[llvm-bugs] [Bug 36220] New: Static analyzer checks have the wrong check name

via llvm-bugs llvm-bugs at lists.llvm.org
Fri Feb 2 18:21:31 PST 2018


https://bugs.llvm.org/show_bug.cgi?id=36220

            Bug ID: 36220
           Summary: Static analyzer checks have the wrong check name
           Product: clang
           Version: unspecified
          Hardware: Macintosh
                OS: MacOS X
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Static Analyzer
          Assignee: dcoughlin at apple.com
          Reporter: michael.eisel at gmail.com
                CC: llvm-bugs at lists.llvm.org

Hi,

I'm working with the version of clang shipped with Xcode 9.0.1 (Apple LLVM
version 9.0.0 (clang-900.0.38)) and I seem to have found a bug where the
check_name parameter of the clang static analysis plist file is incorrect.
Since I use this check_name as an identifier for automatic linting in CI, it's
important to get it right (e.g., to track the change in violations across
commits). When I enable the lint nullability.NullPassedToNonnull, that check
name is returned for a violation that is actually
nullability.NullReturnedFromNonnull. When I disable
nullability.NullPassedToNonnull, the check_name is correctly reported.

Command where check_name is correct (nullability.NullReturnedFromNonnull):

/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/clang
-isysroot
/Applications/Xcode.app/Contents/Developer/Platforms/iPhoneSimulator.platform/Developer/SDKs/iPhoneSimulator.sdk
-arch x86_64 -mios-simulator-version-min\=10.0 -std\=gnu99 -c a.m -o out
-Wno-error -Xclang -analyzer-config -Xclang mode\=deep -Xclang -analyzer-config
-Xclang path-diagnostics-alternate\=true -Xclang -analyzer-checker -Xclang
nullability.NullReturnedFromNonnull -Xclang -analyzer-disable-checker -Xclang
nullability.NullPassedToNonnull -Xclang -analyzer-output\=plist -Xclang
-analyzer-opt-analyze-nested-blocks -D__clang_analyzer__ --analyze && cat out 


Command where check_name is NullPassedToNonnull:

/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/clang
-isysroot
/Applications/Xcode.app/Contents/Developer/Platforms/iPhoneSimulator.platform/Developer/SDKs/iPhoneSimulator.sdk
-arch x86_64 -mios-simulator-version-min\=10.0 -std\=gnu99 -c a.m -o out
-Wno-error -Xclang -analyzer-config -Xclang mode\=deep -Xclang -analyzer-config
-Xclang path-diagnostics-alternate\=true -Xclang -analyzer-checker -Xclang
nullability.NullReturnedFromNonnull -Xclang -analyzer-checker -Xclang
nullability.NullPassedToNonnull -Xclang -analyzer-output\=plist -Xclang
-analyzer-opt-analyze-nested-blocks -D__clang_analyzer__ --analyze && cat out


a.m:
#import <Foundation/Foundation.h>

NS_ASSUME_NONNULL_BEGIN

@interface MyObject : NSObject

- (id)foo;

@end

NS_ASSUME_NONNULL_END

@implementation MyObject

- (id)foo
{
    return nil;
}

@end

Here's one response from a mailing list:
"It seems that we're using the same default CheckName for all nullability
checks, even though we retrieve the correct names in NullabilityChecker.cpp's
REGISTER_CHECKER - that's missing boilerplate for having the same checker
object in C++ code pretend to be multiple user-facing checkers. And we never
noticed only because nobody was using this field."

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20180203/ac02bc1a/attachment.html>


More information about the llvm-bugs mailing list