Introduction to Revise
Revise.jl
may help you keep your Julia sessions running longer, reducing the need to restart when you make changes to code. With Revise, you can be in the middle of a session and then edit source code, update packages, switch git branches, and/or stash/unstash code; typically, the changes will be incorporated into the very next command you issue from the REPL. This can save you the overhead of restarting, loading packages, and waiting for code to JIT-compile.
Using Revise also improves your experience when using the debuggers. Revise will keep track of changed locations of your methods in file, and ensure that the debugger displays the source code of what you're actually debugging.
Many users automatically load Revise on startup. On versions of Julia older than 1.5, this is slightly more involved than just adding using Revise
to .julia/config/startup.jl
: see Using Revise by default for details.
Installation
You can obtain Revise using Julia's Pkg REPL-mode (hitting ]
as the first character of the command prompt):
(v1.0) pkg> add Revise
or with using Pkg; Pkg.add("Revise")
.
Usage example
We'll make changes to Julia's "Example" package (a trivial package designed to illustrate the file and directory organization of typical packages). We have to "develop" it in order to make changes:
(v1.0) pkg> dev Example
[...output related to installation...]
Now we load Revise (if we haven't already done so) and Example:
julia> using Revise # importantly, this must come before `using Example`
julia> using Example
julia> hello("world")
"Hello, world"
Now we're going to check that the Example
module currently lacks a function named f
:
julia> Example.f()
ERROR: UndefVarError: f not defined
But say we really want f
, so let's add it. You can either navigate to the source code (at .julia/dev/Example/src/Example.jl
) in an editor manually, or you can use Julia to open it for you:
julia> edit(hello) # opens Example.jl in the editor you have configured
Now, add a function f() = π
and save the file. Go back to the REPL (the same REPL, don't restart Julia) and try this:
julia> Example.f()
π = 3.1415926535897...
Voila! Even though we'd loaded Example before adding this function, Revise noticed the change and inserted it into our running session.
Revise's first revision has latency of several seconds–it's compiling all of its internal code, which includes a complete Julia interpreter and all of Revise's parse/diff/patch/cache machinery. After your first revision, future revisions will generally be fast enough that they will seem nearly instantaneous. (There are exceptions, but they occur only in specific circumstances, for example when Revise's own code gets invalidated by your changes.)
Now suppose we realize we've made a horrible mistake: that f
method will mess up everything, because it's part of a more complicated dispatch process and incorrectly intercepts certain f
calls. No problem, just delete f
in your editor, save the file, and you're back to this:
julia> Example.f()
ERROR: UndefVarError: f not defined
all without restarting Julia. While you can evaluate new methods without Revise using inline evaluation through your IDE, method deletion is just one example of a change that can only be made easily by Revise.
If you need more examples, see Revise usage: a cookbook.
Other key features of Revise
Revise updates its internal paths when you change versions of a package. To try this yourself, first re-insert that definition of f
in the dev
version of Example
and save the file. Now try toggling back and forth between the dev
and released versions of Example
:
(v1.0) pkg> free Example # switch to the released version of Example
julia> Example.f()
ERROR: UndefVarError: f not defined
(v1.0) pkg> dev Example
julia> Example.f()
π = 3.1415926535897...
Revise is not tied to any particular editor. (The EDITOR or JULIA_EDITOR environment variables can be used to specify your preference for which editor gets launched by Julia's edit
function.)
Some editors (like vim
) may require configuration to function properly.
If you don't want to have to remember to say using Revise
each time you start Julia, see Using Revise by default.
What Revise can track
Revise is fairly ambitious: if all is working, subject to a few Limitations you should be able to track changes to
- any package that you load with
import
orusing
- any script you load with
includet
(see Configuring the revise mode for important default restrictions onincludet
) - any file defining
Base
julia itself (withRevise.track(Base)
) - any of Julia's standard libraries (with, e.g.,
using Unicode; Revise.track(Unicode)
). Some stdlibs may require special handling; see, for example, a trick for modifying REPL. - any file defining
Core.Compiler
(withRevise.track(Core.Compiler)
)
The last one requires that you clone Julia and build it yourself from source.
Secrets of Revise "wizards"
Revise can assist with methodologies like test-driven development. While it's often desirable to write the test first, sometimes when fixing a bug it's very difficult to write a good test until you understand the bug better. Often that means basically fixing the bug before your write the test. With Revise, you can
- fix the bug while simultaneously developing a high-quality test
- verify that your test passes with the fixed code
- commit the test, but not the fix
git stash
your fix and check that your new test fails on the old code, thus verifying that your test captures the essence of the former bug (if it doesn't fail, you need a better test!)git stash pop
, test again, commit the fix, and submit
all without restarting your Julia session.
Other Revise workflows
Revise can be used to perform work when files update. For example, let's say you want to regenerate a set of web pages whenever your code changes. Suppose you've placed your Julia code in a package called MyWebCode
, and the pages depend on "file.js" and all files in the "assets/" directory; then
entr(["file.js", "assets"], [MyWebCode]) do
build_webpages(args...)
end
will execute build_webpages(args...)
whenever you save updates to the listed files or MyWebCode
.
If you want to regenerate the web page as soon as any change is detected, not only in MyWebCode
but also in any package tracked by Revise, you can provide the all
keyword argument to entr
:
entr(["file.js", "assets"]; all=true) do
build_webpages(args...)
end
Taking advantage of Revise in other packages
To make it easier for other packages to benefit from Revise without needing to add it as a dependency or understand Revise's internals, Revise interfaces with CodeTracking, which is a small package acting as Revise's "query" interface.
What else do I need to know?
Except in cases of problems (see below), that's it! Revise is a tool that runs in the background, and when all is well it should be essentially invisible, except that you don't have to restart Julia so often.
Revise can also be used as a "library" by developers who want to add other new capabilities to Julia; the sections How Revise works and Developer reference are particularly relevant for them.
If Revise doesn't work as expected
If Revise isn't working for you, here are some steps to try:
- See Configuration for information on customization options. In particular, some file systems (like NFS) and current users of WSL2 might require special options.
- Revise can't handle all kinds of code changes; for more information, see the section on Limitations.
- Try running
test Revise
from the Pkg REPL-mode. If tests pass, check the documentation to make sure you understand how Revise should work. If they fail (especially if it mirrors functionality that you need and isn't working), see Debugging problems with paths for one set of suggestions.
If you still encounter problems, please file an issue. Especially if you think Revise is making mistakes in adding or deleting methods, please see the page on Debugging Revise for information about how to attach logs to your bug report.