RE: goto considered essential
Wasn't there an exit thingabobbers I saw mentioned somewhere? I thought
it would be neat to use, because it doesn't exist in QBasic, and I had
to setup escape sequences that would abort my loops by resetting
variables and stuff. Really messy and unclear. Exit looked like a
great alternative.
Love & Friendship & Blessed Be!
Lynn Erika Kilroy
Vincent wrote:
>
>
> posted by: Vincent <darkvincentdude at yahoo.com>
>
> Tom Dailey wrote:
> >
> >
> > Consider the following program fragment, which copies one text file into
> > another text file:
> >
> > sequence cmd_args
> > sequence cmd_line_words
> > sequence input_file_name
> > integer input_file_nr
> > object line
> > sequence output_file_name
> > integer output_file_nr
> >
> > cmd_line_words = command_line ( )
> > if length ( cmd_line_words ) != 4 then
> > Printf ( "Exactly two arguments expected -- %d supplied.\n",
> > { length ( cmd_line_words ) - 2 } )
> > else
> > cmd_args = cmd_line_words [ 3..4 ]
> > input_file_name = cmd_args [ 1 ]
> > input_file_nr = open ( input_file_name, "r" )
> > if input_file_nr = -1 then
> > Printf ( "Can't open input file %s.\n",
> > { input_file_name } )
> > else
> > output_file_name = cmd_args [ 2 ]
> > output_file_nr = open ( output_file_name, "w" )
> > if output_file_nr = -1 then
> > Printf ( "Can't open output file %s.\n",
> > { output_file_name } )
> > else
> > while 1 do
> > line = gets ( input_file_nr )
> > if atom ( line ) then
> > exit
> > end if
> > puts ( output_file_nr, line )
> > end while
> > Printf ( "%s copied to %s.\n",
> > { input_file_name, output_file_name } )
> > end if
> > end if
> > end if
> >
> > Without a goto statement, or something like it, we have the creeping
> > margin
> > problem. In production code, with many error checks in the main logic,
> > this
> > can make code very hard to understand. With a goto, we have
> >
> > cmd_line_words = command_line ( )
> > if length ( cmd_line_words ) != 4 then
> > goto arg_count_is_wrong
> > end if
> > cmd_args = cmd_line_words [ 3..4 ]
> > input_file_name = cmd_args [ 1 ]
> > input_file_nr = open ( input_file_name, "r" )
> > if input_file_nr = -1 then
> > goto cant_open_input_file
> > end if
> > output_file_name = cmd_args [ 2 ]
> > output_file_nr = open ( output_file_name, "w" )
> > if output_file_nr = -1 then
> > goto cant_open_output_file
> > end if
> > while 1 do
> > line = gets ( input_file_nr )
> > if atom ( line ) then
> > exit
> > end if
> > puts ( output_file_nr, line )
> > end while
> > Printf ( "%s copied to %s.\n",
> > { input_file_name, output_file_name } )
> > goto halt
> >
> > arg_count_is_wrong:
> > Printf ( "Exactly two arguments expected -- %d supplied.\n",
> > { length ( cmd_line_words ) - 2 } )
> > goto halt
> > cant_open_input_file:
> > Printf ( "Can't open input file %s.\n", { input_file_name } )
> > goto halt
> > cant_open_output_file:
> > Printf ( "Can't open output file %s.\n", { output_file_name } )
> > goto halt
> > halt:
> >
> > Here, the main (non-error) logic is much more apparent. But what, you
> > ask,
> > is then to prevent evil programmers from generating spaghetti code?
> > Well,
> > nothing. So perhaps a more constrained construct would be better. In
> > the
> > spirit of Ada exceptions, we might try something like this:
> >
> > --
> > -- Function/procedure body or main program starts here.
> > --
> > cmd_line_words = command_line ( )
> > exception arg_count_is_wrong if length ( cmd_line_words ) != 4
> > cmd_args = cmd_line_words [ 3..4 ]
> > input_file_name = cmd_args [ 1 ]
> > input_file_nr = open ( input_file_name, "r" )
> > exception cant_open_input_file if input_file_nr = -1
> > output_file_name = cmd_args [ 2 ]
> > output_file_nr = open ( output_file_name, "w" )
> > exception cant_open_output_file if output_file_nr = -1
> > while 1 do
> > line = gets ( input_file_nr )
> > exit if atom ( line )
> > puts ( output_file_nr, line )
> > end while
> > Printf ( "%s has been copied to %s.\n",
> > { input_file_name, output_file_name } )
<snip>
|
Not Categorized, Please Help
|
|
