SQLAnywhere client and server v. 10.0.1.3976

We recently had a client update before the server had the
changes they needed, and started getting this error (which
is expected for what happened) in the mlsrv log:

E. 2010-01-27 10:35:05. <116148> [-10280] The number of
columns in client table 'empl_timeclock' does not agree with
the number defined in ml_column
E. 2010-01-27 10:35:05. <116148> [-10058] Unable to open
upload_insert
I. 2010-01-27 10:35:10. <116148> ROLLBACK Transaction:
end_upload
I. 2010-01-27 10:35:10. <116148> ROLLBACK Transaction:
end_synchronization
I. 2010-01-27 10:35:10. <116148> COMMIT Transaction:
end_connection
I. 2010-01-27 10:35:10. <116148> Disconnected from
consolidated database
I. 2010-01-27 10:35:11. <116148> Synchronization failed


I have both report_error and report_odbc_error defined as
connection scripts. They call a stored proc that takes the
error and inserts it into a table, which fires a trigger
that sends us a formatted email. We do not have handle_error
or handle_odbc_error defined.

This behavior confuses me. Why didn't either report_error
or report_odbc_error pass us the [-10280] error we saw in
the log? Is this another type of error altogether that we
need to catch?