Re: gdb/196: completer bugs

[Fernando Nasser <fnasser at cygnus dot com>]

The following reply was made to PR gdb/196; it has been noted by GNATS.

From: Fernando Nasser <fnasser@cygnus.com>
To: keiths@cygnus.com
Cc: gdb-gnats@sources.redhat.com, Eli Zaretskii <eliz@is.elta.co.il>
Subject: Re: gdb/196: completer bugs
Date: Thu, 16 Aug 2001 17:43:25 -0400

 keiths@cygnus.com wrote:
 > 
 > >Description:
 > Symbol completion appears to be slightly broken. It often
 > includes things like type names in the symbol list.
 > >How-To-Repeat:
 > 1) Load gdb into gdb
 > 2) At prompt type "break comple<TAB><TAB>" and look at
 >    what comes out:
 > (gdb) break comple
 > complete.c                       completedieinfo
 > complete_command                 completer.c
 > complete_execution               completer.h
 > complete_object_allocating_ctor  completion_changed_buffer
 > complete_object_ctor             completion_list_add_name
 > complete_object_dtor             completion_matches
 > complete_on_cmdlist              complex double
 > complete_on_enum                 complex float
 > complete_symtab                  complex int
 > completed.1                      complex long double
 > (gdb) break comple
 > 
 > "complex long double"? completed.1? complete.c?
 
 That is the way it is (what doesn't mean it can't be improved).
 It doesn't know what you want to type in there so it presents you
 with a list of symbols that start with the substring you've typed.
 
 Commands can specify a special completer, like the filename_completer()
 that we use for commands that expect a filename.
 
 Maybe we could have a linespec_completer() and some other more
 specific completers so the list gets shorter.  But probably a 
 completer for linespecs will be as difficult as the parser for it is ;-)
 
 Regards,
 
 -- 
 Fernando Nasser
 Red Hat - Toronto                       E-Mail:  fnasser@redhat.com
 2323 Yonge Street, Suite #300
 Toronto, Ontario   M4P 2C9