[Close #961] Add --no-debug <regex> flag

[ivanovaleksey]

Close #961

Add --no-debug <regex> and Builder::no_debug(String) to supply regexes for matching against types that we should not derive or implement Debug for.

Adding this feature involves:

• Adding a new RegexSet member to bindgen::Builder (similar to the whitelisted_types set).

• A Builder method to add strings to that RegexSet.

• Plumbing in src/options.rs to convert --no-debug <regex> CLI flags into invocations of the builder method.

• Making the MonotoneFramework::constrain function in src/ir/analysis/derive_debug.rs check if the given item is explicitly marked not to be Debug, and if so, inserting it into the self.cannot_derive_debug set via return self.insert(id).

• Skipping any item explicitly marked not to be Debug in src/codegen/derive_debug.rs so that we don't generate an impl Debug for ... for these types when --impl-debug is passed.

• Tests!

  • When --impl-debug is also supplied

  • When --impl-debug is not also supplied

  • When the no-debug type is transitively referenced by a whitelisted item

  • When the no-debug type is explicitly whitelisted

  • When the no-debug type is marked opaque

[highfive]

Warning

• These commits modify unsafe code. Please review it carefully!

[ivanovaleksey]

r? @pepyakin

[pepyakin]

Thanks for your pull request! Sorry for not getting to this earlier!

Looks like there are some tests are failing because of headers differs from expectations, but the only difference I can see is formatting! Looks strange, do you know guys what's going on @emilio / @fitzgen ?

Also, looks like --impl-debug thing is not there, is there any particular reason for that?

Aside of that, your PR looks great!

[fitzgen]

Its possible that we need to re-run rustfmt on the test expectations, and after that there won't be formatting differences between the generated code and the expectations.

[ivanovaleksey]

Thanks guys,

I have just opened this PR to ensure I am on the right way.
--impl-debug is going to be done.
I will check failing specs.

[ivanovaleksey]

I am not sure about why tests failed.
According to the job there was installed rustfmt-nightly v0.3.1.

I installed the same version locally and experimented on a file tests/headers/func_ptr_in_struct.h.
After cargo test it becomes reformatted with the following diff:

index 290a148c..3d3439ab 100644
--- a/tests/expectations/tests/func_ptr_in_struct.rs
+++ b/tests/expectations/tests/func_ptr_in_struct.rs
@@ -1,11 +1,10 @@
 /* automatically generated by rust-bindgen */
                 concat!("Alignment of ", stringify!(Type))
             );
         }
     }
     #[derive(Debug, Copy, Clone, PartialEq, Eq, Hash)]
-    pub enum a {}
+    pub enum a {
+}
 }

But if I run rustfmt directly:

rustfmt tests/expectations/tests/func_ptr_in_struct.rs --config-path tests/rustfmt.toml

it doesn't reformat enum a at all.

I don't know where rustfmt is getting run by bindgen but I found something similar in tests/tests.rs. But it invoke the same nightly version. Maybe these formatting differences might be due to some additional configuration I have missed?

[bors-servo]

☔ The latest upstream changes (presumably #1330) made this pull request unmergeable. Please resolve the merge conflicts.

[pvdrz]

I'm closing this PR as the merged PR #1858 was based over this one..