Dynamic Code Loading and Execution

These functions either execute code contained in an arbitrary, possibly generated-at-runtime string or load code from a file potentially chosen at runtime.




execute("statement", objref)

parse and execute the command in the context of the object. If second arg not present then execute it at the top level of the interpreter. If command begins with a '~' then the tilda is removed and the rest of the command is executed without enclosing it in {}. This allows one to create a func or proc dynamically.

See also



err = execute1("statement")

err = execute1("statement", objref)

err = execute1("statement", show_err_mes)

err = execute1("statement", objref, show_err_mes


Same as execute() but returns 0 if there was an interpreter error during execution of the statement and returns 1 if successful. Does not surround the command with {}.

If the show_err_mes arg is present and equal to 0 then the normal interpreter error message printing is turned off for the scope of the statement.

Error messages can be turned on even inside the statement with show_errmess_always().

Parse and execute the command in the context of the object. If second arg not present then execute it at the top level of the interpreter.


Execute1 is heavily used in the construction of the fitter widgets. It is also useful to objects in gaining information about the outside with the idiom

sprint(cmd, "%s.var = outside_var", this)

Here, outside_var is unavailable from within the object and so a command is constructed which can be executed at the top level where that variable is available and sets the public var in the object.



load_proc("name1", ...)

load_func("name1", ...)

load_template("name1", ..._


Load the file containing a definition.

For each name in the list all the *.oc and *.hoc files will be searched and the first file that contains the appropriate proc, func, or begintemplate will be loaded. Loading only takes place if the name has not previously been defined. The search path consists of the current working directory, followed by the paths in the environment variable HOC_LIBRARY_PATH (space separated), followed by $NEURONHOME/lib/hoc. Remember that only entire files are loaded-- not just the definition of the name. And nothing is loaded if the name is already defined. Inadvertent recursion will use up all the file descriptors. For efficiency, on the first load, all the names are cached in a temporary file and the file is scanned on subsequent loads for that session.


This command is very slow under mswindows. Therefore it is often useful to explicitly load the standard run library with the statement:




load_file("filename", "name")

load_file(0or1, "filename")


Similar to load_proc() but loads files and so does not have the search overhead. Suitable for loading packages of files.

The functionality is identical to xopen() except that the xopen takes place only if if a file of that name has not already been loaded with the load_file, load_proc(), load_template(), or load_func() functions. The file is searched for in the current working directory, $HOC_LIBRARY_PATH (a colon or space separated list of directories), and $NEURONHOME/lib/hoc directories (in that order) for the file if there is no directory prefix. Before doing the xopen on the file the current working directory is temporarily changed to the directory containing the file so that it can xopen files relative to its location.

If the second string arg exists, the file is xopen'ed only if the name is not defined as a variable AND the file has not been loaded with load_file. This is useful in those cases where the package was first xopen'ed without going through the load_file function.

If the first arg is a number and is 1, then the file is loaded again even if it has already been loaded.



see load_proc()


see load_proc()