Netscape 2.0 bus error

I'm getting a bus error from a module I'm developing using the Oracle Call Interface to access data from an NSAPI plug-in. I developed it under Netscape Enterprise 3.0 on HP/UX 10.20, and now for various reasons need to migrate downward to work with Netscape Enterprise 2.01.

The application works fine under 3.x, but when I try to run it under 2.x, it gives me a bus error while in an initialization routine. I have done my best to make the routines thread-safe, but still am facing a potential bus error. Here's the offending routine:

/***************************************************************

• int sm_data_load_grouplist(Lda_Def *lda, char* err_msg)
  *
  
• Loads group data from the database into our security system.
  *
  ***************************************************************/
  int sm_data_load_grouplist(Lda_Def *lda, char *err_msg) { int id; text *name; int cont = 1; sword rv = 0;

  DB_SAY("sm_data_load_grouplist", "Started");

  /* Allocate memory for the name variable */   name = malloc( GROUP_MAX_LENGTH * sizeof(text) );

  /* Bind to the variables */
  if (! odef(lda, get_grouplist_curs, 1, &id, sizeof(int), SQLT_INT, err_msg) )

    return FALSE;
  if (! odef(lda, get_grouplist_curs, 2, name, GROUP_MAX_LENGTH, SQLT_STR, err_msg) )

    return FALSE;

  DB_SAY("sm_data_load_grouplist", "Bound variables");

  /* Execute the query */
  if (rv = oexec(get_grouplist_curs)) {
    DB_SAY("sm_data_load_grouplist", "Error");     get_error(err_msg, rv, lda, __FILE__, __LINE__);     return FALSE;

  }

  DB_SAY("sm_data_load_grouplist", "Executed query");

• cut -------

From this, I get the following messages:

[24/Mar/1998:16:48:46] info: sm_data_load_grouplist reports: Started [24/Mar/1998:16:48:46] info: sm_data_load_grouplist reports: Bound variables
/home/steven/ns-home/https-test-serv/start[3]: 12191 Bus error(coredump)

DB_SAY is a debug macro that writes to the log file; it's just a wrapper over log_error. odef() is a wrapper over obindps. The routine is run under thread lock, so I don't see how it could be a thread safety issue. I've set up 2.0 MaxThreads 1 MinThreads 1-- it seems to make no difference. My questions are-- what could the problem be? Why does it work under 3.0 and not 2.0? How can I debug this further?

Thanks in advance...

Steve Nelson
steven_at_jubal.com Received on Tue Mar 24 1998 - 00:00:00 CST