FreeTDS Development Group

["James K. Lowden" <jklowden AT freetds.org>]

Florian Kornhoffer wrote:
> 
> executing the php5 script shows this output on command line:
> Starte NeuesDatum:<br>SELECT: SELECT DATEFORMAT(AddArbeitstage2Datum(
> '2008-11-24', 1, 'KP623', 1, 'D', '80333' ), 'dd.mm.yyyy')<br>Anzahl
> Ergebnisse: 1<br><pre>string(10) "25.11.2008"
> </pre>Speicherzugriffsfehler
>  
> 'Speicherzugriffsfehler' can be transleted with 'memory access failure'.
> This error does not appear when called with php4 - here the script runs
> til the end.

I would look more carefully into the PHP layer.  I don't see any
indication of a problem in the logs.  The pointers seem reasonable (and
similar between the logs) and no error is produced by the server or the
library.  

I don't mean the error is with PHP.  These problems are usually a matter
of interpretation: PHP expects behavior X and FreeTDS exhibits behavior Y,
which may or may not be the same as Sybase's ct-lib, and may or may not be
allowable according to the specs.  

But the discussion has to be at the level of, "we call this function with
these arguments and do not expect FreeTDS to return that value".  I would
be happy to correct ct-lib if a bug can be isolated.  

You may learn more from a backtrace by running your script inside gdb.  

HTH.  

--jkl