smengcl edited a comment on pull request #2710: URL: https://github.com/apache/hadoop/pull/2710#issuecomment-783214670
Thanks @GauthamBanasandra for the patch. I'm wondering if the [author](https://issues.apache.org/jira/browse/HDFS-11028) uses syscall write on purpose. According to the comment, heap could be corrupted if the `printf` call uses `malloc` or equivalent (depending on C/C++ library's implementation): https://github.com/apache/hadoop/blob/8783461e2ec3aff2630ea3574a69beb1d5c61e84/hadoop-hdfs-project/hadoop-hdfs-native-client/src/main/native/libhdfspp/examples/c/connect_cancel/connect_cancel.c#L43 https://github.com/apache/hadoop/blob/869317be0a1fdff23be5fc500dcd9ae4ecd7bc29/hadoop-hdfs-project/hadoop-hdfs-native-client/src/main/native/libhdfspp/examples/cc/connect_cancel/connect_cancel.cc#L47-L48 How about wrapping the `<unistd.h>` header import line and the existing approach in `#ifndef WINDOWS`? And use `#else` to wrap the new approach. Assuming Windows is the platform you are aiming for. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
