Re: RFR: JDK-8223309: test failing due to self-assign-overloaded
Erik: Looks good. Tim I agreed to sponsor this off list. While doing a sanity build locally with the compiler we use at Oracle (Xcode 10.1), I discovered that clang does not let us use unknown warnings like we do with GCC: error: unknown warning option '-Wno-self-assign-overloaded'; did you mean '-Wno-self-assign-field'? [-Werror,-Wunknown-warning-option] I'm guessing -Wall is enabling that warning. Since we rely on being able to disable new warnings in new versions of the compilers we use, I added unknown-warning-option to the default set of disabled warnings for clang. New webrev: http://cr.openjdk.java.net/~erikj/8223309/webrev.01/ /Erik On 2019-05-03 08:02, Erik Joelsson wrote: Hello, Change looks good, but I would prefer a line break. We try to keep the lines in the makefiles/build system at least close to 80 chars when possible. No need for new webrev if you make this change. /Erik On 2019-05-03 07:44, Johan Vos wrote: Using MacOS X 10.14.4 and Xcode 10.2.1 I could not build jdk/jdk anymore: /Users/johan/openjdk/jdk/test/hotspot/gtest/classfile/test_symbolTable.cpp:62:6: error: explicitly assigning value of variable of type 'TempNewSymbol' to itself [-Werror,-Wself-assign-overloaded] s1 = s1; // self assignment ~~ ^ ~~ 1 error generated. I added an entry to the list of disabled warnings, and that fixed the problem. Bug: https://bugs.openjdk.java.net/browse/JDK-8223309 Webrev: http://cr.openjdk.java.net/~jvos/8223309/webrev.00/ - Johan
Re: RFR: JDK-8223309: test failing due to self-assign-overloaded
I agreed to sponsor this off list. While doing a sanity build locally with the compiler we use at Oracle (Xcode 10.1), I discovered that clang does not let us use unknown warnings like we do with GCC: error: unknown warning option '-Wno-self-assign-overloaded'; did you mean '-Wno-self-assign-field'? [-Werror,-Wunknown-warning-option] I'm guessing -Wall is enabling that warning. Since we rely on being able to disable new warnings in new versions of the compilers we use, I added unknown-warning-option to the default set of disabled warnings for clang. New webrev: http://cr.openjdk.java.net/~erikj/8223309/webrev.01/ /Erik On 2019-05-03 08:02, Erik Joelsson wrote: Hello, Change looks good, but I would prefer a line break. We try to keep the lines in the makefiles/build system at least close to 80 chars when possible. No need for new webrev if you make this change. /Erik On 2019-05-03 07:44, Johan Vos wrote: Using MacOS X 10.14.4 and Xcode 10.2.1 I could not build jdk/jdk anymore: /Users/johan/openjdk/jdk/test/hotspot/gtest/classfile/test_symbolTable.cpp:62:6: error: explicitly assigning value of variable of type 'TempNewSymbol' to itself [-Werror,-Wself-assign-overloaded] s1 = s1; // self assignment ~~ ^ ~~ 1 error generated. I added an entry to the list of disabled warnings, and that fixed the problem. Bug: https://bugs.openjdk.java.net/browse/JDK-8223309 Webrev: http://cr.openjdk.java.net/~jvos/8223309/webrev.00/ - Johan
Re: RFR: JDK-8223309: test failing due to self-assign-overloaded
Hello, Change looks good, but I would prefer a line break. We try to keep the lines in the makefiles/build system at least close to 80 chars when possible. No need for new webrev if you make this change. /Erik On 2019-05-03 07:44, Johan Vos wrote: Using MacOS X 10.14.4 and Xcode 10.2.1 I could not build jdk/jdk anymore: /Users/johan/openjdk/jdk/test/hotspot/gtest/classfile/test_symbolTable.cpp:62:6: error: explicitly assigning value of variable of type 'TempNewSymbol' to itself [-Werror,-Wself-assign-overloaded] s1 = s1; // self assignment ~~ ^ ~~ 1 error generated. I added an entry to the list of disabled warnings, and that fixed the problem. Bug: https://bugs.openjdk.java.net/browse/JDK-8223309 Webrev: http://cr.openjdk.java.net/~jvos/8223309/webrev.00/ - Johan