diff --git a/yiiguxing.translation.xml b/yiiguxing.translation.xml
index b3c5f56..a4afef1 100644
--- a/yiiguxing.translation.xml
+++ b/yiiguxing.translation.xml
@@ -4,6 +4,8 @@
   </component>
   <component name="Translation.States">
     <histories>
+      <item value="subtract" />
+      <item value="Extend this type to create a [Generator] that invokes [generateForAnnotatedElement] for every top level element in the source file annotated with [T]. When all annotated elements have been processed, the results will be combined into a single output with duplicate items collapsed. For example, this will allow code generated for all top level elements which are annotated with `@Deprecated`: ```dart class DeprecatedGenerator extends GeneratorForAnnotation&lt;Deprecated&gt; { @override Future&lt;String&gt; generateForAnnotatedElement( Element element, ConstantReader annotation, BuildStep buildStep) async { Return a string representing the code to emit. } } ``` Elements which are not at the top level, such as the members of a class or extension, are not searched for annotations. To operate on, for instance, annotated fields of a class ensure that the class itself is annotated with [T] and use the [Element] to iterate over fields. The [TypeChecker] utility may be helpful to check which elements have a given annotation." />
       <item value="Computes the set of features for the latest language version known to the analyzer, without any experiments. Use it only if you really don't care which language version you want to use, and sure that the code that you process is valid for the latest language version. Otherwise, it is recommended to use [FeatureSet.fromEnableFlags2]." />
       <item value="Return the result of parsing the file at the given [path]. If a [resourceProvider] is given, it will be used to access the file system. [featureSet] determines what set of features will be assumed by the parser. This parameter is required because the analyzer does not yet have a performant way of computing the correct feature set for a single file to be parsed. Callers that need the feature set to be strictly correct must create an [AnalysisContextCollection], query it to get an [AnalysisContext], query it to get an [AnalysisSession], and then call `getParsedUnit`. Callers that don't need the feature set to be strictly correct can pass in `FeatureSet.latestLanguageVersion()` to enable the default set of features; this is much more performant than using an analysis session, because it doesn't require the analyzer to process the SDK. If [throwIfDiagnostics] is `true` (the default), then if any diagnostics are produced because of syntactic errors in the file an `ArgumentError` will be thrown. If the parameter is `false`, then the caller can check the result to see whether there are any errors." />
       <item value="Return the result of resolving the file at the given [path]. If a [resourceProvider] is given, it will be used to access the file system. Note that if more than one file is going to be resolved then this function is inefficient. Clients should instead use [AnalysisContextCollection] to create one or more contexts and use those contexts to resolve the files." />
@@ -52,8 +54,6 @@
       <item value="they" />
       <item value="Recomand Friend" />
       <item value="fix: wrong onTap ussage." />
-      <item value="is Auto Generated" />
-      <item value="The child can be at most as large as the available space (but is allowed to be smaller)." />
     </histories>
     <option name="languageScores">
       <map>