subprocess.Popen howto?

[norseman]

Øystein ;
         Down below: change >yourPy.py  to | yourPy.py
I just noticed the typo.

Sorry

Steve

norseman wrote:
> Øystein Johansen (OJOHANS) wrote:
>> Hi,
>>  
>> I have problems understanding the subprocess.Popen object. I have a 
>> iterative calculation in a process running and I want to pipe the 
>> output (stdout) from this calculation to a Python script.
>>  
>> Let me include a simple code that simulates the calculating process:
>> /* This code simulates a big iterative calculation */
>> #include <stdio.h>
>> #include <math.h>
>>  
>> int main()
>> {
>>  float val[2] = { M_PI, M_E };
>>  int i;
>>  
>>  for ( i = 0; i < 2 i++) {
>>   sleep( 15 );   /* It's a hard calculation. It take 15 seconds */
>>   printf("Value: %5.6f\n", val[i] );
>>   fflush( stdout );
>>  }
>>  return 0;
>> }
>>  
>> let's compile this to mycalc: gcc -o mycalc calc.c ... (untested code)
>>  
>> In C I have this code which starts the mycalc process and handles the 
>> output from it:
>>  
>> #include <stdio.h>
>> #include <assert.h>
>> #define BUF_SIZE 256
>>  
>> int main()
>> {
>>  FILE *pip;
>>  char line[BUF_SIZE];
>>  
>>  pip = popen("mycalc", "r");
>>  assert( pip != NULL );
>>  
>>  while ( fgets( line, BUF_SIZE, pip )) {
>>   printf( "Hello; I got: %s \n", line );
>>   fflush( stdout );
>>  }
>>  pclose( pip );
>>  return 0;
>> }
>> How can I make such while-loop in Python? I assume I should use 
>> subprocess.Popen(), but I can't figure out how?
>>  
>> -Øystein
>>  
>>  
>>
>> -------------------------------------------------------------------
>> The information contained in this message may be CONFIDENTIAL and is
>> intended for the addressee only. Any unauthorised use, dissemination 
>> of the
>> information or copying of this message is prohibited. If you are not the
>> addressee, please notify the sender immediately by return e-mail and 
>> delete
>> this message.
>> Thank you.
>>
>>
>> ------------------------------------------------------------------------
>>
>> -- 
>> http://mail.python.org/mailman/listinfo/python-list
> ================================
> 
> If you don't like a lot of typing that obscures the process,
> take a look at os.Popen2  Pg.39 or so in Lib.pdf for 2.5.2
> In this case - the popen3 is probably your best bet.
> 
> I took a test run on "subprocess" a few months ago. My review:
> excessive typing to accomplish the simple.  BlackBox stuff is supposed 
> to be short and sweet.
> 
> BlackBox was the term applied to drivers and couplers back when.
> Back when: one started by writing a subroutine that got used.
>            next one got to write a program
>            next one got to write an application
>            then a BlackBox
>            from there one could graduate to OS
> That was back when!
> 
> 
> to repeat from another of my postings:
>                         ---
> processPython 2.5.2 (r252:60911, Mar  4 2008, 10:40:55)
> [GCC 3.3.6] on linux2
> Type "help", "copyright", "credits" or "license" for more information.
> 
>  >>> import os
>  >>> AbWd = os.spawnlp( os.P_WAIT,"abiword","abiword","")
> 
> The P_WAIT stops python until the program (abiword in this case) 
> completes.  The "" at the end are for tokens to be given to the program 
> and yes - contrary to manual, the program MUST be there TWICE (complete 
> with any path needed).
> 
> for windows this works:
> (for cut and paste from cmd.exe, see:
>   Re: Copy & Paste in a Dos box  from norseman  05/06/2009 4:28PM
> )
> 
> Python 2.5.1 ... on win32
>  >>> import os
>  >>> result = os.spawnl( os.P_WAIT, "d:\\winmcad\\mcad","")
> 
> Runs the program mcad. Returns to python when mcad exits.
>                         ---
> 
> 
> In your case:  substitute ..."your_compiled_c_program", " >yourPy.py")
>                 at the obvious places.
> 
> In case this isn't clear;
>   method 1:  py1.py starts compiled.c which redirects to yourPy.py
>   method 2:  py1.py uses os.popen3(compiled.c...) & the two work it out.
> In either case the receiver at the end stays open until completion so 
> sleep() things are not needed.  (may need to test for EOL and EOT)
> 
> HTH
> 
> Steve
> norseman at hughes.net
> -- 
> http://mail.python.org/mailman/listinfo/python-list
>