Some ancient history: I wrote those ProcessBuilder tests back in 2005 and tripped over the changes to NLSPATH and XFILESEARCHPATH. I filed a bug back then, but it did not get fixed. I rediscovered those changes a decade later when putenv/getenv was found to be the root cause of SIGSEGV in real world usage.
I had forgotten about my workarounds in the ProcessBuilder test; rediscovered by recursive grep, which is still my workhorse code understanding tool. On Thu, Jul 27, 2017 at 7:13 AM, Philip Race <philip.r...@oracle.com> wrote: > Ok by me too ... it would not have crossed my mind to look at > ProcessBuilder > but I suppose it was trying to support the now deleted behaviours. > > -phil. > > > On 7/27/17, 6:29 AM, Alan Bateman wrote: > >> >> >> On 26/07/2017 22:20, Martin Buchholz wrote: >> >>> 1. JDK-8185365 <https://bugs.openjdk.java.net/browse/JDK-8185365> >>> >>> Tidy up leftover dead code after JDK-8136570 >>> http://cr.openjdk.java.net/~martin/webrevs/openjdk10/post-8136570-tidy/ >>> >> This looks okay to me. >> >> -Alan >> >
