clang 37: expression with side effects will be evaluated despite being used as an operand to 'typeid' [-Wpotentially-evaluated-expression]

[saper]

clang++37 -g -DDEBUG -DDEBUG_LVL="NONE" -Wall -DLIBSASS_VERSION="\"3.2.5-266-ge12f9\"" -std=c++0x -I /home/saper/sw/libsass/include -fPIC -fPIC -c -o src/ast.o src/ast.cpp
In file included from src/ast.cpp:1:
src/ast.hpp:2054:18: warning: expression with side effects will be evaluated despite being used as an operand to 'typeid' [-Wpotentially-evaluated-expression]
      if (typeid(*(*this)[0]) == typeid(Type_Selector))
                 ^
src/ast.hpp:2073:21: warning: expression with side effects will be evaluated despite being used as an operand to 'typeid' [-Wpotentially-evaluated-expression]
             typeid(*(*this)[0]) == typeid(Parent_Selector);
                    ^
src/ast.cpp:276:21: warning: expression with side effects will be evaluated despite being used as an operand to 'typeid' [-Wpotentially-evaluated-expression]
        if ((typeid(*(*rhs)[i]) == typeid(Pseudo_Selector) || typeid(*(*rhs)[i]) == typeid(Wrapped_Selector)) && (*rhs)[L-1]->is_pseudo_element())
                    ^
src/ast.cpp:276:70: warning: expression with side effects will be evaluated despite being used as an operand to 'typeid' [-Wpotentially-evaluated-expression]
        if ((typeid(*(*rhs)[i]) == typeid(Pseudo_Selector) || typeid(*(*rhs)[i]) == typeid(Wrapped_Selector)) && (*rhs)[L-1]->is_pseudo_element())
                                                                     ^
src/ast.cpp:284:20: warning: expression with side effects will be evaluated despite being used as an operand to 'typeid' [-Wpotentially-evaluated-expression]
        if (typeid(*(*rhs)[i]) == typeid(Pseudo_Selector) || typeid(*(*rhs)[i]) == typeid(Wrapped_Selector))
                   ^
src/ast.cpp:284:69: warning: expression with side effects will be evaluated despite being used as an operand to 'typeid' [-Wpotentially-evaluated-expression]
        if (typeid(*(*rhs)[i]) == typeid(Pseudo_Selector) || typeid(*(*rhs)[i]) == typeid(Wrapped_Selector))
                                                                    ^
6 warnings generated.
clang++37 -g -DDEBUG -DDEBUG_LVL="NONE" -Wall -DLIBSASS_VERSION="\"3.2.5-266-ge12f9\"" -std=c++0x -I /home/saper/sw/libsass/include -fPIC -fPIC -c -o src/sass_context.o src/sass_context.cpp
clang++37 -g -DDEBUG -DDEBUG_LVL="NONE" -Wall -DLIBSASS_VERSION="\"3.2.5-266-ge12f9\"" -std=c++0x -I /home/saper/sw/libsass/include -fPIC -fPIC -c -o src/sass_interface.o src/sass_interface.cpp
clang++37 -g -DDEBUG -DDEBUG_LVL="NONE" -Wall -DLIBSASS_VERSION="\"3.2.5-266-ge12f9\"" -std=c++0x -I /home/saper/sw/libsass/include -fPIC -fPIC -c -o src/util.o src/util.cpp
In file included from src/util.cpp:2:
src/ast.hpp:2054:18: warning: expression with side effects will be evaluated despite being used as an operand to 'typeid' [-Wpotentially-evaluated-expression]
      if (typeid(*(*this)[0]) == typeid(Type_Selector))
                 ^
src/ast.hpp:2073:21: warning: expression with side effects will be evaluated despite being used as an operand to 'typeid' [-Wpotentially-evaluated-expression]
             typeid(*(*this)[0]) == typeid(Parent_Selector);
                    ^

May or may be not harmless.

[mgreter]

Is there a "definition of done" for this Bug? I didn't see this warning so far myself? Maybe you could give some more information of used compiler and environment?

[cvrebert]

Environment:
OS X El Capitan 10.11.1

$ c++ --version
Apple LLVM version 7.0.0 (clang-700.1.76)
Target: x86_64-apple-darwin15.0.0
Thread model: posix

libsass: d18808d
sass/sassc@7efa9c4

Full build output: https://gist.github.com/cvrebert/bd95d26561e14de2cead

[mgreter]

We don't see this on the Travis Mac OSX CI build and I don't have access or connection to any Mac, so someone else needs to step in here to analyze this further ...

[cvrebert]

Looks like the Travis build is using an older version of clang (and possibly of OS X?):

$ clang --version
Apple LLVM version 6.0 (clang-600.0.54) (based on LLVM 3.5svn)
Target: x86_64-apple-darwin13.4.0

[saper]

On Fri, 6 Nov 2015, Marcel Greter wrote: Is there a "definition of done" for this Bug? I didn't see this warning so far myself? Maybe you could give some more information of used compiler and environment?

this was clang 3.7 as the title says... definition of done: potential issue reported by the warning is resolved.

[saper]

On Fri, 6 Nov 2015, Marcel Greter wrote: We don't see this on the Travis [Mac OSX CI build] [1] and I don't have access or connection to any Mac, so someone else needs to step in here to analyze this further ...

you need clang at least 3.7 on any platform to see this. It's related to rtti and is probably mit platform-dependent.

[mgreter]

Fixed in #1707