In e.g. QTBUG-48206 using the filename part of QWebEngineDownloadItem::path() is the proposed way of getting a filename of a download.

However, when I want to store a download named foo in a custom directory, and a ~/Downloads/foo happens to exist, I'd get foo(1) as proposed name.

From an API perspective, I can see two ways of solving this, and IMHO both should exist:

• Being able to pass a directory only to setPath, and let Chromium select a filename based on that.
• Being able to get the filename part without any collision logic (i.e. only the part inferred from content-disposition/URL/etc.)