Hi Kumar,
I can't comment on the details of the actual expansion logic, nor the tests.
Looking at the overall structure I'm still unclear why more of this
isn't just hidden in win32 only files. Why do the new JLI_* methods have
to be JLI methods? I would have hoped that everything could be
hidden/handled inside CreateApplicationArgs/
One specific comment:
share/bin/main.c:
99 #ifdef _WIN32
100 {
101 int i = 0;
102 if (getenv(JLDEBUG_ENV_ENTRY) != NULL) {
103 printf("Windows original main args:\n");
104 for (i = 0 ; i < __argc ; i++) {
105 printf("wwwd_args[%d] = %s\n", i, __argv[i]);
106 }
107 }
108 }
Does MSC not permit declaration of i inside the for loop? It avoids the
need for the extra scope.
David
-----
On 27/07/2012 10:41 PM, Kumar Srinivasan wrote:
Hi,
Please review the fix
http://cr.openjdk.java.net/~ksrini/7146424/webrev.0/
to address:
7146424: Wildcard expansion for single entry classpath
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7146424
and
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7167744
Notes:
a. cmdtoargs.c will be pushed as a separate changeset using a separate CR
and with contributor attribution to akhil.ar...@oracle.com
b. src/solaris/bin/java_md.c is a redundant file and will be removed,
webrev for whatever reason is not reporting it.
Thanks
Kumar
