You should also not use the VM crash protection with multi-threaded
code. It overwrites VM signal handlers for the duration of native
calls.
On Jan 30, 2009, at 4:25 PM, Erik Earle wrote:
I tried to duplicate the c example for dnotify.c here:
http://souptonuts.sourceforge.net/code/dnotify.c.html
The attached is what I came up with. It works... somewhat.
- I get segmentation faults if I do anything other than touch or rm
a file, (opening a file with vi or 'echo foo > foo.txt' causes the
fault)
- I don't get modified events.
- In spite of setting the flag for the calling the 3 arg callback,
it still only calls the single arg callback.
I'm running:
- Ubuntu Gutsy, Linux xxxxx 2.6.24-23-generic #1 SMP Thu Nov 27
18:44:42 UTC 2008 i686 GNU/Linux
- java version "1.6.0_07"
- jna-3.0.7
I've tried to get a memory dump using the instructions here:
https://jna.dev.java.net/
#struct_debug (the "-Djna.dump_memory=true" prop) but I the output
just show this:
sigaction:
DNotify$SigAction(allocated@0x81155f0 (144 bytes)) {
DNotify$__sighandler_t
sa_sighandler@0=com.sun.jna.examples.linux.DNotify$1@1cd8669
DNotify$__sigaction_t
sa_sigaction@4=com.sun.jna.examples.linux.DNotify$2@337838
DNotify$__sigset_t sa_mask@8=DNotify$__sigset_t(allocated@0x8158810
(128 bytes)) {
int __val[32]@0=[I@119cca4
}
int sa_flags@88=4
DNotify$__sigrestorer_t
sa_restorer@8c=com.sun.jna.examples.linux.DNotify$3@ca2dce
}
memory dump
[00000000]
[00000000]
[00000000]
.... (just keeps repeating the zeros)
sometimes I get a core dump, that's also attached.
Any thoughts?
<
hs_err_pid28433
.log
<
DNotify
.java
To unsubscribe, e-mail: user...@jna.dev.java.net
For additional commands, e-mail: user...@jna.dev.java.net