feat: privateModule linter

[thorimur]

This linter lints against nonempty modules that have only private declarations, and suggests adding @[expose] public section to the top.

This linter found 1 violation in Mathlib, which was fixed in #31822.

---

• depends on: chore: turn on module system in MathlibTest #31823

[github-actions]

PR summary 6ff93aa9da

Import changes for modified files

No significant changes to the import graph

Import changes for all files

Files	Import difference
Mathlib.Tactic.Linter.PrivateModule (new file)	2

---

Declarations diff

+ bar
+ foo
+ privateModule

You can run this locally as follows

## summary with just the declaration names:
./scripts/declarations_diff.sh <optional_commit>

## more verbose report:
./scripts/declarations_diff.sh long <optional_commit>

The doc-module for script/declarations_diff.sh contains some details about this script.

---

No changes to technical debt.

You can run this locally as

./scripts/technical-debt-metrics.sh pr_summary

• The relative value is the weighted sum of the differences with weight given by the inverse of the current value of the statistic.
• The absolute value is the relative value divided by the total sum of the inverses of the current values (i.e. the weighted average of the differences).

[adomani]

By the way, another standard request would be to add a test of this linter! 😄

[thorimur]

By the way, another standard request would be to add a test of this linter! 😄

I would love to do this, but the module system is not enabled in MathlibTest, so I can't activate the linter (hence the TODO)! 😅 (I had to test locally in a file in regular Mathlib.)

I'm experimenting with turning on the module system in MathlibTest in #31823, so if that's easy (🤞, not sure if it has been attempted yet, though), maybe let's merge that first and write some tests!

[grunweg]

Medium-term, we would like to add this to Mathlib.Init, right? Then this linter should be disabled by default, enabled in the lakefile, and this register_option should be public...

[thorimur]

Ah, makes sense. I'll do the latter three things here, right? I could also add it to Mathlib.Init here if you think so, or we can battle-test it first.

[grunweg]

Yes, the latter three can be done here - adding to Mathlib.Init can also happen once this linter is tested properly.

[thorimur]

#31823 builds just fine, so I've made this a dependent PR and added a couple of basic tests. :) Should I go ahead and add it to Mathlib.Init, or should we see if we see any edge cases first?

[thorimur]

0729c34 adds the linter to Mathlib.Init, but I can cherry pick this to another branch if I've misunderstood. :)

[adomani]

By the way, another standard request would be to add a test of this linter! 😄

I would love to do this, but the module system is not enabled in MathlibTest, so I can't activate the linter (hence the TODO)! 😅 (I had to test locally in a file in regular Mathlib.)

🤦 Sorry for not realising that there was a TODO about this in the PR already!

[thorimur]

By the way, another standard request would be to add a test of this linter! 😄

I would love to do this, but the module system is not enabled in MathlibTest, so I can't activate the linter (hence the TODO)! 😅 (I had to test locally in a file in regular Mathlib.)

🤦 Sorry for not realising that there was a TODO about this in the PR already!

Haha, no problem—better to err on the side of tests than otherwise, after all! 😁

[grunweg]

Thanks! The tests look good to me (I haven't looked at the linter's implementation so far).

[mathlib4-dependent-issues-bot]

This PR/issue depends on:

• chore: turn on module system in MathlibTest #31823
  By Dependent Issues (🤖). Happy coding!