Procdef manager

Procdef manager

[ download ] [ xml source ] [ discussion ]
The procdef manager manages (and edits) process definitions. For the time being, I'm only implementing a local filesystem procdef repository -- but the procdef manager will almost certainly be the basis for a remote-serving repository.

The procdef repository, then, is a directory. Process definitions are located in XML files stored in that directory; the filename of each is of the form _.xml. So for example, version 1 of the "chair" process is in chair_1.xml.

Versions are related to one another via an item file, which is likewise an XML file, but is named .xml with no version number. Note the possibility of a naming collision if your identifier has an underscore in it. I'm not even going to worry about that right now.

Finally, the entire procdef repository is catalogued with the file index.xml, which lists all the procdefs in the repository, their current versions, and their owners. At some point, we'll also have to include a permission setup. But all in good time.

Note that the PDM is at once a document management system (in that it provides version control) and an editor.

Table of contents: [##itemlist##]

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. This presentation was prepared using LPML.
Project name: [##label##]

[##label##]

[Previous: [##prevlabel##]] [Top: [##indexlabel##]] [Next: [##nextlabel##]]
[##body##]
[Previous: [##prevlabel##]] [Top: lpml alpha] [Next: [##nextlabel##]]

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.
Well, this is the third component of the wftk, so my environment is getting a little more mature. I've moved XML manipulation into a library to cut down on overhead; eventually I'll redo the core engine to use that library instead of simply defining everything right in the same file.

But this program is really simple. It takes a command on the command line, and it returns some HTML representing some facet of the procdef repository (or some individual procdef in the repository.) A separate program will serve to make changes to procdefs and such; this one will just show things. (I think.) #include #include #include "xmlapi.h" #include "localdefs.h" XML * directory; XML * item; XML * version; XML * datasheet; FILE * file; char sbuf[1024]; char * mark; char * format; XML * xml; XML * holder; int main (int argc, char *argv[]) { if (argc #^lt# 2) { printf ("Usage: pdm [[command> [#^lt#args>]\n"); return (1); } directory = NULL; item = NULL; version = NULL; if (!strcmp (argv[1], "list") ) { } else if (!strcmp (argv[1], "starter")) { } else if (!strcmp (argv[1], "datasheet")) { } else if (!strcmp (argv[1], "editor")) { } else { printf ("Unknown command '%s'\n", argv[1]); } if (directory) xml_free (directory); if (item) xml_free (item); if (version) xml_free (version); return (0); } The list is the first one I'm implementing. It simply takes a URL format and builds a set of <li> lines, one for each procdef in the directory. Note that at this moment there's no mechanism for folders, or viewing permissions, or really anything sophisticated at all. sprintf (sbuf, "%s%s", PROCESS_DEFINITION_REPOSITORY, "index.xml"); file = fopen (sbuf, "r"); if (!file) { printf ("Unable to open directory file %s\n", sbuf); return (1); } directory = xml_read (file); fclose (file); if (!directory) { printf ("Corrupt directory file.\n"); return (1); } if (argc > 2) { format = argv[2]; } else { format = "edit?item=%s"; } item = xml_firstelem (directory); while (item) { if (!strcmp (item->name, "item")) { mark = (char *) xml_attrval (item, "title"); if (!*mark) mark = (char *) xml_attrval (item, "id"); printf ("[[li>[[strong>[[a href=\""); printf (format, xml_attrval (item, "id")); printf ("\">%s[[/a>[[/strong>[[br>\n", mark); xml_writecontent (stdout, item); printf ("[[br>\n"); } item = xml_nextelem (item); } The start form generator is pretty straightforward. Given a procdef identifier, we load the procdef item file first and find the current version of the definition. Then we load that definition and get some pertinent information, mostly the data items it needs to get started, but also a title. We return our data on stdout in line-based format; the first line is the title of the process, the second is the current version, and each subsequent line is a data item, formatted as a table row. if (argc < 3) { printf ("Missing procdef identifier in starter command.\n"); return (1); } sprintf (sbuf, "%s%s.xml", PROCESS_DEFINITION_REPOSITORY, argv[2]); file = fopen (sbuf, "r"); if (!file) { printf ("Unable to open procdef file %s.\n", sbuf); return (1); } item = xml_read (file); fclose (file); if (!item) { printf ("Corrupt procdef file %s.\n", sbuf); return (1); } sprintf (sbuf, "%s%s_%s.xml", PROCESS_DEFINITION_REPOSITORY, argv[2], xml_attrval (item, "ver")); file = fopen (sbuf, "r"); if (!file) { printf ("Unable to open procdef version file %s.\n", sbuf); return (1); } version = xml_read (file); fclose (file); if (!version) { printf ("Corrupt version file %s.\n", sbuf); return (1); } if (strcmp (xml_attrval (version, "name"), "")) { printf ("%s\n", xml_attrval (version, "name")); } else { printf ("%s\n", argv[2]); } printf ("%s\n", xml_attrval (item, "ver")); xml = xml_firstelem (version); while (xml) { if (!strcmp (xml->name, "data")) { } else if (!strcmp (xml->name, "sequence")) { break; } else if (!strcmp (xml->name, "parallel")) { break; } else if (!strcmp (xml->name, "task")) { break; } xml = xml_nextelem (xml); } A starter data item is one which we encounter before encountering some action (really before encountering anything that can block.) When we find one, we need to emit a line of HTML table which will format it. How we do that depends on its type. printf ("[[tr>[[td>%s[[/td>\n", xml_attrval (xml, "name")); printf ("[[td>"); if (!strcmp (xml_attrval (xml, "type"), "text")) { printf ("[[textarea name=\"%s\" rows=5 cols=40>", xml_attrval (xml, "name")); xml_writecontent (stdout, xml); printf ("[[/textarea>\n"); } else { printf ("[[input name=\"%s\" value=\"", xml_attrval (xml, "name")); xml_writecontent (stdout, xml); printf ("\">\n"); } printf ("[[/td>[[/tr>\n"); This is basically a straight variation on the starter command. if (argc < 5) { printf ("Missing arguments in datasheet command.\n"); return (1); } sprintf (sbuf, "%s%s_%s.xml", PROCESS_DEFINITION_REPOSITORY, argv[2], argv[3]); file = fopen (sbuf, "r"); if (!file) { printf ("Unable to open procdef version file %s.\n", sbuf); return (1); } version = xml_read (file); fclose (file); if (!version) { printf ("Corrupt version file %s.\n", sbuf); return (1); } datasheet = xml_create( "datasheet"); sprintf (sbuf, "%s_%s.xml", argv[2], argv[3]); xml_set (datasheet, "procdef", sbuf); xml_set (datasheet, "process", argv[4]); xml = xml_firstelem (version); while (xml) { if (!strcmp (xml->name, "data")) { holder = xml_create ("data"); xml_set (holder, "id", xml_attrval (xml, "name")); xml_set (holder, "type", xml_attrval (xml, "type")); xml_append (datasheet, holder); } else if (!strcmp (xml->name, "sequence")) { break; } else if (!strcmp (xml->name, "parallel")) { break; } else if (!strcmp (xml->name, "task")) { break; } xml = xml_nextelem (xml); } xml_write (stdout, datasheet);