| Re: Incredible interaction between graphics and DLM: SOS! [message #61863] |
Thu, 07 August 2008 07:22 |
fabio.tosetti
Messages: 2
Registered: August 2008
|
Junior Member |
|
|
Now the problem is more clear to me, thank you.
Of course I don't like to make a mess with signal handlers... I think
that the best solution is to run the application that uses Pyro in e
separate process, and do inter-process communication, that we have
already implemented.
|
|
|
|
| Re: Incredible interaction between graphics and DLM: SOS! [message #61878 is a reply to message #61863] |
Tue, 05 August 2008 09:40  |
Allan Whiteford
Messages: 117
Registered: June 2006
|
Senior Member |
|
|
Fabio,
IDL will be sending signals back and forward with all sorts of other
stuff to handle graphics. Note that the man page for sleep says:
"sleep() may be implemented using SIGALRM; mixing calls to alarm() and
sleep() is a bad idea.
Using longjmp() from a signal handler or modifying the handling of
SIGALRM while sleeping will cause undefined results."
The IDL command line will have signals to deal with the graphics
subsystem which can/will cause sleep(20) to return early. There may also
be "longjmp"s involved.
This shouldn't affect most other things you want to do, for instance if
instead of sleep(20) you called a function which happened to take 20
seconds to do something then it should not return early.
Pyro is presumably also using signals and the like to handle
communication; I don't know anything about it.
Not really a solution to your problem but maybe it'll help you to look
in the right place. Mess with the signal handlers at your own risk though!
Thanks,
Allan
fabio.tosetti@gmail.com wrote:
> I have a very simple file set (see below), with only one function,
> idl_test(), written just to isolate the problem.
>
> The following instruction simply sleep for 20 seconds and correctly
> returns:
>
> IDL> print, idl_test()
>
>
> The following sequence cause the idl_test() to return BEFORE the
> expected time:
>
> IDL> err = dialog_message("A dialog")
> IDL> print, idl_test()
>
> The idl_test() returns after about 3-4 seconds.
> The same happens if I plot a graphic before calling idl_test().
>
> Of course in my application I don't want to sleep(), but to call some
> Python code that do remote communication (Pyro), but the result is
> exactly the same, and it's very frustrating :-(...
>
> THE EXAMPLE FILES ARE:
>
> 1) idl_test.dlm:
>
> MODULE idl_test
> DESCRIPTION idl test library
> VERSION 1.0
> SOURCE LB
> BUILD_DATE AUG 05 2008
>
> FUNCTION IDL_TEST 0 0
>
> 2) idl_test.h:
>
> #ifndef IDL_TEST_H_INCLUDE
> #define IDL_TEST_H_INCLUDE
>
> extern "C"{
> int IDL_Load(void);
> }
>
> #endif /*IDL_TEST_H_INCLUDE*/
>
>
> 2) idl_test.cpp file:
>
> #include "idl_test.h"
>
> #include <stdio.h>
> #include <string.h>
> #include <errno.h>
> #include <stdlib.h>
> #include <sys/types.h>
> #include <unistd.h>
>
> extern "C"{
> #include "idl_export.h"
> }
>
> IDL_VPTR idl_test(int lArgc, IDL_VPTR Argv[]) {
>
> // TEST
> int sec = 20;
> printf("Sleeping for %d seconds... will timeout earlier!\n", sec);
> sleep(sec);
> return IDL_StrToSTRING("idl_test returned before 20 seconds!!!");
> }
>
>
> int IDL_Load(void) {
>
> // These tables contain information on the functions and procedures
> // that make up the TESTMODULE DLM. The information contained in
> these
> // tables must be identical to that contained in testmodule.dlm.
> //
> static IDL_SYSFUN_DEF2 function_addr[] = {
>
> { {(IDL_SYSRTN_GENERIC) idl_test}, "IDL_TEST", 0, 0, 0, 0},
>
> };
>
> // Register my routines: the routines must be specified exactly
> the same
> // as in .dlm.
> return IDL_SysRtnAdd(function_addr, TRUE,
> IDL_CARRAY_ELTS(function_addr));
> }
>
> 4) Makefile:
>
>
> all: idl_test.so
>
> # ITT had the great idea of changing the name of
> # the installation directory since idl6.3
> # So if you have idl < 6.4 use
> #IDLDIR = /usr/local/rsi/idl/external/include
> # else use
> IDLDIR = /usr/local/itt/idl/external/include
>
> CPPFLAGS = -Wall -W -Wreturn-type -Wunused -D_GNU_SOURCE
> MORE_INCLUDE = -I$(IDLDIR) -I/usr/include/python
>
> %.o:%.h
>
> .cpp.o:
> g++ $(CPPFLAGS) $(MORE_INCLUDE) -c $< -o $@
>
> idl_test.so: idl_test.o $(IDLDIR)/idl_export.h
> g++ $(CPPFLAGS) -shared -o idl_test.so idl_test.o -lstdc++
>
> clean:
> rm -f *.o *.so
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
