Skip to content

A5-2-6: Exclude cases with the same binary operator #231

Closed

Description

@lcartey

Affected rules

  • A5-2-6

Description

It's not well defined by the standard, but I think it's reasonable to exclude -> and . when considering whether an operand of a logical operator is binary.

Update: We thought the query was incorrectly reporting -> and . as binary operators. However, the problem was a confusion over this example:

foo->bar() && foo->baz() && foo->bang()

Where the query is reporting that this should be bracketed like so:

(foo->bar() && foo->baz()) && foo->bang()

This is a grey area in the rule itself - technically the title implies we should report this case. However, the rationale states:

Parentheses are required to add clarity in logical expressions making code easier to review versus code based only C++ operator precedence rules.

Which I think argues against asking for the developers to add unnecessary brackets in this case.

Proposed next steps:

  • Exclude cases where the nested binary operator is the same as the parent binary operator.
  • Refine the wording of the alert to ensure that it's clear where the brackets should be applied.

Example

foo->bar() && foo->baz() && foo->bang()

Activity

  1. added
    false positive/false negativeAn issue related to observed false positives or false negatives.
    Difficulty-LowA false positive or false negative report which is expected to take <1 day effort to address
    user-reportIssue reported by an end user of CodeQL Coding Standards
    on Mar 3, 2023
  2. changed the title [-]`A5-2-6`: Do not consider `->` and `.` as binary oprators[/-] [+]`A5-2-6`: Do not consider `->` and `.` as binary operators[/+] on Mar 3, 2023
  3. lcartey commented on Apr 19, 2023

    @lcartey
    ContributorAuthor

    Reviewing the query, the description above is not accurate. The query already only reports BinaryOperations, which does not include FieldAccesses (such as -> and .). I have investigated whether these could be caused by the use of operator->, either independently or in conjunction with templates but cannot reproduce the problem.

  4. changed the title [-]`A5-2-6`: Do not consider `->` and `.` as binary operators[/-] [+]`A5-2-6`: Consider excluding cases with the same binary operator[/+] on May 26, 2023
  5. lcartey commented on May 26, 2023

    @lcartey
    ContributorAuthor

    Issue body updated: the problem has been identified as related to (a) confusion caused by the alert message about where the brackets should apply and (b) an over-zealous rule.

  6. changed the title [-]`A5-2-6`: Consider excluding cases with the same binary operator[/-] [+]`A5-2-6`: Exclude cases with the same binary operator[/+] on May 26, 2023
  7. self-assigned this
    on Nov 5, 2023
  8. nbusser commented on Jun 2, 2024

    @nbusser

    I would like to report a similar over-zealous case when using the same comparison operator:

    (a == 0) && (b == 0) // True negative, OK
    
    (a == 0) && (b == 0) && ( c == 0) // Triggers A5-2-6 warning (zealous)
    
    ((a == 0) && (b == 0)) && ( c == 0) // Needs to do this to avoid A5-2-6 warning, which alters readability of code

    AUTOSAR C++14 is not really clear for this specific case.

  9. moved this from Assigned to Ready for review in Coding Standards Public Development Boardon Oct 17, 2024
  10. lcartey commented on Oct 17, 2024

    @lcartey
    ContributorAuthor

    This will be addressed by: #753

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

Difficulty-LowA false positive or false negative report which is expected to take <1 day effort to addressImpact-Highfalse positive/false negativeAn issue related to observed false positives or false negatives.top-25-fpsuser-reportIssue reported by an end user of CodeQL Coding Standards

Type

No type

Projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions