A fix on CPAN module MVS::JESFTP

Last week, I tried to do some work through MVS::JESFTP.  Actually, I knew this module for a long time,  just had  an idea to automate the process of submitting JCL job and fetching joblog, so I decided to reuse the module rather than invent a new wheel.  However, firstly I met problem on invoking function wait_for_results while trying out it.  I searched related info online, although some similar questions posted,  still could't find any clue to solve them. No way except dirty my hands.  Source code of   wait_for_results is as below.

 
 

  
  sub 
  
  wait_for_results 
  
  { 
  
  #------------------------------------------------
 
 
 
 
	
  
  my
  
  (
  
  $self
  
  , 
  
  $JOB
  
  , 
  
  $TIMEOUT
  
  ) 
  
  = 
  
  @_
  
  ;
 
 
 
 

  
  

 
 
 
 
	
  
  $JOB 
  
  =~ 
  
  s/\..+$//
  
  ;
 
 
 
 
	
  
  $TIMEOUT 
  
  ||= 
  
  60
  
  ;
 
 
 
 

  
  

 
 
 
 
	
  
  my 
  
  @results 
  
  = 
  
  ();
 
 
 
 
	
  
  my 
  
  $i 
  
  = 
  
  0
  
  ;
 
 
 
 
	
  
  while 
  
  (
  
  ++
  
  $i 
  
  <= 
  
  $TIMEOUT
  
  ) 
  
  {
 
 
 
 
		
  
  # print "$i: waiting for $JOB...\n";
 
 
 
 
		
  
  last 
  
  if 
  
  (
  
  @results 
  
  = 
  
  grep 
  
  /^$JOB\s+JOB\d+\s+OUTPUT/
  
  , 
  
  $self
  
  ->
  
  dir
  
  );  
 
 
 
 
		
  
  sleep
  
  (
  
  1
  
  );
 
 
 
 
	
  
  }
 
 
 
 
	
  
  return 
  
  (
  
  @results
  
  ) 
  
  ? 
  
  \
  
  @results 
  
  : 
  
  undef
  
  ;
 
 
 
 

  
  } 
  
  #---------------------------------------------------------------------
 
 

By investigating the code, found that culprit is regexp for grep "^$JOB\s+JOB\d+\s+OUTPUT" .  It looks like good, unfortunately, it can't work on my system which issues following job info. 

         JOBFTP   JOB00254 THINKHY OUTPUT A        RC=0256 4 spool files

The third segment is username but not  OUTPUT as the regexp shows.  What I need to do next is very simple,  to relax the match condition and fix a line as below.

         Before:  

lastif(@results=grep/^$JOB\s+JOB\d+\s+OUTPUT/,$self->dir);

         After:      

lastif(@results=grep/^$JOB\s+JOB\d+/i,$self->dir);          

In addition, to ensure the code can work in various environment, add two lines to specify jesstatus and jesjobname.

 
 
	
  
  $self
  
  ->
  
  quot
  
  (
  
  'SITE'
  
  , 
  
  'jesstatus=all'
  
  ) 
  
  or 
  
  return 
  
  undef
  
  ;
 
 
 
 
	
  
  $self
  
  ->
  
  quot
  
  (
  
  'SITE'
  
  , 
  
  "jesjobname=$JOB"
  
  ) 
  
  or 
  
  return 
  
  undef
  
  ;
 
 

Now,  the code of wait_for_results is 

   

 
 

  
  sub 
  
  wait_for_results 
  
  { 
  
  #------------------------------------------------
 
 
 
 
	
  
  my
  
  (
  
  $self
  
  , 
  
  $JOB
  
  , 
  
  $TIMEOUT
  
  ) 
  
  = 
  
  @_
  
  ;
 
 
 
 

  
      
 
 
 
 
	
  
  $JOB 
  
  =~ 
  
  s/\..+$//
  
  ;
 
 
 
 
	
  
  $TIMEOUT 
  
  ||= 
  
  60
  
  ;
 
 
 
 

  
  

 
 
 
 
	
  
  my 
  
  @results 
  
  = 
  
  ();
 
 
 
 

  
  

 
 
 
 
	
  
  $self
  
  ->
  
  quot
  
  (
  
  'SITE'
  
  , 
  
  'jesstatus=all'
  
  ) 
  
  or 
  
  return 
  
  undef
  
  ;
 
 
 
 
	
  
  $self
  
  ->
  
  quot
  
  (
  
  'SITE'
  
  , 
  
  "jesjobname=$JOB"
  
  ) 
  
  or 
  
  return 
  
  undef
  
  ;
 
 
 
 

  
  

 
 
 
 
	
  
  my 
  
  $i 
  
  = 
  
  0
  
  ;
 
 
 
 
	
  
  while 
  
  (
  
  ++
  
  $i 
  
  <= 
  
  $TIMEOUT
  
  ) 
  
  {
 
 
 
 
		
  
  # print "$i: waiting for $JOB...\n";
 
 
 
 

  
          
  
  #last if (@results = grep /^$JOB\s+JOB\d+.*?\s+OUTPUT/, $self->dir);
 
 
 
 

  
  

 
 
 
 

  
          
  
  # Fix the regexp, relax the match condition. [6/9/2012 thinkhy]
 
 
 
 
		
  
  last 
  
  if 
  
  (
  
  @results 
  
  = 
  
  grep 
  
  /^$JOB\s+JOB\d+/i
  
  , 
  
  $self
  
  ->
  
  dir
  
  );  
 
 
 
 

  
  

 
 
 
 
		
  
  sleep
  
  (
  
  1
  
  );
 
 
 
 
	
  
  }
 
 
 
 

  
  

 
 
 
 
	
  
  return 
  
  (
  
  @results
  
  ) 
  
  ? 
  
  \
  
  @results 
  
  : 
  
  undef
  
  ;
 
 
 
 

  
  

 
 
 
 

  
  } 
  
  #---------------------------------------------------------------------
 
 

Finally this CPAN works perfectly on my system. Anyway, it is truly appreciated that Mike Owens implemented the useful module for us.  Hope my work is also helpful for you. Please contact me(think.hy@gmail.com)  freely if any feedback.

Reference

       1  Submit batch jobs from Java on z/OS

            http://www.ibm.com/developerworks/systems/library/es-batch-zos.html

       2 Access z/OS batch jobs from Java 
            http://www.ibm.com/developerworks/systems/library/es-zosbatchjavav/index.html

6/11/2012 Update

   

 In the regular expression lastif(@results=grep/^$JOB\s+JOB\d+/i,$self->dir);  ,  I  missed the OUTPUT segment that would lead to error for retrieval of  data stream when job is in active state.  Modified the regular expression as following.

        @results=grep/^$JOB\s+JOB\d+.*?OUTPUT/i,@tmp_results;

     Another thing is about multiple jobs with the same job name, I think only If all these jobs are finished(in OUTPUT state) , we can download joblogs.   So I added some code in function wait_for_results to ensure job logs can be downloaded in a correct way.

 
 

  
  sub wait_for_results { #------------------------------------------------
 
 
 
 

  
  	my($self, $JOB, $TIMEOUT) = @_;
 
 
 
 

  
      
 
 
 
 

  
  	$JOB =~ s/\..+$//;
 
 
 
 

  
  	$TIMEOUT ||= 60 * 60 * 24; # Default timeout value
 
 
 
 

  
  

 
 
 
 

  
  	my @results = ();
 
 
 
 

  
  

 
 
 
 

  
  	$self->quot('SITE', 'jesstatus=all') or return undef;
 
 
 
 

  
  	$self->quot('SITE', "jesjobname=$JOB") or return undef;
 
 
 
 

  
  

 
 
 
 

  
  	my $i = 0;
 
 
 
 

  
  	while (++$i <= $TIMEOUT) {
 
 
 
 

  
  	    print "$i: waiting for $JOB...\n";
 
 
 
 

  
  

 
 
 
 

  
          # Fix the regexp, relax the match condition. [6/9/2012 thinkhy]
 
 
 
 

  
          my @tmp_results = $self->dir;
 
 
 
 

  
  

 
 
 
 

  
          shift @tmp_results; # Remove title row
 
 
 
 

  
          #print join "\n", @tmp_results;
 
 
 
 

  
  

 
 
 
 

  
          next unless(@tmp_results); # If job log is unavaliable.
 
 
 
 

  
  

 
 
 
 

  
          @results = grep /^$JOB\s+JOB\d+.*?OUTPUT/i, @tmp_results;
 
 
 
 

  
  

 
 
 
 

  
          # If all the jobs are in OUTPUT state , that way we can download joblog.
 
 
 
 

  
          last if ($#tmp_results == $#results);
 
 
 
 

  
  

 
 
 
 

  
  		sleep(1);
 
 
 
 

  
  	}
 
 
 
 

  
  

 
 
 
 

  
  	return (@results) ? \@results : undef;
 
 
 
 

  
  

 
 
 
 

  
  } #---------------------------------------------------------------------
 
 

 Complete code of my jesftp.pm has pasted to github: https://gist.github.com/2900792