Enabled errors on unused variables.

[orizi]

Stack:

• Moved new diagnostics from enabled at edition to warning. #4462
• Added warning diagnostics severity. #4459
• Enabled errors on unused variables. #4457 ⬅

⚠️ Part of a stack created by spr. Do not merge manually using the UI - doing so may have unexpected results.

---

This change is 

[mkaput]

Reviewable status: 0 of 53 files reviewed, 1 unresolved discussion (waiting on @gilbens-starkware and @orizi)

a discussion (no related file):
Errors? Has this been discussed with the community members, and has feedback been gathered about this idea?

I believe most languages implement this as warnings; the only language I know that emits fatal errors is Golang, so I expect this to be a rather unexpected behaviour.

My personal feedback is that this should be a warning, as I tend to often leave unfinished code and run tests over it. But I'm happy to hear others voice!

---

[orizi]

Reviewable status: 0 of 53 files reviewed, 1 unresolved discussion (waiting on @gilbens-starkware and @mkaput)

a discussion (no related file):

Previously, mkaput (Marek Kaput) wrote…

Errors? Has this been discussed with the community members, and has feedback been gathered about this idea?

I believe most languages implement this as warnings; the only language I know that emits fatal errors is Golang, so I expect this to be a rather unexpected behaviour.

My personal feedback is that this should be a warning, as I tend to often leave unfinished code and run tests over it. But I'm happy to hear others voice!

since it is very easy to prevent the error (just adding an _ before) - I see no reason for this to not be an error.
warning are currently not a concept we have - and i see no reason to add it just for that.

---

[gilbens-starkware]

Reviewed 53 of 53 files at r1, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @mkaput)

[mkaput]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @orizi)

a discussion (no related file):

Previously, orizi wrote…

since it is very easy to prevent the error (just adding an _ before) - I see no reason for this to not be an error.
warning are currently not a concept we have - and i see no reason to add it just for that.

I understand the implementation reasons for this, but I just want to say, that even this simple fix with _ requires manual work in the code. It will have a tendency to kick the developer out of their flow of thought (for example, in the scenario I outlined in the previous comment). That's a subtle UX inconvenience, that may affect users' perception of the language.

I, honestly, strongly suggest gathering community feedback before releasing this.

---

[milancermak]

100% agree with what Marek said. This shouldn't be an error.

[orizi]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @mkaput)

a discussion (no related file):

Previously, mkaput (Marek Kaput) wrote…

I understand the implementation reasons for this, but I just want to say, that even this simple fix with _ requires manual work in the code. It will have a tendency to kick the developer out of their flow of thought (for example, in the scenario I outlined in the previous comment). That's a subtle UX inconvenience, that may affect users' perception of the language.

I, honestly, strongly suggest gathering community feedback before releasing this.

makes sense - pre-adding a warning concept.

---

[orizi]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @mkaput)

a discussion (no related file):

Previously, orizi wrote…

makes sense - pre-adding a warning concept.

happens in the following PR.
the PR after it will be removing the edition dependency from it.
scarb wrapper will need to support skipping the diagnostics for these cases.

---

[mkaput]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @orizi)

a discussion (no related file):

scarb wrapper will need to support skipping the diagnostics for these cases.

I don't fully follow, but I expect this to get more clear when I'll see the code. If DiagnosticReporter gets the information about the level of the diagnostic, we will print them as warn:. The question is, what should be the out coming Result of the compilation. I think perhaps we could introduce a custom enum: Ok, Warnings, Err here.

---

[mkaput]

Reviewed 53 of 53 files at r1, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @orizi)

[gilbens-starkware]

Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @orizi)