On Jul 3, 4:21 am, yhe <[email protected]> wrote: > In 'OpenFilesCallback': > * if a directory file is selected then all of the files in the > directory(s) will become > candidates for selection after being 'filtered' for passing to the > 'openFiles' method; > * if 'singleFile' attribute is true as set in 'OpenFileOptions' > class the the first > 'filtered' file in the directory is passed; > > In 'File': > * needs a 'URL' string attribute;
There's a tricky case with recursive directory traversal. Suppose that my workplace puts everybody's home directory under /home on a network file system. If a recursive filepicker API is implemented synchronously, and I accidentally select the "/home" dir, then you might lock up the whole browser for literally hours as it traverses thousands of peoples' home dirs. As for a File's URL attribute, we've tried to avoid exposing a user's username (e.g. the "joe" in "/home/joe/Photos/foo.jpeg") if we can help it. If you really need a URL for a File's contents, there's a workaround where you can capture the File's Blob to a ResourceStore, which gives you a URL. Just make sure to remove the capture, when you're done.
