I was able to configure PyCharm to use the Python version provided by the Conda version installed when executing “using PythonPlot”.
I can use it to create plots, even though I need to explicitly call plt.show() in the Python console.
When using Julia, this worked with the plugin version 2026.4.0, but it no longer works with 2026.4.1 or 2026.4.2.
I get:
julia> using PythonPlot
julia> plot(rand(3))
Python: [<matplotlib.lines.Line2D object at 0x77a3e23a1a90>]
julia> PythonPlot.show()
<sys>:0: UserWarning: FigureCanvasAgg is non-interactive, and thus cannot be shown
So we are adding inbuilt code validation. Combined with unused variable detection, you get much better feedback while writing code.
I also added Revise.jl to auto-load on startup. This was a request by @xgdgsc and I think it makes complete sense. You can turn it off in the settings of course.
Other than that: a range of bugfixes as usual…
@xgdgsc: Huge thanks for all the detailed bug reports and feature requests - this release is basically your wishlist! @ufechner7: Thanks for catching the PythonPlot regression!
Plotting with PythonPlot works now!
Zoom and pan work, too!
Small issue: Clicking the “Fit to window” button never enlarges the figure beyond 100%. So it works only if the window is smaller than the figure.
You need to call figure() to create a new figure, but I think this is normal. Perhaps figure() should be called automatically when the “Clear” button is clicked.
Is there a way to have support for syntax highlighting for LaTeXStrings.jl?
It basically provides a string type L"…", where you don’t have to escape the LaTeX special characters ($,)
It usually takes JetBrains up to 48hrs to approve our version. In the meanwhile you can manually install it via Settings - Plugins - install from disk:
Big inspection cleanup in this release. The false positive warnings for do-blocks, tuple unpacking, string interpolation, and keyword arguments were getting annoying - all fixed. Also nested struct completion finally works now.
As for new features: Thanks to @taaem, we now also support LatexStrings: L"..."
And thanks to RGerzaguet we now also support terminal debugging via @run. This autoloads Debugger.jl - so you only need to remember @run. This is only a first implementation of it though, I will see that I integrate it further into the actual IDE debugging processes for one of the future releases (if possible). Oh and we fixed a bug that limited the size of plots to 10k. Now you can have much larger plots.
@xgdgsc: Thanks for the detailed bug reports on variable inspection! @ufechner7: Appreciate the PyPlot investigation! @RGerzaguet: Plot limits feature noted - and thanks for the @run request! @taaem: LaTeX strings are in - enjoy!
Full Debugger Integration: @run macro now provides complete IDE debugging experience with stack frames, variables panel, and step/continue buttons (#36)
Cell/Block Detection: “Send to REPL” now auto-detects and sends entire code blocks instead of single lines (#22)
Remote Variable Restore: Variables panel automatically refreshes after reconnecting to tmux/screen sessions (#27)
Fixes:
LanguageServer crash after initializing / Lambda hints NPE (#37)
False positive “undefined variable” for where clauses, selective imports, and module-level const (#34)
Struct dot-completion now shows both fields AND methods (ats. shows a,b,c plus filtA(ats::Ats)) (#9)
Parameter type inference now uses direct PSI accessors instead of fragile tree search (#9)
This was a bugfix release for the most part (with some major grammar improvement), but I also completed the full integration of repl debugging with the JetBrains debugger. So @run will now open a new debugsession inside JetBrains IDEs that is tied to the repl debug session. Variables will be mirrored in both the Debug panel and our Julia panel. Plots and Datatables will be shown inside our Julia window too. You can use F8,F9 etc to jump over lines and it will call n/c in the repl debug. Terminating one session automatically closes the other.
Apart from that, many bugfixes and ones that came back due to recent restructuring. Send to repl works now as expected and we get better autocomplete.
@RGerzaguet: Thank you for the debugger feedback and thorough testing of the @run macro integration! @taaem: Thanks for catching the LanguageServer crash! @xgdgsc: Continued appreciation for the detailed bug reports across multiple issues - the inspection fixes wouldn’t have happened without you! @Petr: I will look into how ways for us to include external folders next - thanks for reporting
Update: I published a quick bugfix release 2026.5.2, replacing 2026.5.1 (it fixes tmux errors and one with repl behavior and some minor remaining grammar issues)
Just a note: the “external folders” mention refers to the workspace concept in VS Code: there it is possible to add the entire folder of a package (or packages) to the workspace to enable symbol lookup. Very handy when LS gives trouble.
Breakpoints in @run Sessions: Full breakpoint support during REPL debugging - bp add "file":line, bp show, bp rm (#36)
Bidirectional Breakpoint Sync: Breakpoints now sync between IDE gutter and Julia REPL - set them wherever you prefer (#36)
Line-Based Breakpoint Removal: New bp remove "file":line command (in addition to Debugger.jl’s index-based approach)
Manifest.toml SDK Detection: Julia version now read from Manifest.toml for smarter SDK selection (#44)
Improved Completion Priority: LS/JETLS results now ranked higher in merged mode
Fixes:
REPL freeze after documentation hover due to EDT blocking (#43)
Plugin starting LSP in non-Julia projects (#41)
Parser error with end-(m-i) arithmetic in array indexing (#42)
Bug that caused some hover documentation to not show
Made big progress with synchronizing IDE and repl debug sessions. You can now set breakpoints directly from the REPL with bp add "file":line - and they sync bidirectionally with the IDE gutter. Set a breakpoint in the gutter, it shows up in your REPL session. Add one via bp add, it appears in the IDE etc etc
Also fixed an annoying freeze that happened when hovering for documentation - that one was blocking the EDT and locking up the REPL. And again some issues with our own on hover documentation and some minor grammar issues.
@RGerzaguet: Thanks for suggesting breakpoint support and the detailed feedback on the debugger workflow - I think we really got something working now
@xgdgsc: As always, thank you for the testing! @smartiegangs: Thanks for reporting the REPL crash with hover documentation - because of it, I also discovered some issues in our own on hover documentation files
Question:
What is the suggested way to execute a Julia script?
In the REPL I would just type
include("examples/my_example.jl")
or similar.
It would be nice if there were a run button that does this, executing the currently open script using include in the REPL with the project folder as the current directory.
