[ANN] ProfileSVG v0.2.0

Dear all,

I am pleased to announce the release of ProfileSVG v0.2.0.

At JuliaCon 2020, held after the release of ProfileSVG v0.1.3, there were talks on VS Code, Pluto.jl etc., which had a significant impact on the Julia development environment.
I do not have a clear answer as to the role of ProfileSVG in these revolutionary changes, but I do believe in the importance of “diversity”. :smile:

One of the important changes is that ProfileSVG now supports HTML display. It is just the original SVG enclosed in HTML tags, and in many environments there is little or no change. However, this means that you can now use the interactive feature within Pluto.jl.

Also, the default color scheme has been changed. You can customize the color scheme. In addition, support for several other options has been added.

In the next v0.3 series, I plan to improve the interactive UI.

Although the development speed is slow, I am looking forward to your feedback!

21 Likes

Thanks for the contribution! It would be nice to embed a profile viewer in VS Code! Either ProfileSVG.jl or ProfileView.jl. Can you elaborate on the differences between the two packages?

See also Julia VS Code extension version v0.17 released
This is the reason why I mentioned the VS Code. :smile:

The difference between ProfileSVG.jl and ProfileView.jl is that the former outputs SVG “image”, while the latter launches a stand-alone GTK-based widget.

The feature allows ProfileSVG to display profiling results in the Plot Pane of VS Code, just like any other plotting package.

1 Like

I thought that Juno had integration with ProfileView.jl at some point without a Gtk window, but I may be confusing something. What I meant is that users could just select a script and click a button in VS Code to profile. No need to write the profile calls and to warm up the compiler. These details should be hidden from the end user.

1 Like

Yes, Juno had (has) its own profile-viewing feature.

Just like Juno, VS Code has also its own profile-viewing feature. Without any additional package you can use the @profview macro and by adding the VS Code profile visualizer you can generate flame graphs.