This method is the central method of the text_item's completion mechanism. This completion mechanism both deals selecting in hierarchical organised data (like the Unix file-system) and with simple lists of possible values. The second case is generally handled by text_item<->value_set. For selecting from hierarchical organised data the following methods need to be redefined:

• text_item<-split_completion Split the currently typed value in a directory part and a file part.
• text_item<-completions Return a set of possible files in the indicated directory using the result of text_item<-split_completion.
• text_item->indicate_directory Indicate the match of a directory.

  In this documentation we refer to files and directories, but in general this may refer to any organised data using any syntactical convention. The Prolog library file file_item.pl illustrated how the methods described above may be redefined to select a file from the Unix file-system.

  When text_item->complete is invoked, it attempts to complete the current entry. To do so, it locates the first choice-point and the set of objects that may be selected after this choice-point using chain<-complete_name. If the current point is a choice-point it will pop-up the browser object @completer with all possible completions. If the current point is a unique description of a completed value it will indicate such using the error sole_completion. Otherwise it will replace text_item<-displayed_value with the text leading to the first choice-point or completion.

  See also

  - text_item->indicate_directory
  - @completer
  - text_item-value_set