Thanks David and Alan, shall I find some one to commit it ?
On Thu, Apr 19, 2012 at 8:53 AM, David Holmes <david.hol...@oracle.com>wrote: > On 19/04/2012 4:05 AM, Alan Bateman wrote: > >> On 18/04/2012 14:02, David Holmes wrote: >> >>> On 18/04/2012 10:23 PM, Sean Chou wrote: >>> >>>> Hi David, Alan, >>>> >>>> So is the patch acceptable ? >>>> >>> >>> There is still the matter of the unexpected NULL if strdup fails. I'd >>> need to see the clients for this code to see how they handle failure. >>> My concern is the case where the caller sees a NULL return which >>> indicates an error, and so accesses the msg and now potentially hits >>> another NULL. It's unlikely but ... >>> >> I looked through the usages and nothing obvious jumps out. HotSpot will >> invoke ZIP_Open and that now returns a fixed string in the event of a >> failure. The library code opens the zip file directly and then uses >> ZIP_Put_In_Cache0 which seems to be handling this case. >> > > Yes looks like we only have two bits of client code here and ZipFile.c > checks for NULL. > > Ok. > > David > ----- > > -Alan. >> > -- Best Regards, Sean Chou
