Perf: Improve performance of multithreaded --check and undefined-field diagnostic

changelog.md

@@ -5,6 +5,7 @@
 * `NEW` Add postfix snippet for `unpack`
 * `FIX` `diagnostics.severity` defaulting to "Warning" when run using `--check` [#2730](https://github.com/LuaLS/lua-language-server/issues/2730)
 * `NEW` Add support for lambda style functions, `|paramList| expr` is syntactic sugar for `function(paramList) return expr end` 
+* `CHG` Improve performance of multithreaded `--check` and `undefined-field` diagnostic
 
 ## 3.9.3
 `2024-6-11`

script/cli/check_worker.lua

@@ -43,15 +43,6 @@ local checkLevel = define.DiagnosticSeverity[CHECKLEVEL] or define.DiagnosticSev
 
 util.enableCloseFunction()
 
--- Hash function used to distribute work.
-local function hashString(str)
-    local hash = 0
-    for i = 1, #str do
-        hash = (hash * 37 & 0xFFFFFFFF) + str:byte(i, i)
-    end
-    return hash
-end
-
 local lastClock = os.clock()
 local results = {}
 
@@ -109,9 +100,9 @@ xpcall(lclient.start, errorhandler, lclient, function (client)
 
     local uris = files.getChildFiles(rootUri)
     local max  = #uris
+    table.sort(uris)    -- sort file list to ensure the work distribution order across multiple threads
     for i, uri in ipairs(uris) do
-        local hash = hashString(uri) % numThreads + 1
-        if hash == threadId then
+        if (i % numThreads + 1) == threadId then
             files.open(uri)
             diag.doDiagnostic(uri, true)
             -- Print regularly but always print the last entry to ensure that logs written to files don't look incomplete.

script/core/diagnostics/undefined-field.lua

@@ -21,7 +21,7 @@ return function (uri, callback)
     local function checkUndefinedField(src)
         await.delay()
 
-        if #vm.getDefs(src) > 0 then
+        if vm.hasDef(src) then
             return
         end
         local node = src.node

script/vm/def.lua

@@ -92,3 +92,42 @@ function vm.getDefs(source)
 
     return results
 end
+
+local HAS_DEF_ERR = {}  -- the error object for comparing
+local function checkHasDef(checkFunc, source, pushResult)
+    local _, err = pcall(checkFunc, source, pushResult)
+    return err == HAS_DEF_ERR
+end
+
+---@param source parser.object
+function vm.hasDef(source)
+    local mark = {}
+    local hasLocal
+    local function pushResult(src)
+        if src.type == 'local' then
+            if hasLocal then
+                return
+            end
+            hasLocal = true
+            if  source.type ~= 'local'
+            and source.type ~= 'getlocal'
+            and source.type ~= 'setlocal'
+            and source.type ~= 'doc.cast.name' then
+                return
+            end
+        end
+        if not mark[src] then
+            mark[src] = true
+            if guide.isAssign(src)
+            or guide.isLiteral(src) then
+                -- break out on 1st result using error() with a unique error object
+                error(HAS_DEF_ERR)
+            end
+        end
+    end
+
+    return checkHasDef(searchBySimple, source, pushResult)
+        or checkHasDef(searchByLocalID, source, pushResult)
+        or checkHasDef(vm.compileByNodeChain, source, pushResult)
+        or checkHasDef(searchByNode, source, pushResult)
+end