Use node.position in add_message if available

[DanielNoord]

• Add yourself to CONTRIBUTORS if you are a new contributor.
• Add a ChangeLog entry describing what your PR does.
• If it's a new feature, or an important bug fix, add a What's New entry in
  doc/whatsnew/<current release.rst>.
• Write a good description on what the PR does.

Type of Changes

	Type
✓	✨ New feature

Description

Closes #5466.

I'd like to get approval on the code before I do the automatic updating of the tests. So please review this already even if tests fail 😄

[Pierre-Sassoulas]

The current change LGTM :)

[cdce8p]

The current change LGTM :)

👍🏻

[DanielNoord]

@Pierre-Sassoulas Let's merge this first before we add all the regression test PRs. Otherwise we need to keep rebasing this.

[cdce8p]

Missed that earlier, do we have an option for error messages to either opt out of or opt into using the position attribute? Say for example one message should be emitted for the whole function and not just the def <name>. Maybe a new parameter for add_message?

[DanielNoord]

Missed that earlier, do we have an option for error messages to either opt out of or opt into using the position attribute? Say for example one message should be emitted for the whole function and not just the def <name>. Maybe a new parameter for add_message?

If you supply line etc. when you add the message nothing will be overwritten. Is that enough?

Note: apparently 3.8 is broken... I'll merge the other test PRs first while working on this.

[coveralls]

Pull Request Test Coverage Report for Build 1973639791

• 34 of 34 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.02%) to 94.044%

---

Totals
Change from base Build 1973532638:	0.02%
Covered Lines:	15049
Relevant Lines:	16002

---

💛 - Coveralls

[Pierre-Sassoulas]

Thank you @DanielNoord !

[cdce8p]

Missed that earlier, do we have an option for error messages to either opt out of or opt into using the position attribute? Say for example one message should be emitted for the whole function and not just the def <name>. Maybe a new parameter for add_message?

If you supply line etc. when you add the message nothing will be overwritten. Is that enough?

My initial idea was a parameter to toggle between both options. Not passing each value individually. However, after comparing the changes here, I think it might no longer be necessary. So far I haven't found a case were it would be noticeable better to mark the whole code block. Let's leave it as it is for now.