Re: wrong documentation and others ..... - Mailing list pgsql-interfaces

Hola Gabi,I have an idea.  I've read in some Java Usenet groups that class
files are not as Write-Once-Read-Everywhere as Sun would like us
to believe.  If you are using one of the Jar files provided by Peter Mount
it might be a good idea to compile the JDBC class from source.Compiling on your systems could help in two ways.  The
compiler
could spot the problem e.g. a name collison.  Also, the solaris javac might
produce class files which do not cause the error in your solaris JVM.Another thought, is there any chance you have an
oldversion of
 
the postgresql JDBC drivers in the classpath of your solaris system?
I had an annoying bug once myself when an old version of a library
was in my classpath before the proper version, so the old-buggy
version was the library being used.
Hasta Luego,Dave

On Wed, 11 Oct 2000 09:52:28 +0200, you wrote:
> David Huttleston Jr escribi�:
> 
> > Hola Gaberiel,
> >         I have a some questions to try to help track down your problem.
> >
> 
>     Hola David. Thanks to reponse. I will try to related my problem in detail.
> 
> >
> > 1) Which program of mine did you run?  Was is the shelf I sent first, or the
> > better cleaned up version I posted as 'An Example of Large Objects and JDBC'?
> 
>             The second. The only change is to add a "for" sentence.
> 
>     for (int i = 0; i < 50; i++) {
>       // build an sample object to save on the shelf
>       ArrayList stuff = new ArrayList();
>      ........
>      Integer key = new Integer(i);
>       ...........
>      }
> 
>         Even the original code, when is running serveral times also fail.
> 
> > 2) What version of the postgresql JDBC driver are you using?
> 
>              I'm using postgresql-7.0.2 with the original jdbc drivers in the
> distribution. This have installed the jdbc6.3.tar.gz patch. I'm not installed ANY patch
> or others drivers.
> 
> >
> > 3) What version of JVM 1.2?
> 
>             As I say in the previous message in Linux the code work fine, but it fail
> in Solaris. I'm using
>             SunOS myhost 5.7 Generic_106541-02 sun4u sparc SUNW,Ultra-5_10
>             And yes, I'm using jvm1.2
> 
> >
> > 4) Are you using a JIT?
> 
>             NO
> 
> >
> > 5) How are you running the program?  (Stand-alone, as a Servlet, ...)
> 
>             Your code in running stand-alone.
> 
> > 6) When the program fails, can you print a stacktrace?
> 
>     Whit the above interation:
> 
> bash-2.03# java TestShelf
>  The second item within stuff: my 2nd thing
>  The second item within stuff: my 2nd thing
>  The second item within stuff: my 2nd thing
> FATAL 1:  my bits moved right off the end of the world!
>         Recreate index pg_attribute_relid_attnam_index.
> FastPath call returned FATAL 1:  my bits moved right off the end of the world!
>         Recreate index pg_attribute_relid_attnam_index.
> 
>         at org.postgresql.fastpath.Fastpath.fastpath(Compiled Code)
>         at org.postgresql.fastpath.Fastpath.fastpath(Compiled Code)
>         at org.postgresql.fastpath.Fastpath.getInteger(Compiled Code)
>         at org.postgresql.largeobject.LargeObjectManager.create(Compiled Code)
>         at org.postgresql.jdbc2.PreparedStatement.setBytes(Compiled Code)
>         at Shelf.putObject(Compiled Code)
>         at TestShelf.main(Compiled Code)
> 
> >
> >
> > Problems that happen after a few executions, in-my-experience, usually happen
> > when resources (like connections) are not being released correctly.  Hopefully
> > we can track down why if we know your enviroment better.
> >
> 
>         I think that is not a problem of code. It seem that is a problem with solaris
> 
> >
> > Adios,
> >         David
> >
> 
>     Thanks for your time, Gabi.
> 
> --
> Gabriel L�pez Mill�n
> Facultad de Inform�tica -Universidad de Murcia
> 30001 Murcia - Espa�a (Spain)
> Telf: +34-968-364644 E-mail: gabilm@dif.um.es
> 
> 
>