The fix looks reasonable to me, basically, if the command argument format is not legal, it’s possible we won’t find the main class when doing argument file expansion, and the index is -1 which could cause crash on Windows platform for the wildcard processing.
I think we should add a test case for this, probably in test/jdk/tools/launcher/Arrrghs.java. As I understand it, the argument validation is done in a later stage after calling JLI_Launch. Cheers, Henry > On Dec 2, 2019, at 2:12 AM, Alan Bateman <alan.bate...@oracle.com> wrote: > > On 20/11/2019 19:42, Nikola Grcevski wrote: >> : >> >> Please let me know if this approach is acceptable for the current bug report >> and I'll make a webrev and include appropriate launcher tests. I was >> thinking the tests should do extra validation on the output from >> _JAVA_LAUNCHER_DEBUG on Windows. >> > I think you're in the right area but I would have expected the arg index to 0 > here. Henry Jen (cc'ed) might have some comments on this. > > -Alan
