Dealing with processes

[ Previous: main() ] [ Top:  ] [ Next: Working with tasks ]
What can we do with processes and datasheets? Create them and delete them, define a process by attaching a procdef, and that's about it.

create: Creating a new datasheet
Creation of a process is almost too easy.
  
datasheet = wftk_process_new (session, argsleft > 0 ? PERIOD_TO_NULL(argv[argp]) : (char *) 0, argsleft > 1 ? PERIOD_TO_NULL(argv[argp+1]) : (char *) 0);  argp++; argp++;

if (datasheet) {
   wftk_process_save (session, datasheet);
   printf ("Created process %s\n", xml_attrval (datasheet, "id"));
}
A caveat: if I specify an ID which is already taken, I'll get another (but unique) ID, because the semantics of the wftk library consider the specified ID as a request, not a demand. Obviously, you can code around that if you're using the library yourself, but the command-line utility is a very thin and simple wrapper, so it just goes with the flow.

delete: Deleting an existing process
Here's one place where I start to ask, where should I refuse to do things? At what point can a process say, "No, I'm active, you can't delete me?" I'm not sure yet. Another bridge to burn when I get to it. In the meantime, this is a good way to clean up after testing. But somewhere, either here or in wftk_process_delete, we really need to look aside at a defined deletion process instead of deleting willy-nilly.
  
if (argsleft < 2) {
   printf ("wftk delete: repository and ID required (use . for default repository)\n");
   exit(1);
}

wftk_process_delete (session, argsleft > 0 ? PERIOD_TO_NULL(argv[argp]) : (char *) 0, argv[argp+1]); argp++; argp++;


define: Attach a procdef
Defining a process means attaching a process definition to it. The procdef is stored in a procdef repository, and the current version is always attached to new processes. A process will thus always work from the same version of the procdef; if new versions are created, they apply only to newly created processes. Otherwise havoc would ensue.
  
if (argsleft < 4) {
   printf ("wftk define: dsrep and ID, pdrep and ID are all required (use . for default repositories)\n");
   exit (1);
}

datasheet = wftk_process_load (session, PERIOD_TO_NULL(argv[argp]), argv[argp+1]); argp++; argp++;
if (!datasheet) {
   printf ("Datasheet repository %s can't find datasheet %s", argv[argp-2], argv[argp-1]);
} else {
   if (wftk_process_define (session, datasheet, PERIOD_TO_NULL (argv[argp]), argv[argp+1])) {
      wftk_process_save (session, datasheet);
      printf ("Process started\n");
   }
}




start: Start a procdef

Starting is just like defining, except that (assuming all data requirements have been met) the workflow is started immediately instead
of setting up a start task.



 

if (argsleft < 4) {
   printf ("wftk start: dsrep and ID, pdrep and ID are all required (use . for default repositories)\n");
   exit (1);
}

datasheet = wftk_process_load (session, PERIOD_TO_NULL(argv[argp]), argv[argp+1]); argp++; argp++;
if (!datasheet) {
   printf ("Datasheet repository %s can't find datasheet %s", argv[argp-2], argv[argp-1]);
} else {
   if (wftk_process_start (session, datasheet, PERIOD_TO_NULL (argv[argp]), argv[argp+1])) {
      wftk_process_save (session, datasheet);
      printf ("Process started\n");
   }
}




show: Displaying process information





adhoc: Running ad-hoc workflow

Ad-hoc workflow is something that most workflow systems handle only begrudgingly.  I'm finding that it's a convenient way
to handle a lot of things, though, like requests, error situations, and so on.  Basically, the adhoc command
just reads a file or stdin, and routes the XML there to wftk_process_adhoc, which tosses it into the datasheet
and starts it up.  The ad-hoc code is really ad-hoc, that is, any valid workflow structure can be in it, and once it's 
queued, we really don't care where it resides.




 

if (argsleft < 2) {
   printf ("wftk adhoc: dsrep and ID are required.  Use . for default repository\n");
   exit (1);
}

datasheet = wftk_process_load (session, PERIOD_TO_NULL(argv[argp]), argv[argp+1]); argp++; argp++;
if (!datasheet) {
   printf ("Datasheet repository %s can't find datasheet %s", argv[argp-2], argv[argp-1]);
} else {
   file = stdin;
   if (argsleft > 0) {
      file = fopen (argv[argp], "r");
      if (!file) {
         printf ("Can't open ad-hoc code file %s\n", argv[argp]);
      }
   }
   if (file) {
      mark = xml_read (file);
      if (!mark) {
         printf ("Can't read XML in ad-hoc input file %s\n", argv[argp]);
      }
      if (file != stdin) fclose (file);
      if (mark) {
         wftk_process_adhoc (session, datasheet, mark);
      }
   }
}







[ Previous: main() ]
[ Top:  ]
[ Next: Working with tasks ]





This code and documentation are released under the terms of the GNU license.  They are
additionally copyright (c) 2000, Vivtek.  All rights reserved except those explicitly
granted under the terms of the GNU license.