Add NullableToStringCall rule #2903

t-kameyama · 2020-07-28T12:26:31Z

codecov · 2020-07-28T13:11:46Z

Codecov Report

Merging #2903 into master will increase coverage by 0.01%.
The diff coverage is 75.75%.

@@             Coverage Diff              @@
##             master    #2903      +/-   ##
============================================
+ Coverage     80.24%   80.26%   +0.01%     
- Complexity     2455     2467      +12     
============================================
  Files           422      423       +1     
  Lines          7420     7452      +32     
  Branches       1357     1367      +10     
============================================
+ Hits           5954     5981      +27     
+ Misses          762      759       -3     
- Partials        704      712       +8

Impacted Files	Coverage Δ	Complexity Δ
...turbosch/detekt/rules/bugs/NullableToStringCall.kt	`74.19% <74.19%> (ø)`	`10.00 <10.00> (?)`
...turbosch/detekt/rules/bugs/PotentialBugProvider.kt	`100.00% <100.00%> (ø)`	`2.00 <0.00> (ø)`
.../gitlab/arturbosch/detekt/api/internal/BaseRule.kt	`100.00% <0.00%> (+6.25%)`	`6.00% <0.00%> (ø%)`
...rturbosch/detekt/api/internal/CompilerResources.kt	`100.00% <0.00%> (+66.66%)`	`3.00% <0.00%> (+2.00%)`

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update fc769b5...beb3028. Read the comment docs.

cortinico

Thanks for your contribution @t-kameyama 🙏
There are several things that need to be addressed. Moreover, I advice for splitting this PR in maybe two.

Ideally, the rule should be simplified:

A simpler rule should just check if you're invoking kotlin.Any?.toString(). If so, report to the user that you can potentially return a "null".
Ideally the string template handling should be added afterwards. Potentially this can also be a separate rule as the name NullableToStringCall becomes confusing. Something like NullableStringInTemplate or something along this line.

cortinico · 2020-07-28T20:59:52Z

detekt-api/src/main/kotlin/io/gitlab/arturbosch/detekt/api/internal/BaseRule.kt


    open val ruleId: RuleId = javaClass.simpleName
    var bindingContext: BindingContext = BindingContext.EMPTY
+    var languageVersionSettings: LanguageVersionSettings? = null


I'm unsure we want to have a LanguageVersionSettings stored in the baserule + extending to API to always pass it around just for a single rule.

If possible, let's come up with a simpler version of this rule that doesn't need a LanguageVersionSettings explicitly. We can still update BaseRule if it brings value, but it should probably be discussed with the other maintainers first.

cortinico · 2020-07-28T21:02:27Z

...es-errorprone/src/main/kotlin/io/gitlab/arturbosch/detekt/rules/bugs/NullableToStringCall.kt

+                val descriptor = expression.descriptor() ?: return
+                val originalType = descriptor.returnType ?.takeIf { it.isNullable() } ?: return
+                @Suppress("DEPRECATION")
+                val dataFlowValueFactory = DataFlowValueFactoryImpl(languageVersionSettings)


This class is deprecated and has a clear deprecation message on top:
https://github.com/JetBrains/kotlin/blob/687d13a320b4a5697aedca3ed31f6568a40deafa/compiler/frontend/src/org/jetbrains/kotlin/resolve/calls/smartcasts/DataFlowValueFactoryImpl.kt#L28

I don't think we want to directly depend on this class.

We can get a container with the following function and get an instance of DataFlowValueFactory from it.
https://github.com/JetBrains/kotlin/blob/017f640f26e29f1b1d0909d89e812aca9173691c/compiler/cli/src/org/jetbrains/kotlin/cli/jvm/compiler/TopDownAnalyzerFacadeForJVM.kt#L130

However, I'm not using a container here because we only need an instance of DataFlowValueFactory.

cortinico · 2020-07-28T21:02:37Z

...es-errorprone/src/main/kotlin/io/gitlab/arturbosch/detekt/rules/bugs/NullableToStringCall.kt

+ *
+ * <compliant>
+ * fun foo(a: Any?): String {
+ *     return a?.toString ?: "-"


Suggested change

* return a?.toString ?: "-"

* return a?.toString() ?: "-"

cortinico · 2020-07-28T21:07:21Z

...rrorprone/src/test/kotlin/io/gitlab/arturbosch/detekt/rules/bugs/NullableToStringCallSpec.kt

+            """
+            val actual = subject.compileAndLintWithContext(env, code)
+            Assertions.assertThat(actual).hasSize(1)
+            Assertions.assertThat(actual.first().issue.id).isEqualTo("NullableToStringCall")


You can remove all the Assertions.assertThat(actual.first().issue.id)... as your subject contains only NullableToStringCall

cortinico · 2020-07-28T21:10:57Z

...rrorprone/src/test/kotlin/io/gitlab/arturbosch/detekt/rules/bugs/NullableToStringCallSpec.kt

+        it("reports when a nullable toString is implicitly called in a raw string template") {
+            val code = """
+                fun test(a: Any?) {
+                    println(${'"'}${'"'}${'"'}${'$'}a${'"'}${'"'}${'"'})


It feels like this test is a bit too complicated. I'm having a hard time finding a real-world scenario where such code would be used.

This code just test println("""$a""") so ${'"'}""${'$'}a""${'"'} should be enought.

BraisGabin · 2020-07-29T08:05:36Z

Why do you need LanguageVersionSettings? If you want to extend the detekt API we will need to move it to another PR and add some explanation for the change.

Ideally the string template handling should be added afterwards. Potentially this can also be a separate rule as the name NullableToStringCall becomes confusing. Something like NullableStringInTemplate or something along this line.

Why do you think that this should be two different rules? I mean, it's the same error. When you do this: "hello $any" you are doing an implicit call of toString(). NullableToStringCall doesn't seem a bad name thinking of this fact.

cortinico · 2020-07-29T17:24:48Z

Why do you think that this should be two different rules? I mean, it's the same error. When you do this: "hello $any" you are doing an implicit call of toString(). NullableToStringCall doesn't seem a bad name thinking of this fact.

I don't have a strong opinion on having two separate rules here. On the other hand I do believe the explicit/implicit toString() call should be distinguished (if is not through a separate rule, a config option will also work).

t-kameyama · 2020-08-04T05:04:46Z

Why do you need LanguageVersionSettings?

We don't want the following cases to be false positives.

fun test(a: Any?) {
    if (a != null) {
        println("$a") // 'a' is cast to 'String'
    }
}

To do so, we need to know if the variables have been smart cast. To know if it has been cast, we need the languageVersionSettings as follows.
https://github.com/t-kameyama/detekt/blob/16d7dd4dfbecfbc2b3ae0bb15cb1f3f1714cb7d5/detekt-rules-errorprone/src/main/kotlin/io/gitlab/arturbosch/detekt/rules/bugs/NullableToStringCall.kt#L69-L75

t-kameyama · 2020-08-04T05:34:31Z

Ideally the string template handling should be added afterwards. Potentially this can also be a separate rule as the name NullableToStringCall becomes confusing. Something like NullableStringInTemplate or something along this line.

I don't think it's necessary to separate the two because I think it's the same error.

cortinico · 2020-08-04T21:10:17Z

If you want to extend the detekt API we will need to move it to another PR and add some explanation for the change.

@t-kameyama please follow what @BraisGabin already suggested.

arturbosch · 2020-08-05T18:03:12Z

@t-kameyama thanks for the rule!
I agree with @cortinico that we should prefer a simpler version of this rule without handling string templates for now as we want to go into a different direction with the rule api (I wrote a longer comment in #2929 (comment)).

BraisGabin · 2020-08-09T11:11:32Z

@t-kameyama please refactor this PR so we can merge it :). And thanks again for this great contribution!

t-kameyama · 2020-08-10T00:32:31Z

PTAL

...es-errorprone/src/main/kotlin/io/gitlab/arturbosch/detekt/rules/bugs/NullableToStringCall.kt

...rrorprone/src/test/kotlin/io/gitlab/arturbosch/detekt/rules/bugs/NullableToStringCallSpec.kt

t-kameyama force-pushed the issue_2901 branch from 0686308 to a62cddd Compare July 28, 2020 12:58

cortinico requested changes Jul 28, 2020

View reviewed changes

t-kameyama added 3 commits August 4, 2020 13:41

Add NullableToStringCall rule

ac8d4f3

Fix document

16d7dd4

Fix test

b066aa8

t-kameyama force-pushed the issue_2901 branch from a62cddd to b066aa8 Compare August 4, 2020 04:59

t-kameyama requested review from BraisGabin and cortinico August 4, 2020 10:51

Add dataFlowValueFactory to BaseRule

439efbc

t-kameyama mentioned this pull request Aug 5, 2020

Add LanguageVersionSettings and DataFlowValueFactory to BaseRule #2929

Merged

arturbosch modified the milestones: 1.11.0, 1.11.0-RC2 Aug 9, 2020

Merge branch 'master' into issue_2901

041d52c

arturbosch approved these changes Aug 10, 2020

View reviewed changes

...es-errorprone/src/main/kotlin/io/gitlab/arturbosch/detekt/rules/bugs/NullableToStringCall.kt Show resolved Hide resolved

...rrorprone/src/test/kotlin/io/gitlab/arturbosch/detekt/rules/bugs/NullableToStringCallSpec.kt Outdated Show resolved Hide resolved

t-kameyama added 2 commits August 10, 2020 19:58

Add @since 1.11.0

01134a3

Add import for Assertions#assertThat

beb3028

t-kameyama force-pushed the issue_2901 branch from 1e54b17 to beb3028 Compare August 10, 2020 10:59

t-kameyama requested a review from arturbosch August 10, 2020 11:11

arturbosch approved these changes Aug 10, 2020

View reviewed changes

BraisGabin approved these changes Aug 10, 2020

View reviewed changes

arturbosch merged commit 3bfec16 into detekt:master Aug 11, 2020

t-kameyama deleted the issue_2901 branch August 11, 2020 10:49

Uh oh!

Add NullableToStringCall rule #2903

Add NullableToStringCall rule #2903

Uh oh!

Conversation

t-kameyama commented Jul 28, 2020

Uh oh!

codecov bot commented Jul 28, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

cortinico left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cortinico Jul 28, 2020

Choose a reason for hiding this comment

Uh oh!

cortinico Jul 28, 2020

Choose a reason for hiding this comment

Uh oh!

t-kameyama Aug 4, 2020

Choose a reason for hiding this comment

Uh oh!

cortinico Jul 28, 2020

Choose a reason for hiding this comment

Uh oh!

cortinico Jul 28, 2020

Choose a reason for hiding this comment

Uh oh!

cortinico Jul 28, 2020

Choose a reason for hiding this comment

Uh oh!

BraisGabin Jul 29, 2020

Choose a reason for hiding this comment

Uh oh!

BraisGabin commented Jul 29, 2020

Uh oh!

cortinico commented Jul 29, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

t-kameyama commented Aug 4, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

t-kameyama commented Aug 4, 2020

Uh oh!

cortinico commented Aug 4, 2020

Uh oh!

arturbosch commented Aug 5, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

BraisGabin commented Aug 9, 2020

Uh oh!

t-kameyama commented Aug 10, 2020

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

codecov bot commented Jul 28, 2020 •

edited

Loading

cortinico left a comment •

edited

Loading

cortinico commented Jul 29, 2020 •

edited

Loading

t-kameyama commented Aug 4, 2020 •

edited

Loading

arturbosch commented Aug 5, 2020 •

edited

Loading