FreeTDS Development Group

[Bob Kline <bkline AT rksystems.com>]

On Wed, 21 Feb 2001, Paul Schaap wrote:

> Can someone with CVS access patch the README

While we're patching the README:

> The warning chain is not implemented yet.

I think this sentence can be dropped.  Wasn't this one of the gaps you
plugged, Stefan?

> Query timeouts and the cancel() method are working for SQLServer but
> not for Sybase.

I'm not convinced that this is 100% true even for SQL Server.  For
TDS=4.2, test t0005 passes, with or without logging turned on.  For
TDS=7.0, test t0005 passes only if logging is not enabled.  If logging
is enabled, the test runs (taking much longer than 3 seconds), throws an
exception, reports failure, and exits.  Although the client process --
and its connections -- no longer exist, sp_who shows one of the three
threads which were created for the client to be still running, eating
all available cycles on all the server machine's processors.  Here's the
tail end of the output from the test:

>>>>>>>>
row - row 4 2001-02-22 08:28:34.713
row - row 3 2001-02-22 08:28:34.703
512 rows were fetched
Setting timeout to 0
The timeout is 0
Starting query.  Will cancel in 3000 milliseconds
Performing big query
Results of big query are-
Exception in thread "main" java.lang.NullPointerException
        at t0005.doBigQuery(t0005.java:94)
        at t0005.main(t0005.java:218)
background thread is canceling
The query timeout didn't work.  Killing the test!
Failed t0005.
<<<<<<<<

This is with SQL Servers version 7.0, SP3 (one running on a
single-processor W2K machine @ SP1, another on a quad-processor WinNT
4.0, SP5 box).  I'm not sure what's going on here.

-- 
Bob Kline
mailto:bkline AT rksystems.com
http://www.rksystems.com