that may not reach CheckImportPath (e.g. no module).\nIsPackage uses the presence of \"/\" in the path to distinguish package\npaths from filetype scopes when the qualifier isn't a valid identifier.\n\nFixes #2859.\n\nSigned-off-by: Daniel Martí \nChange-Id: I92e4e9e50e6f56dc35bf51c97f073ca83b1248c5\nReviewed-on: https://review.gerrithub.io/c/cue-lang/cue/+/1233779\nUnity-Result: CUE porcuepine \nReviewed-by: Roger Peppe \nTryBot-Result: CUEcueckoo \n","tree":"00ac74365eb6fb9bcf88f4d3af134333ccd4b04b"}}