On Fri, 14 Jan 2022 09:18:43 GMT, Johannes Bechberger <d...@openjdk.java.net> wrote:
>> make/autoconf/toolchain.m4 line 237: >> >>> 235: if test -n "$XCODEBUILD"; then >>> 236: # On Mac OS X, default toolchain to clang after Xcode 5 >>> 237: XCODE_VERSION_OUTPUT=`"$XCODEBUILD" -version 2>&1 | $HEAD -n 1` >> >> What's the warning? If it is truly expected, shouldn't it go to /dev/null ? > > The warning is > > objc[3881]: Class AMSupportURLConnectionDelegate is implemented in both > /usr/lib/libauthinstall.dylib (0x1e32b6b90) and > /Library/Apple/System/Library/PrivateFrameworks/MobileDevice.framework/Versions/A/MobileDevice > (0x1087e42c8). One of the two will be used. Which one is undefined. > objc[3881]: Class AMSupportURLSession is implemented in both > /usr/lib/libauthinstall.dylib (0x1e32b6be0) and > /Library/Apple/System/Library/PrivateFrameworks/MobileDevice.framework/Versions/A/MobileDevice > (0x1087e4318). One of the two will be used. Which one is undefined. > > as noted in the issues referenced by my bug report and we can ignore it (as > build scripts of other projects do). > > The problem with fully discarding the output is that there are other errors > that might cause the exit code to be non-zero and in this case displaying the > warning should be helpful for debugging. > Of course one could either just remove the expected warning or only display > the error output when the return code is non-zero, but these alternatives > seemed to cumbersome for me (and the first alternative to error-prone). OK. I've seen this a couple of times, and I hate warnings that can't be fixed. But it is what it is, thanks. ------------- PR: https://git.openjdk.java.net/jdk18/pull/95