[#242] No scan symlinks as md files

Problem: when the repository contains a symlink to a markdown file, it
is processed by xrefcheck as if it was the same markdown file but in the
symlink's location. This leads to broken references and can be avoided
because neither GitHub nor GitLab try to render symlinks as the file
they point to.

Solution: consider symlinks as no scannable files. In the future, we
will consider to include a new dedicated scanner for symlinks if it
works.

src/Xrefcheck/Scan.hs

@@ -264,7 +264,10 @@ scanRepo scanMode rw formatsSupport config root = do
     gatherFileStatuses = map (second fst)
 
     processFile :: CanonicalPath -> IO (FileStatus, [ScanError 'Parse])
-    processFile canonicalFile = case mscanner canonicalFile of
+    processFile canonicalFile =
+      ifM (pathIsSymbolicLink canonicalFile)
+      (pure (NotScannable, []))
+      case mscanner canonicalFile of
         Nothing -> pure (NotScannable, [])
         Just scanner -> scanner canonicalFile <&> _1 %~ Scanned
 

src/Xrefcheck/System.hs

@@ -14,6 +14,7 @@ module Xrefcheck.System
   , getPosixRelativeChild
   , getPosixRelativeOrAbsoluteChild
   , hasIndirectionThroughParent
+  , pathIsSymbolicLink
   , takeDirectory
   , takeExtension
   , (</)
@@ -94,6 +95,10 @@ takeDirectory (UnsafeCanonicalPath p) = UnsafeCanonicalPath $ FP.takeDirectory p
 takeExtension :: CanonicalPath -> String
 takeExtension (UnsafeCanonicalPath p) = FP.takeExtension p
 
+-- | 'System.Directory.pathIsSymbolicLink' version for 'CanonicalPath'.
+pathIsSymbolicLink :: CanonicalPath -> IO Bool
+pathIsSymbolicLink (UnsafeCanonicalPath p) = Directory.pathIsSymbolicLink p
+
 -- | Get the list of directories, canonicalized, between two given paths.
 getDirsBetweenRootAndFile :: CanonicalPath -> CanonicalPath -> [CanonicalPath]
 getDirsBetweenRootAndFile (UnsafeCanonicalPath rootPath) file =

tests/golden/check-symlinks/check-symlinks.bats

@@ -0,0 +1,17 @@
+#!/usr/bin/env bats
+
+# SPDX-FileCopyrightText: 2022 Serokell <https://serokell.io>
+#
+# SPDX-License-Identifier: MPL-2.0
+
+load '../helpers/bats-support/load'
+load '../helpers/bats-assert/load'
+load '../helpers/bats-file/load'
+load '../helpers'
+
+
+@test "Checking that symlinks are not processed" {
+  to_temp xrefcheck -v
+
+  assert_diff expected.gold
+}

tests/golden/check-symlinks/dir/b.md

@@ -0,0 +1,11 @@
+<!--
+ - SPDX-FileCopyrightText: 2022 Serokell <https://serokell.io>
+ -
+ - SPDX-License-Identifier: MPL-2.0
+ -->
+
+[Empty file](a)
+
+[Symlink ref](../s.md)
+
+[Symlink ref with anchor](../s.md#a)

tests/golden/check-symlinks/expected.gold

@@ -0,0 +1,20 @@
+=== Repository data ===
+
+  dir/b.md:
+    - references:
+        - reference (relative) at src:7:1-15:
+            - text: "Empty file"
+            - link: a
+            - anchor: -
+        - reference (relative) at src:9:1-22:
+            - text: "Symlink ref"
+            - link: ../s.md
+            - anchor: -
+        - reference (relative) at src:11:1-36:
+            - text: "Symlink ref with anchor"
+            - link: ../s.md
+            - anchor: a
+    - anchors:
+        none
+
+All repository links are valid.

tests/golden/check-symlinks/s.md

@@ -0,0 +1 @@
+dir/b.md