Гапон А.К. 381806-3 #47
Conversation
|
The program does not compile. Please fix errors. |
|
@eshulankina поправил |
| Result.Context->getLangOpts()); | ||
| auto res = ("static_cast<" + tSource + ">(").str(); | ||
| rwr.ReplaceText(loc, res); | ||
| auto closePL = Lexer::getLocForEndOfToken( |
There was a problem hiding this comment.
Please describe why you are using Lexer::getLocForEndOfToken(expr->getSubExprAsWritten()->IgnoreImpCasts()->getEndLoc(),...) here.
What would be the difference if expr->getSubExprAsWritten()->IgnoreImpCasts()->getEndLoc().getLocOffset(1) was used as closePL?
There was a problem hiding this comment.
Lexer::getLocForEndOfToken определяет локацию за последним элементом токена
There was a problem hiding this comment.
What is the difference in the use with expr->getSubExprAsWritten()->IgnoreImpCasts()->getEndLoc().getLocOffset(1)?
There was a problem hiding this comment.
Ну эта функция возвращает изначальное местоположение, сразу за концом токена, на который ссылается переданное местоположение expr-> getSubExprAsWritten () -> IgnoreImpCasts () -> getEndLoc ()
There was a problem hiding this comment.
What happens if we use expr->getSubExprAsWritten()->IgnoreImpCasts()->getEndLoc().getLocOffset(1) instead of Lexer::getLocForEndOfToken()?
When does the approach with expr->getSubExprAsWritten()->IgnoreImpCasts()->getEndLoc().getLocOffset(1) not work?
There was a problem hiding this comment.
SubExpr->getEndLoc().getLocWithOffset(1) захватил бы только первый символ имени, в отличие от реализованного варианта.
| auto res = ("static_cast<" + tSource + ">(").str(); | ||
| rwr.ReplaceText(loc, res); | ||
| auto closePL = Lexer::getLocForEndOfToken( | ||
| expr->getSubExprAsWritten()->IgnoreImpCasts()->getEndLoc(), 0, |
There was a problem hiding this comment.
Please describe what getSubExprAsWritten() does.
What is the difference with getSubExpr() method?
There was a problem hiding this comment.
Получает подвыражение приведения как написано в исходном коде, просматривая неявные приведения или другие промежуточные узлы, которые были добавлены семантическим анализом.
getSubExpr возвращает итоговый результат, с учетом неявных приведений и промежуточных узлов
There was a problem hiding this comment.
If you try to print the AST of the test.cpp example, you will see that you have several ImplicitCastExpr nodes:
Depending on which approaches you use to get the subexpression node, you will get different result nodes for the test.cpp example:
Code:
// getSubExpr() retrieves the first cast subexpression
llvm::outs() << "getSubExpr(): " << castNode->getSubExpr()->getStmtClassName();
// getSubExprAsWritten() retrieves the cast subexpression as it was written in the source
// code, looking through any implicit casts or other intermediate nodes
// introduced by semantic analysis.
llvm::outs() << "getSubExprAsWritten(): " << castNode->getSubExprAsWritten()->getStmtClassName();
// getSubExpr() retrieves the first cast subexpression.
// ignoreImpCasts() skips past any implicit casts which might surround this expression until
// reaching a fixed point. It skips the following AST nodes: ImplicitCastExpr, FullExpr.
llvm::outs() << "getSubExpr()->ignoreImpCasts(): " << castNode->getSubExpr()->IgnoreImpCasts()->getStmtClassName();Output:
getSubExpr(): ImplicitCastExpr
getSubExprAsWritten(): DeclRefExpr
getSubExpr()->ignoreImpCasts(): DeclRefExpr
No description provided.