feat(parser): handle commit ids in tree URIs

README.md

@@ -438,18 +438,24 @@ parsed fields.
   - `https://github.com/<repo>`
   - `https://github.com/<repo>.git`
   - `https://github.com/<repo>/tree/<branch>`
+  - `https://github.com/<repo>/tree/<commit>`
   - `https://github.com/<repo>/tree/<tag>`
   - `https://github.com/<repo>/blob/<branch>`
   - `https://github.com/<repo>/blob/<branch>/<file_path>`
+  - `https://github.com/<repo>/blob/<commit>`
+  - `https://github.com/<repo>/blob/<commit>/<file_path>`
   - `https://github.com/<repo>/blob/<tag>`
   - `https://github.com/<repo>/blob/<tag>/<file_path>`
 - GitLab
   - `https://gitlab.com/<repo>`
   - `https://gitlab.com/<repo>.git`
   - `https://gitlab.com/<repo>/-/tree/<branch>`
+  - `https://gitlab.com/<repo>/-/tree/<commit>`
   - `https://gitlab.com/<repo>/-/tree/<tag>`
   - `https://gitlab.com/<repo>/-/blob/<branch>`
   - `https://gitlab.com/<repo>/-/blob/<branch>/<file_path>`
+  - `https://gitlab.com/<repo>/-/blob/<commit>`
+  - `https://gitlab.com/<repo>/-/blob/<commit>/<file_path>`
   - `https://gitlab.com/<repo>/-/blob/<tag>`
   - `https://gitlab.com/<repo>/-/blob/<tag>/<file_path>`
 - Gitea

lua/git-dev/init.lua

@@ -229,7 +229,7 @@ M.open = function(repo, ref, opts)
     or parsed_repo.commit
     or parsed_repo.branch
 
-  if not branch and parsed_repo.full_blob then
+  if not branch then
     ui:print "Could not detect a ref in given URI, falling back to default."
   end
 

lua/git-dev/parser.lua

@@ -29,6 +29,10 @@ local function is_git_bundle(file_path)
   return sig == "git\001\n"
 end
 
+local function is_commit_id(text)
+  return text:find("^" .. ("%x"):rep(40) .. "$")
+end
+
 ---@class Parser
 ---@field gitcmd table
 ---@field base_uri_format string
@@ -50,7 +54,6 @@ end
 
 ---@class GitDevParsedRepo
 ---@field repo_url string
----@field full_blob? string
 ---@field commit? string
 ---@field branch? string
 ---@field selected_path? string
@@ -59,10 +62,10 @@ end
 -- Since branch / tag might include slashes, it is impossible to determine how
 -- to separate it from the optional trailing file path. `git ls-remote` is used
 -- to find the longest ref name that can be deducted from `text`. If none is
--- found, it will be tested as a commit ID. If it is invalid as a commit, it
--- will be notated as `full_blob`. Otherwise, it will be separated to `branch`
--- and `selected_path`. If there are no slashes, it can be assumed that the
--- full blob is actually a branch / tag name, and there is no trailed file path.
+-- found, it will be treated as an optionally partial commit ID and the rest of
+-- the blob will be treated as a selected path. If there are no slashes, it can
+-- be assumed that the full blob is actually a branch / tag name, and there is
+-- no trailed file path.
 function Parser:parse_full_blob(text, repo_url)
   if not text or text == "" then
     return {}
@@ -90,18 +93,12 @@ function Parser:parse_full_blob(text, repo_url)
     end
     ref = new_ref
   end
-  -- No such ref, first part might still be a commit ID.
+  -- No such ref, fallback to a commit ID (or part of it).
   if not ref then
-    -- If the first part looks like a valid commit ID, it probably is.
-    -- TODO: Find a better way. The repository is not cloned yet.
-    if parts[1]:find("^" .. ("%x"):rep(40) .. "$") then
-      return {
-        commit = parts[1],
-        selected_path = vim.fn.join(vim.list_slice(parts, 2), "/"),
-      }
-    else
-      return { full_blob = text }
-    end
+    return {
+      commit = parts[1],
+      selected_path = vim.fn.join(vim.list_slice(parts, 2), "/"),
+    }
   end
 
   -- Generate a selected path and trim leading slash.
@@ -114,9 +111,14 @@ end
 
 function Parser:parse_tree_or_full_blob(text, repo_url)
   local res = {}
-  local branch_or_tag = text:match "^/tree/(.*)$"
-  if branch_or_tag then
-    res.branch = branch_or_tag
+  local ref = text:match "^/tree/(.*)"
+  if ref and ref ~= "" then
+    local matches = self.gitcmd:list_refs_sync(repo_url, ref)
+    if #matches == 0 then
+      res.commit = ref
+    else
+      res.branch = ref
+    end
   else
     -- Full blob contains both the ref and the file path.
     -- Due to possible branch / tag names that contain `/`, the parser cannot

tests/parser_spec.lua

@@ -11,6 +11,9 @@ T.run = function()
         repo_url = "https://github.com/moyiz/git-dev.nvim.git",
         type = "http",
       },
+      remote_refs = {
+        { commit_id = "1", ref = "refs/heads/main" },
+      },
     },
     {
       url = "https://github.com/moyiz/git-dev.nvim/tree/url-parsing",
@@ -19,6 +22,9 @@ T.run = function()
         repo_url = "https://github.com/moyiz/git-dev.nvim.git",
         type = "http",
       },
+      remote_refs = {
+        { commit_id = "1", ref = "refs/heads/url-parsing" },
+      },
     },
     {
       url = "https://github.com/moyiz/git-dev.nvim/blob",
@@ -62,6 +68,9 @@ T.run = function()
         branch = "stable",
         type = "http",
       },
+      remote_refs = {
+        { commit_id = "1", ref = "refs/heads/stable" },
+      },
     },
     {
       url = "https://github.com/echasnovski/mini.nvim/tree/v0.12.0",
@@ -70,30 +79,45 @@ T.run = function()
         branch = "v0.12.0",
         type = "http",
       },
+      remote_refs = {
+        { commit_id = "1", ref = "refs/tags/v0.12.0" },
+      },
     },
     {
       url = "https://github.com/echasnovski/mini.nvim/blob/main/README.md",
       expected = {
         repo_url = "https://github.com/echasnovski/mini.nvim.git",
-        full_blob = "main/README.md",
+        branch = "main",
+        selected_path = "README.md",
         type = "http",
       },
+      remote_refs = {
+        { id = "1", ref = "refs/heads/main" },
+      },
     },
     {
       url = "https://github.com/echasnovski/mini.nvim/blob/stable/README.md",
       expected = {
         repo_url = "https://github.com/echasnovski/mini.nvim.git",
-        full_blob = "stable/README.md",
+        branch = "stable",
+        selected_path = "README.md",
         type = "http",
       },
+      remote_refs = {
+        { id = "1", ref = "refs/heads/stable" },
+      },
     },
     {
       url = "https://github.com/echasnovski/mini.nvim/blob/main/lua/mini/basics.lua",
       expected = {
         repo_url = "https://github.com/echasnovski/mini.nvim.git",
-        full_blob = "main/lua/mini/basics.lua",
+        branch = "main",
+        selected_path = "lua/mini/basics.lua",
         type = "http",
       },
+      remote_refs = {
+        { id = "1", ref = "refs/heads/main" },
+      },
     },
     {
       url = "https://github.com/echasnovski/mini.nvim/blob/0cdf41afab4f13b7529d9965cff9c3d2d67b0bfb/lua/mini/statusline.lua",
@@ -112,12 +136,61 @@ T.run = function()
         branch = "features/reuse-only",
         type = "http",
       },
+      remote_refs = {
+        { commit_id = "1", ref = "refs/heads/features/reuse-only" },
+      },
     },
     {
       url = "https://github.com/spack/spack/blob/features/reuse-only/.github/dependabot.yml",
       expected = {
         repo_url = "https://github.com/spack/spack.git",
-        full_blob = "features/reuse-only/.github/dependabot.yml",
+        branch = "features/reuse-only",
+        selected_path = ".github/dependabot.yml",
+        type = "http",
+      },
+      remote_refs = {
+        { commit_id = "1", ref = "refs/heads/features/reuse-only" },
+      },
+    },
+    {
+      url = "https://github.com/moyiz/git-dev.nvim/tree/d5f9302226852b80f58ef13e8353fda0cd78184e",
+      expected = {
+        repo_url = "https://github.com/moyiz/git-dev.nvim.git",
+        commit = "d5f9302226852b80f58ef13e8353fda0cd78184e",
+        type = "http",
+      },
+    },
+    {
+      url = "https://github.com/moyiz/git-dev.nvim/tree/d5f9302226852b8",
+      expected = {
+        repo_url = "https://github.com/moyiz/git-dev.nvim.git",
+        commit = "d5f9302226852b8",
+        type = "http",
+      },
+    },
+    {
+      url = "https://github.com/moyiz/git-dev.nvim/tree/d5f93022",
+      expected = {
+        repo_url = "https://github.com/moyiz/git-dev.nvim.git",
+        commit = "d5f93022",
+        type = "http",
+      },
+    },
+    {
+      url = "https://github.com/moyiz/git-dev.nvim/blob/d5f9302226852b80f58ef13e8353fda0cd78184e/.gitignore",
+      expected = {
+        repo_url = "https://github.com/moyiz/git-dev.nvim.git",
+        commit = "d5f9302226852b80f58ef13e8353fda0cd78184e",
+        selected_path = ".gitignore",
+        type = "http",
+      },
+    },
+    {
+      url = "https://github.com/moyiz/git-dev.nvim/blob/d5f9302226852/.gitignore",
+      expected = {
+        repo_url = "https://github.com/moyiz/git-dev.nvim.git",
+        commit = "d5f9302226852",
+        selected_path = ".gitignore",
         type = "http",
       },
     },
@@ -160,6 +233,9 @@ T.run = function()
         branch = "1.1.0",
         type = "http",
       },
+      remote_refs = {
+        { commit_id = "1", ref = "refs/tags/1.1.0" },
+      },
     },
     {
       url = "https://gitlab.com/gitlab-org/code-creation/repository-x-ray/-/blob/010e6076637568b291b7563e089175130ce72369/cmd/scan/main.go",