Issue context (addLocation / addFlow): What's the difference?

[guwirth (Günter Wirth)]

Hi,

I like to understand what the difference between issue addLocation and addFlow is?

• I found in your documentation: https://docs.sonarqube.org/latest/user-guide/issues/#header-1
• … and the API: https://github.com/SonarSource/sonarqube/blob/master/sonar-plugin-api/src/main/java/org/sonar/api/batch/sensor/issue/NewIssue.java

But for me it seems that addLocation is the same as addFlow with only one flow?

• When should I use which of the API function?
• What is the difference in the UI?

The items in both cases are NewIssueLocation. Each individual NewIssueLocation can have an own InputComponent (file, dir, module) and message. In some cases messages belong to locations outside of the project (e.g. a library):

• How can I handle this? Is it allowed to set file paths that do not exist in the project?

Regards,
Günter

[Julien_HENRY]

Hi @guwirth

@guwirth

it seems that addLocation is the same as addFlow with only one flow?

addLocation is technically the same as addFlow with only one location. It is currently stored the same way in the scanner report and in DB. Say otherwise, a flow with a single location is a secondary location.

@guwirth

• When should I use which of the API function?

Depends if you report a flow or a secondary location

• flow = code path that will lead to the issue
• secondary location = other locations that may help to understand the issue

@guwirth

• What is the difference in the UI?

All secondary locations are displayed at the same time when you select an issue (see for example some issues about cognitive complexity).

Flows can be collapsed, navigated, and in recent SQ versions we are supporting cross files locations, with code snippets.

@guwirth

The items in both cases are NewIssueLocation . Each individual NewIssueLocation can have an own InputComponent (file, dir, module) and message. In some cases messages belong to locations outside of the project (e.g. a library):

• How can I handle this? Is it allowed to set file paths that do not exist in the project?

No, you can only report locations on InputFile or on the project itself. We have deprecated InputDir for a while. One solution is to report the location on the InputProject, and provide more details in the location message.

[guwirth (Günter Wirth)]

Hello Julien,

thanks a lot for your detailed explanation!

addLocation is technically the same as addFlow with only one location

So the picture looks like this?

Issue (primary location)
  |-- Locations (List) <== addLocation
  |      |-- IssueLocation 1 (secondary location)
  |      |-- IssueLocation 2 (secondary location)
  |      |-- ...
  |-- Flows (List)
        |-- Flow 1 <== addFlow
        |      |-- IssueLocation 1 (secondary location)
        |      |-- IssueLocation 2 (secondary location)
        |      |-- ...
        |-- Flow 2 <== addFlow
        |      |-- IssueLocation 1 (secondary location)
        |      |-- IssueLocation 2 (secondary location)
        |      |-- ...
        |-- ...

Means for me the addFlow is the modern addLocation?

Regards,
Günter

[Julien_HENRY]

Depends at which level you look. From a functional point of view, the two concepts are different, and may in the future be handled technically in a different way.

But currently, technically, the underlying model is:

issue <--- 0 - n ---> flow <--- 1 - n ---> issue location

Adding a secondary location will ‘just’ add a flow with a single location.

[mason]

Hi Julien,
I found an issue here:

when we set primary issue in file a.cpp and add a secondary location in b.cpp, secondary location will become a primary issue in Sonarqube UI if user set sonar.exclusion=a.cpp

is this by desgin?