[9] RFR of 8074821: Resolve disabled warnings for libnio

[Brian Burkhalter]

I would like to reprise this thread

http://mail.openjdk.java.net/pipermail/nio-dev/2015-May/003126.html

concerning https://bugs.openjdk.java.net/browse/JDK-8074821 starting from the message included below.

Please review this new patch at your convenience:

http://cr.openjdk.java.net/~bpb/8074821/webrev.01/

It differs from the prior one in the addition of the lines

+// GetVersionEx is deprecated; disable the warning until a replacement is found
+#pragma warning(disable : 4996)

just before the functions in which GetVersionEx() is invoked in the Windows source code of Iocp.c and Net.c. There is precedence elsewhere in the source for using this approach. I don’t think that in the Net.c source it is necessary to repeat the pragma directive [1] but it does serve to emphasize the functions which call GetVersionEx().

For lack of a fully functional Solaris VM image I have not run the Selector tests as suggested below. I an unsure as to how important that is to the purpose of this issue.

Thanks,

Brian

[1] https://msdn.microsoft.com/en-us/library/2c8f766e.aspx

On May 8, 2015, at 12:01 AM, Alan Bateman <Alan.Bateman at oracle.com> wrote:

> On 07/05/2015 19:15, Brian Burkhalter wrote:
>> :
>> 
>> Would something like this make more sense?
>> 
>> --- a/src/java.base/unix/native/libnio/ch/IOUtil.c
>> +++ b/src/java.base/unix/native/libnio/ch/IOUtil.c
>> @@ -1,5 +1,5 @@
>>  /*
>> - * Copyright (c) 2000, 2012, Oracle and/or its affiliates. All rights reserved.
>> + * Copyright (c) 2000, 2015, Oracle and/or its affiliates. All rights reserved.
>>   * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
>>   *
>>   * This code is free software; you can redistribute it and/or modify it
>> @@ -129,7 +129,8 @@
>>          JNU_ThrowIOExceptionWithLastError(env, "getrlimit failed");
>>          return -1;
>>      }
>> -    if (rlp.rlim_max < 0 || rlp.rlim_max > java_lang_Integer_MAX_VALUE) {
>> +    if (rlp.rlim_max == RLIM_INFINITY ||
>> +        rlp.rlim_max > (rlim_t)java_lang_Integer_MAX_VALUE) {
>>          return java_lang_Integer_MAX_VALUE;
>>      } else {
>>          return (jint)rlp.rlim_max;
>> 
> This looks good as it doesn't make any assumption on the value of RLIM_INFINITY. If you have VM images with Solaris then it's easy to edit /etc/system, reboot, and then attempt to run the Selector tests.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/nio-dev/attachments/20150817/d62e6695/attachment.html>