Code minings are not correctly redrawn when switching between in-line and line header mode

[travkin79]

Issue description

I'm working on an Eclipse plug-in that creates code minings. While testing these code minings, I found out that switching between in-line code minings and line header code minings (in my case by changing preference settings) causes some minings not to be redrawn.

In more detail:

1. case: When I have in-line code minings in my editor and I switch preferences to use line header code minings instead, some code minings are redrawn as line header code minings, but some stay beeing drawn as in-line code minings, instead of redrawing all code minings as line header code minings.
2. case: When I have line header code minings in my editor and I switch preferences to use in-line code minings instead, some code minings are redrawn as in-line code minings, but some stay beeing drawn as line header code minings, instead of redrawing all code minings as in-line code minings.

First analysis

I debugged the executed code and found out that switching preferences between in-line and line header code minings resulted in correctly re-creating all code minings, but the org.eclipse.jface.internal.text.codemining.CodeMiningManager mixes up CodeMiningLineContentAnnotations with CodeMiningLineHeaderAnnotations. If the manager finds an existing code mining annotation for some position, it reuses the annotation without checking if it has the right type. It seems, the CodeMiningManager should instead replace CodeMiningLineContentAnnotations with CodeMiningLineHeaderAnnotations when switching from in-line code minings to line header code minings and vice versa.

Steps to reproduce

Unfortunately, the plug-in I'm working on is not open source. Thus, I cannot provide simple steps to reproduce the issue.

In my case, I have a file with plenty code minings, often multiple code minings per line and often several consecutive lines with at least one code mining per line. Having this file open in an editor and then switching the code mining preferences for the custom editor causes the issue.

Tested under this environment:

• Eclipse 2025-03 (started via launch configuration from another instance of the IDE)
• Windows 11 Pro

[BeckerWdf]

by changing preference settings

Which setting does change this?

I debugged the executed code and found out that switching preferences between in-line and line header code minings resulted in correctly re-creating all code minings, but the org.eclipse.jface.internal.text.codemining.CodeMiningManager mixes up CodeMiningLineContentAnnotations with CodeMiningLineHeaderAnnotations

As you already have performed an details analysis. Would you like to provide a fix via a pull request? As your plug-in is not open source that would have the advantage that you directly can test your fix in combination with your plug-in.

[travkin79]

Hello @BeckerWdf,

Which setting does change this?

It's a custom preferences setting in my plug-in that switches between creating either in-line code minings or line header code minings.

As you already have performed an details analysis. Would you like to provide a fix via a pull request? As your plug-in is not open source that would have the advantage that you directly can test your fix in combination with your plug-in.

I performed the analysis to be sure where the issue is coming from. I think, I could implement the fix, but due to a high workload it could take quite a while.

I also found another issue with drawing code minings. Sometimes line header code minings are drawn into wrong lines. Scrolling in the editor triggers redrawing and then the issue usually disappears. Did someone else observe that issue? If it's not a known issue, I'd create another GitHub issue for that.

[BeckerWdf]

I performed the analysis to be sure where the issue is coming from. I think, I could implement the fix, but due to a high workload it could take quite a while.

So just to be transparent here. We are a limited number of developers on Eclipse Platform and we have to prioritise our work. If the bug does not happen on the Eclipse IDE it's unlikely that we can provide a fix soon. Especially if we don't have a reproducer for this. As you cannot share your plugin you might be able to write a small minimal Java / SWT program that shows this issue.

[travkin79]

Hi @BeckerWdf,
I understand. I'll try to find out if I can reproduce the issue with some editor from the Eclipse IDE. If not, I'll try to implement a fix.