Search
Friday, 15th of September 2017, 2:53:53 UTC
11:16:44
hlavaty
hi on windoes, is there a way to get pid for run-program process, which is suitable for taskkill? sb-ext:process-pid gives a number which doesn't seem to be windows pid
11:20:07
jackdaniel
hlavaty: I believe that it is meant to be killed with sb-ext:process-close called on the return value of sb-ext:run-program
11:23:18
hlavaty
i hope it does kill the whole process tree
11:24:51
hlavaty
actually it looks like sb-ext:process-pid returns handle for which GetProcessID returns the actual pid
11:28:46
scymtym
ACTION thought PROCESS-CLOSE detaches the child process without killing it
11:28:49
hlavaty
sb-ext:process-close doesnt say anything about killing the process
11:29:12
jackdaniel
OK, sorry for confusing things then
11:29:40
scymtym
on non-win32 platforms there is PROCESS-KILL
11:31:24
hlavaty
scymtym: yeah, we are using that, but the issue is on windows
11:32:01
jackdaniel
then maybe (defun terminate-win-process (pid) (sb-ext:run-program "taskkill" '(…))) ;? if pid is correct
11:32:46
hlavaty
jackdaniel: yes, but as i said initially, sb-ext:process-pid doesn't return the pid
11:33:24
hlavaty
we'll try sb-ext:process-pid -> handle -> GetProcessId -> pid -> taskkill and see if it works
11:34:01
scymtym
hlavaty: maybe try this via sb-alien: https://msdn.microsoft.com/en-us/library/ms686714(VS.85).aspx
11:35:17
hlavaty
scymtym: yeah that is an option, although it doesnt seem to advertise killing the whole process subtree
11:38:27
scymtym
neither does PROCESS-KILL
11:39:15
scymtym
unless the signal is SIGTERM or SIGINT and the child process takes care of it itself
11:39:52
hlavaty
sb-ext:process-pid -> GetProcessId -> taskkill seems to work
12:16:08
stassats
the hppa build is broken because it has an ldb instruction, the new instruction format defines functions => conflict
12:16:19
stassats
maybe %ldb plus an instruction macro named ldb
12:17:37
stassats
or there's some other way
12:19:12
dougk
op-encoder-name can add an exception for LDB
12:20:08
stassats
now onto unknown type specifier: SYSTEM-AREA-POINTER
12:21:43
stassats
from IMMEDIATE-CONSTANT-SC
12:23:56
stassats
no other backend has sap as immediate
12:42:34
stassats
failed AVER: (= (LOGAND (SB!FASL::DESCRIPTOR-BITS #) SB!VM:WIDETAG-MASK) SB!VM:SIMPLE-FUN-WIDETAG)
12:47:30
stassats
and The variable SB!FASL::DESCRIPTOR-BITS is unbound.
12:48:09
stassats
oh, it's a function, # confused me
13:53:14
stassats
finally, no trace of define-static-fun
13:53:33
stassats
now, making assembly routines into functions
14:08:49
stassats
on darwin, compiling without -O2 breaks SLAD
14:09:54
stassats
probably worth investigating
14:11:50
stassats
during defrag_immobile_space=>set_forwarding_pointer
14:16:18
stassats
and only on darwin, works on linux even with clang
14:19:41
stassats
(i didn't want to deal with lldb)
14:22:59
stassats
Memory fault at 0x38087e8 (pc=0x38087ef
14:25:25
stassats
frame #15: 0x0000000009aaa84f sbcl`set_forwarding_pointer(pointer=0x0000000009aaa84f, newspace_copy=165909120) at forwarding-ptr.h:55
14:25:50
stassats
why would the pointer be at the same place as PC
Friday, 15th of September 2017, 14:53:53 UTC