The error code is in round brackets i.e. (17) in the CLI; my application reports the same error code.

SQLite version 3.39.4 2022-09-29 15:55:41
Enter ".help" for usage hints.
sqlite> .open ./db/chinook.db
sqlite> select form employees;
Parse error: no such column: form (17)
  select form employees;
         ^--- error here
sqlite>

The documentation: (17) SQLITE_SCHEMA The SQLITE_SCHEMA result code indicates that the database schema has changed.

I am using a free-standing laptop with no other application using SQLite: how can the schema change?

The SQL is badly constructed: I would expect a return code of 1.

Interesting. If you don't open a database you get

SQLite version 3.39.4 2022-09-29 16:56:49
Enter ".help" for usage hints.
Connected to a transient in-memory database.
Use ".open FILENAME" to reopen on a persistent database.
sqlite> select form employees;
Parse error: no such column: form
  select form employees;
         ^--- error here
sqlite> 

Are we sure that the "17" means SQLite error 17, and not, for example, that the error is found at the 17th character in the command, or something like that ?

Are we sure that the "17" means SQLite error 17..

Absolutely sure. In my application I capture the return code and it is definitely 17.

no such column (is logical) ought to be a return code 1 ,generic error (but 17 SQLite_SCHEMA is not , perhaps!).

I don't think you understood the question, which has nothing to do with the error text and everything to do with the return code and why we are seeing:

#define SQLITE_SCHEMA      17   /* The database schema changed */

and not:

#define SQLITE_ERROR        1   /* SQL error or missing database */

which is the normal code for SQL errors.

(I don't have the answer, just wanted to affirm the OP that it's a legitimate question)

The parse error is

Parse error: no such column: form

To me that is

SQLITE_ERROR 1 /* SQL error */

and as sqweek said just wanted to affirm the OP that it's a legitimate question

However, getting SQLITE_SCHEMA instead of SQLITE_ERROR is not earth shattering but confusing nonetheless.

I don't get the problem with the program below. Can you post your patch to shell.c so we can try to reproduce that way?

Dan.

#include <sqlite3.h>
#include <stdio.h>
#include <stdlib.h>

int main(int argc, char **argv){
  const char *zDb = "";
  sqlite3 *db = 0;
  sqlite3_stmt *pStmt = 0;
  int rc = 0;

  if( argc>1 ){
    zDb = argv[1];
  }

  rc = sqlite3_open(zDb, &db);
  if( rc!=SQLITE_OK ){
    fprintf(stderr, "error in open\n");
    exit(-1);
  }

  rc = sqlite3_prepare_v2(db, "SELECT form employee", -1, &pStmt, 0);
  printf("error: (rc=%d), (sqlite3_errcode=%d), (sqlite3_errmsg=%s)\n",
    rc, sqlite3_errcode(db), sqlite3_errmsg(db)
  );
  return 0;
}