(maxima.info)Functions and Variables for Debugging


Prev: Keyword Commands Up: Debugging
Enter node , (file) or (file)node

38.3 Functions and Variables for Debugging
==========================================

 -- Option variable: debugmode
     Default value: 'false'

     When a Maxima error occurs, Maxima will start the debugger if
     'debugmode' is 'true'.  The user may enter commands to examine the
     call stack, set breakpoints, step through Maxima code, and so on.
     See 'debugging' for a list of debugger commands.

     Enabling 'debugmode' will not catch Lisp errors.

 -- Option variable: refcheck
     Default value: 'false'

     When 'refcheck' is 'true', Maxima prints a message each time a
     bound variable is used for the first time in a computation.

 -- Option variable: setcheck
     Default value: 'false'

     If 'setcheck' is set to a list of variables (which can be
     subscripted), Maxima prints a message whenever the variables, or
     subscripted occurrences of them, are bound with the ordinary
     assignment operator ':', the '::' assignment operator, or function
     argument binding, but not the function assignment ':=' nor the
     macro assignment '::=' operators.  The message comprises the name
     of the variable and the value it is bound to.

     'setcheck' may be set to 'all' or 'true' thereby including all
     variables.

     Each new assignment of 'setcheck' establishes a new list of
     variables to check, and any variables previously assigned to
     'setcheck' are forgotten.

     The names assigned to 'setcheck' must be quoted if they would
     otherwise evaluate to something other than themselves.  For
     example, if 'x', 'y', and 'z' are already bound, then enter

          setcheck: ['x, 'y, 'z]$

     to put them on the list of variables to check.

     No printout is generated when a variable on the 'setcheck' list is
     assigned to itself, e.g., 'X: 'X'.

 -- Option variable: setcheckbreak
     Default value: 'false'

     When 'setcheckbreak' is 'true', Maxima will present a break prompt
     whenever a variable on the 'setcheck' list is assigned a new value.
     The break occurs before the assignment is carried out.  At this
     point, 'setval' holds the value to which the variable is about to
     be assigned.  Hence, one may assign a different value by assigning
     to 'setval'.

     See also 'setcheck' and 'setval'.

 -- System variable: setval

     Holds the value to which a variable is about to be set when a
     'setcheckbreak' occurs.  Hence, one may assign a different value by
     assigning to 'setval'.

     See also 'setcheck' and 'setcheckbreak'.

 -- Function: timer (<f_1>, ..., <f_n>)
          timer (all)
          timer ()

     Given functions <f_1>, ..., <f_n>, 'timer' puts each one on the
     list of functions for which timing statistics are collected.
     'timer(f)$ timer(g)$' puts 'f' and then 'g' onto the list; the list
     accumulates from one call to the next.

     'timer(all)' puts all user-defined functions (as named by the
     global variable 'functions') on the list of timed functions.

     With no arguments, 'timer' returns the list of timed functions.

     Maxima records how much time is spent executing each function on
     the list of timed functions.  'timer_info' returns the timing
     statistics, including the average time elapsed per function call,
     the number of calls, and the total time elapsed.  'untimer' removes
     functions from the list of timed functions.

     'timer' quotes its arguments.  'f(x) := x^2$ g:f$ timer(g)$' does
     not put 'f' on the timer list.

     If 'trace(f)' is in effect, then 'timer(f)' has no effect; 'trace'
     and 'timer' cannot both be in effect at the same time.

     See also 'timer_devalue'.

 -- Function: untimer (<f_1>, ..., <f_n>)
          untimer ()

     Given functions <f_1>, ..., <f_n>, 'untimer' removes each function
     from the timer list.

     With no arguments, 'untimer' removes all functions currently on the
     timer list.

     After 'untimer (f)' is executed, 'timer_info (f)' still returns
     previously collected timing statistics, although 'timer_info()'
     (with no arguments) does not return information about any function
     not currently on the timer list.  'timer (f)' resets all timing
     statistics to zero and puts 'f' on the timer list again.

 -- Option variable: timer_devalue
     Default value: 'false'

     When 'timer_devalue' is 'true', Maxima subtracts from each timed
     function the time spent in other timed functions.  Otherwise, the
     time reported for each function includes the time spent in other
     functions.  Note that time spent in untimed functions is not
     subtracted from the total time.

     See also 'timer' and 'timer_info'.

 -- Function: timer_info (<f_1>, ..., <f_n>)
          timer_info ()

     Given functions <f_1>, ..., <f_n>, 'timer_info' returns a matrix
     containing timing information for each function.  With no
     arguments, 'timer_info' returns timing information for all
     functions currently on the timer list.

     The matrix returned by 'timer_info' contains the function name,
     time per function call, number of function calls, total time, and
     'gctime', which meant "garbage collection time" in the original
     Macsyma but is now always zero.

     The data from which 'timer_info' constructs its return value can
     also be obtained by the 'get' function:

          get(f, 'calls);  get(f, 'runtime);  get(f, 'gctime);

     See also 'timer'.

 -- Function: trace (<f_1>, ..., <f_n>)
          trace (all)
          trace ()

     Given functions <f_1>, ..., <f_n>, 'trace' instructs Maxima to
     print out debugging information whenever those functions are
     called.  'trace(f)$ trace(g)$' puts 'f' and then 'g' onto the list
     of functions to be traced; the list accumulates from one call to
     the next.

     'trace(all)' puts all user-defined functions (as named by the
     global variable 'functions') on the list of functions to be traced.

     With no arguments, 'trace' returns a list of all the functions
     currently being traced.

     The 'untrace' function disables tracing.  See also 'trace_options'.

     'trace' quotes its arguments.  Thus, 'f(x) := x^2$ g:f$ trace(g)$'
     does not put 'f' on the trace list.

     When a function is redefined, it is removed from the timer list.
     Thus after 'timer(f)$ f(x) := x^2$', function 'f' is no longer on
     the timer list.

     If 'timer (f)' is in effect, then 'trace (f)' has no effect;
     'trace' and 'timer' can't both be in effect for the same function.

 -- Function: trace_options (<f>, <option_1>, ..., <option_n>)
          trace_options (<f>)

     Sets the trace options for function <f>.  Any previous options are
     superseded.  'trace_options (<f>, ...)' has no effect unless 'trace
     (<f>)' is also called (either before or after 'trace_options').

     'trace_options (<f>)' resets all options to their default values.

     The option keywords are:

        * 'noprint' Do not print a message at function entry and exit.
        * 'break' Put a breakpoint before the function is entered, and
          after the function is exited.  See 'break'.
        * 'lisp_print' Display arguments and return values as Lisp
          objects.
        * 'info' Print '-> true' at function entry and exit.
        * 'errorcatch' Catch errors, giving the option to signal an
          error, retry the function call, or specify a return value.

     Trace options are specified in two forms.  The presence of the
     option keyword alone puts the option into effect unconditionally.
     (Note that option <foo> is not put into effect by specifying
     '<foo>: true' or a similar form; note also that keywords need not
     be quoted.)  Specifying the option keyword with a predicate
     function makes the option conditional on the predicate.

     The argument list to the predicate function is always '[level,
     direction, function, item]' where 'level' is the recursion level
     for the function, 'direction' is either 'enter' or 'exit',
     'function' is the name of the function, and 'item' is the argument
     list (on entering) or the return value (on exiting).

     Here is an example of unconditional trace options:

          (%i1) ff(n) := if equal(n, 0) then 1 else n * ff(n - 1)$

          (%i2) trace (ff)$

          (%i3) trace_options (ff, lisp_print, break)$

          (%i4) ff(3);

     Here is the same function, with the 'break' option conditional on a
     predicate:

          (%i5) trace_options (ff, break(pp))$

          (%i6) pp (level, direction, function, item) := block (print (item),
              return (function = 'ff and level = 3 and direction = exit))$

          (%i7) ff(6);

 -- Function: untrace
          untrace (<f_1>, ..., <f_n>)
          untrace ()

     Given functions <f_1>, ..., <f_n>, 'untrace' disables tracing
     enabled by the 'trace' function.  With no arguments, 'untrace'
     disables tracing for all functions.

     'untrace' returns a list of the functions for which it disabled
     tracing.


automatically generated by info2www version 1.2.2.9