On Wed, 8 Mar 2023 14:46:55 GMT, Daniel Jeliński <djelin...@openjdk.org> wrote:
>> permissive- is a pretty big problem for Windows code in general, in >> particular our very jint and jlong typedefs themselves on Windows are >> completely broken on Windows and compiling with permissive- will straight up >> make the entire JDK explode. Change here is particular is a small part of a >> much bigger (JDK wide) fix, if that's what you meant by it happening in >> other files. I'm not sure if the code here could be redefined as a jchar >> from the start though, should I give it a try? > > Hi @TheShermanTanker What complier are you using? What command line? > > `security.cpp` compiles just fine with `-permissive-` here, both debug and > release. I'm using this compiler: > >>cl.exe > Microsoft (R) C/C++ Optimizing Compiler Version 19.33.31629 for x64 > Copyright (C) Microsoft Corporation. All rights reserved. > > you can find the cmdline under support/native/jdk.crypto.mscapi/libsunmscapi. @djelinski After quite some time I think I know what the problem is: My fork is somehow missing the -Zc:wchar_t- flag in the configure script, and it's not because of the permissive- flag as I had initially though. Is there a reason we do this though? Barring this particular file it seems like the rest of the JDK compiles fine with it on my end. Maybe I could change this Pull Request to a build one instead of just a security-libs one in that case? ------------- PR: https://git.openjdk.org/jdk/pull/12907
