Revise (really, Julia itself) can handle many kinds of code changes, but a few may require special treatment:
Sometimes you might wish to change a method's type signature or number of arguments, or remove a method specialized for specific types. To prevent "stale" methods from being called by dispatch, Revise automatically accommodates method deletion, for example:
f(x) = 1 f(x::Int) = 2 # delete this method
If you save the file, the next time you call
f(5) from the REPL you will get 1, and
methods(f) will show a single method. Revise even handles more complex situations, such as functions with default arguments: the definition
defaultargs(x, y=0, z=1.0f0) = x + y + z
generates 3 different methods (with one, two, and three arguments respectively), and editing this definition to
defaultargs(x, yz=(0,1.0f0)) = x + yz + yz
requires that we delete all 3 of the original methods and replace them with two new methods.
However, to find the right method(s) to delete, Revise needs to be able to parse source code to extract the signature of the to-be-deleted method(s). Unfortunately, a few valid constructs are quite difficult to parse properly. For example, methods generated with code:
for T in (Int, Float64, String) # edit this line to `for T in (Int, Float64)` @eval mytypeof(x::$T) = $T end
will not disappear from the method lists until you restart.
To delete a method manually, you can use
m = @which foo(args...) to obtain a method, and then call
Macros and generated functions
If you change a macro definition or methods that get called by
@generated functions outside their
quote block, these changes will not be propagated to functions that have already evaluated the macro or generated function.
You may explicitly call
revise(MyModule) to force reevaluating every definition in module
MyModule. Note that when a macro changes, you have to revise all of the modules that use it.
Distributed computing (multiple workers)
Revise supports changes to code in worker processes. The code must be loaded in the main process in which Revise is running, and you must use
@everywhere using Revise.
Revise cannot handle changes in anonymous functions used in
remotecalls. Consider the following module definition:
module ParReviseExample using Distributed greet(x) = println("Hello, ", x) foo() = for p in workers() remotecall_fetch(() -> greet("Bar"), p) end end # module
Changing the remotecall to
remotecall_fetch((x) -> greet("Bar"), p, 1) will fail, because the new anonymous function is not defined on all workers. The workaround is to write the code to use named functions, e.g.,
module ParReviseExample using Distributed greet(x) = println("Hello, ", x) greetcaller() = greet("Bar") foo() = for p in workers() remotecall_fetch(greetcaller, p) end end # module
and the corresponding edit to the code would be to modify it to
greetcaller(x) = greet("Bar") and
remotecall_fetch(greetcaller, p, 1).
Changes that Revise cannot handle
Finally, there are some kinds of changes that Revise cannot incorporate into a running Julia session:
- changes to type definitions
- file or module renames
- conflicts between variables and functions sharing the same name
These kinds of changes require that you restart your Julia session.