sqlrelay-discussion — Discussion of topics related to SQL Relay

[Sqlrelay-discussion] Multiple connection

[Guo, J. <Jin...@pa...>]

David,

I want to set up the sqlrelay connect to the database with 4 connections =
initially and grow by 2 until 10 connections reached. After startup, I =
can only see one sqlr-connection process running, does that mean I have =
only one connection available? Could you please check my conf file and =
tell me what is wrong?

<?xml version=3D"1.0"?>
<!DOCTYPE instances SYSTEM "sqlrelay.dtd">

<instances>

        <instance id=3D"example" port=3D"5559" =
socket=3D"/usr/local/firstworks/var/sqlrelay/tmp/socket.file" =
dbase=3D"oracle7" connections=3D"4" maxconnections=3D"10" =
maxqueuelength=3D"5" growby=3D"2" ttl=3D"60" endofsession=3D"commit" =
sessiontimeout=3D"600" runasuser=3D"nobody" runasgroup=3D"nobody" =
cursors=3D"2" authtier=3D"listener" handoff=3D"pass" deniedips=3D"" =
allowedips=3D"" debug=3D"none">
                <users>
                        <user user=3D"" password=3D""/>
                </users>
                <connections>
                        <connection connectionid=3D"db1" =
string=3D"user=3Dabc;password=3Ddef;oracle_sid=3Dmydb.world;" =
metric=3D"1"/>
                </connections>
        </instance>

</instances>

Re: [Sqlrelay-discussion] query timeout

[Firstworks/4access <dav...@fi...>]

On Fri, 2005-07-15 at 09:52 -0400, simon martin wrote:
> Hi all,
> 
> One of my customers uses postgresql and I am using sqlrelay for
> connection pooling. This is working very well and has been for quite
> some time now. Unfortunately they are having some problems with the
> database and queries that used to take a copule of seconds, now take a
> couple of minutes. They are looking into the problem, but they want me
> to add a 2 second timeout to some of the time critical queries that
> are run (some external clients will only wait 10 seconds for a
> response to a set of queries).
> 
> The only way I can see of implementing this is using the
> STATEMENT_TIMEOUT command. Is there any way to specify a query timeout
> in sqlrelay?
> 
> TIA
> 

There's no good way (yet).  Some database API's (including postgresql's)
provide functions for running asynchronous queries.  You can start a
query, let it run in the background, check on it's progress, and
possibly cancel it.  Using an interface like that is the simplest way to
implement a timeout.  However, since many databases API's don't provide
functions for running asynchronous queries, I never got around to
implementing methods for asynchronous queries in SQL Relay's API.  In
fact, I'm not sure that there even is a safe way to do them if the
database's native API doesn't support them.  At any rate, "api for
running queries in the background" is on the TODO list, I'll update it
with a note about being able to run a query with a timeout, but I'm not
certain when I'll get it implemented.

If it's a feature you really need soon, please consider taking out a
custom development contract (see
http://sqlrelay.sourceforge.net/support.html).  I place much higher
priority on bugs and features requested by support and custom
development contract holders than on regular TODO list items.

David Muse
dav...@fi...

[Sqlrelay-discussion] query timeout

[simon m. <sm...@mi...>]

Hi all,

One of my customers uses postgresql and I am using sqlrelay for
connection pooling. This is working very well and has been for quite
some time now. Unfortunately they are having some problems with the
database and queries that used to take a copule of seconds, now take a
couple of minutes. They are looking into the problem, but they want me
to add a 2 second timeout to some of the time critical queries that
are run (some external clients will only wait 10 seconds for a
response to a set of queries).

The only way I can see of implementing this is using the
STATEMENT_TIMEOUT command. Is there any way to specify a query timeout
in sqlrelay?

TIA

--=20
    __ _   Debian GNU User   Simon Martin
   / /(_)_ __  _   ___  __   Project Manager
  / / | | '_ \| | | \ \/ /   Milliways
 / /__| | | | | |_| |>  <    mailto: sm...@mi...
 \____/_|_| |_|\__,_/_/\_\   ICQ: 81183862

Re: [Sqlrelay-discussion] sqlrelay-0.36.4 is out

[vieonet f. <df...@vi...>]

Great tx a lot 


----- Original Message ----- 
From: "David Muse" <dav...@fi...>
To: <sql...@li...>
Sent: Friday, July 15, 2005 3:35 AM
Subject: [Sqlrelay-discussion] sqlrelay-0.36.4 is out


> Hello All,
> 
> sqlrelay-0.36.4 is out with a short list of important fixes:
> 
> * fixed const error in mdbtools connection
> * added const to *connection::identify()/pingQuery() methods
> * improved efficency of sending lob's (a lot)
> * added Py_INCREF(Py_None) before returning Py_None's or adding them to
> an array
> * fixed php pear db bug that caused NULL values not to be inserted
> * fixed a bug where a null clob could cause crashes
> 
> If any of these fixes are interesting to you, it's upgrade time!  I'm
> using Fedora Core 4 now, so I made some FC4 RPM's too.
> 
> Give it a try, report any bugs you find.
> 
> David Muse
> dav...@fi...
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
> from IBM. Find simple to follow Roadmaps, straightforward articles,
> informative Webcasts and more! Get everything you need to get up to
> speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>

[Sqlrelay-discussion] sqlrelay-0.36.4 is out

[David M. <dav...@fi...>]

Hello All,

sqlrelay-0.36.4 is out with a short list of important fixes:

* fixed const error in mdbtools connection
* added const to *connection::identify()/pingQuery() methods
* improved efficency of sending lob's (a lot)
* added Py_INCREF(Py_None) before returning Py_None's or adding them to
an array
* fixed php pear db bug that caused NULL values not to be inserted
* fixed a bug where a null clob could cause crashes

If any of these fixes are interesting to you, it's upgrade time!  I'm
using Fedora Core 4 now, so I made some FC4 RPM's too.

Give it a try, report any bugs you find.

David Muse
dav...@fi...

RE: [Sqlrelay-discussion] using sql relay with a specific ODBCdriver

[Jeremie L. <jer...@ko...>]

Thanks for your answer.=20

There is no need for a user/password in the connect string. It works =
well, when we try to connect with the command line.

We used the version 0.36.3 compiled from sources. For that, we needed =
the sources of UnixODBC 2.2.11.

We will try to compile again on RedHat Enterprise Linux 4 ES. If it =
doesn't work, we will try with a MySQL database instead of our specific =
driver.

I'm very interested in your 'extra debug' file. It could help :)


	J=E9r=E9mie LEGRAND





-----Message d'origine-----
De=A0: sql...@li... =
[mailto:sql...@li...] De la part de =
Firstworks/4access
Envoy=E9=A0: lundi 11 juillet 2005 19:38
=C0=A0: sql...@li...
Objet=A0: Re: [Sqlrelay-discussion] using sql relay with a specific =
ODBCdriver

Everything looks good except, I noticed your connectstring is:

user=3D;password=3D;dsn=3DTolas;autocommit=3Dyes

Did you just leave out the user and password in this email or are they
really empty in the file?  If they are really empty, that could be the
problem.  Does the database not require a user/password?

It's possible that there is a bug too.  What version of SQL Relay are
you using?  Did you build from source?  I can send you a file with some
extra debug in it that will probably give us enough information to
figure out what's going on, but you'll have to rebuild SQL Relay from
source.  Is that ok?

David Muse
dav...@fi...

On Fri, 2005-07-08 at 15:22 +0200, Jeremie LEGRAND wrote:
> Hello,
>=20
> =20
>=20
>             I would like to know if it is possible to use sqlrelay
> with a specific odbc driver made for one of our servers.
>=20
> The situation is :
>=20
> =20
>=20
> - I have an Alpha with OpenVMS on it
>=20
> - I have a Redhat server, with a specific ODBC driver to connect to
> the Alpha
>=20
> - I have a Debian server, with Apache and PHP
>=20
> =20
>=20
> --> I try to make an ODBC link between RedHat and Debian, in order to
> use databases of the Alpha on PHP (Debian).
>=20
> =20
>=20
> I can't directly connect Debian on the Alpha : the specific ODBC
> driver is only supported on RedHat...
>=20
> For information, the ODBC driver is made by Attunity.
>=20
> =20
>=20
> =20
>=20
> =20
>=20
> Our ODBC driver himself works. We could use it on PHP pages as well.
>=20
> =20
>=20
> --> We defined our odbc.ini as follows :
>=20
> [Tolas]
>=20
> Driver =3D Connect to VMS
>=20
> Driver =3D /usr/local/navroot/lib/navcli.so
>=20
> =20
>=20
> =20
>=20
> Here is the config file
>=20
> =20
>=20
> =
-------------------------------------------------------------------------=
-
>=20
> <?xml version=3D"1.0"?>
>=20
> <instances>
>=20
>             <instance id=3D"toto" port=3D"9000"
> unixport=3D"/tmp/sqlrelay.socket" dbase=3D"odbc" connections=3D"1"
> maxconnections=3D"1" maxqueuelength=3D"1" growby=3D"1" ttl=3D"60"
> endofsession=3D"rollback" sessiontimeout=3D"60" runasuser=3D"root"
> runasgroup=3D"root" cursors=3D"5" authtier=3D"listener" =
handoff=3D"reconnect"
> deniedips=3D"" allowedips=3D"" debug=3D"none">
>=20
>                         <users>
>=20
>                                    <user user=3D"titi" =
password=3D"titi"/>
>=20
>                         </users>
>=20
>                         <connections>
>=20
>                                    <connection connectionid=3D"Tolas"
> string=3D"user=3D;password=3D;dsn=3DTolas;autocommit=3Dyes" =
metric=3D"1"/>
>=20
>                         </connections>
>=20
>             </instance>
>=20
> </instances>
>=20
> =
-------------------------------------------------------------------------=
-
>=20
> =20
>=20
> =20
>=20
> =20
>=20
> =20
>=20
> The problem is that the configuration seems correct, but when we start
> sqlrelay, we got the following message :
>=20
> =20
>=20
> =20
>=20
> =
-------------------------------------------------------------------------=
-
>=20
> Starting listener:
>=20
>   sqlr-listener -id toto -config /root/toto
>=20
> =20
>=20
> Starting 1 connections to Tolas :
>=20
>   sqlr-connection-odbc -id toto -connectionid Tolas -config /root/toto
>=20
> Couldn't log into database.
>=20
> =20
>=20
> sqlr-connection-odbc failed to start.
>=20
> =20
>=20
> =20
>=20
> Thanks to MP3.com for sponsoring:
>=20
>         Clustered/Replicated database support.
>=20
>         Perl API.
>=20
> =
-------------------------------------------------------------------------=
-
>=20
> =20
>=20
> =20
>=20
> Thank you for any suggestion.
>=20
>=20



-------------------------------------------------------
This SF.Net email is sponsored by the 'Do More With Dual!' webinar =
happening
July 14 at 8am PDT/11am EDT. We invite you to explore the latest in dual
core and dual graphics technology at this free one hour event hosted by =
HP,
AMD, and NVIDIA.  To register visit http://www.hp.com/go/dualwebinar
_______________________________________________
Sqlrelay-discussion mailing list
Sql...@li...
https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion

[Sqlrelay-discussion] 0.36.4pre2

[David M. <dav...@fi...>]

Hello everyone,

There's a new prerelease available at
http://www.firstworks.com/sqlrelay-0.36.4pre2.tar.gz

The 0.36.4 release will be out as soon as I'm sure that I actually fixed
the bugs that I intended to fix in this release, which should be pretty
soon :)

Dave
dav...@fi...

Re: [Sqlrelay-discussion] using sql relay with a specific ODBC driver

[Firstworks/4access <dav...@fi...>]

Everything looks good except, I noticed your connectstring is:

user=;password=;dsn=Tolas;autocommit=yes

Did you just leave out the user and password in this email or are they
really empty in the file?  If they are really empty, that could be the
problem.  Does the database not require a user/password?

It's possible that there is a bug too.  What version of SQL Relay are
you using?  Did you build from source?  I can send you a file with some
extra debug in it that will probably give us enough information to
figure out what's going on, but you'll have to rebuild SQL Relay from
source.  Is that ok?

David Muse
dav...@fi...

On Fri, 2005-07-08 at 15:22 +0200, Jeremie LEGRAND wrote:
> Hello,
> 
>  
> 
>             I would like to know if it is possible to use sqlrelay
> with a specific odbc driver made for one of our servers.
> 
> The situation is :
> 
>  
> 
> - I have an Alpha with OpenVMS on it
> 
> - I have a Redhat server, with a specific ODBC driver to connect to
> the Alpha
> 
> - I have a Debian server, with Apache and PHP
> 
>  
> 
> --> I try to make an ODBC link between RedHat and Debian, in order to
> use databases of the Alpha on PHP (Debian).
> 
>  
> 
> I can't directly connect Debian on the Alpha : the specific ODBC
> driver is only supported on RedHat...
> 
> For information, the ODBC driver is made by Attunity.
> 
>  
> 
>  
> 
>  
> 
> Our ODBC driver himself works. We could use it on PHP pages as well.
> 
>  
> 
> --> We defined our odbc.ini as follows :
> 
> [Tolas]
> 
> Driver = Connect to VMS
> 
> Driver = /usr/local/navroot/lib/navcli.so
> 
>  
> 
>  
> 
> Here is the config file
> 
>  
> 
> --------------------------------------------------------------------------
> 
> <?xml version="1.0"?>
> 
> <instances>
> 
>             <instance id="toto" port="9000"
> unixport="/tmp/sqlrelay.socket" dbase="odbc" connections="1"
> maxconnections="1" maxqueuelength="1" growby="1" ttl="60"
> endofsession="rollback" sessiontimeout="60" runasuser="root"
> runasgroup="root" cursors="5" authtier="listener" handoff="reconnect"
> deniedips="" allowedips="" debug="none">
> 
>                         <users>
> 
>                                    <user user="titi" password="titi"/>
> 
>                         </users>
> 
>                         <connections>
> 
>                                    <connection connectionid="Tolas"
> string="user=;password=;dsn=Tolas;autocommit=yes" metric="1"/>
> 
>                         </connections>
> 
>             </instance>
> 
> </instances>
> 
> --------------------------------------------------------------------------
> 
>  
> 
>  
> 
>  
> 
>  
> 
> The problem is that the configuration seems correct, but when we start
> sqlrelay, we got the following message :
> 
>  
> 
>  
> 
> --------------------------------------------------------------------------
> 
> Starting listener:
> 
>   sqlr-listener -id toto -config /root/toto
> 
>  
> 
> Starting 1 connections to Tolas :
> 
>   sqlr-connection-odbc -id toto -connectionid Tolas -config /root/toto
> 
> Couldn't log into database.
> 
>  
> 
> sqlr-connection-odbc failed to start.
> 
>  
> 
>  
> 
> Thanks to MP3.com for sponsoring:
> 
>         Clustered/Replicated database support.
> 
>         Perl API.
> 
> --------------------------------------------------------------------------
> 
>  
> 
>  
> 
> Thank you for any suggestion.
> 
> 

[Sqlrelay-discussion] using sql relay with a specific ODBC driver

[Jeremie L. <jer...@ko...>]

Hello,

=20

            I would like to know if it is possible to use sqlrelay with
a specific odbc driver made for one of our servers.

The situation is :

=20

- I have an Alpha with OpenVMS on it

- I have a Redhat server, with a specific ODBC driver to connect to the
Alpha

- I have a Debian server, with Apache and PHP

=20

--> I try to make an ODBC link between RedHat and Debian, in order to
use databases of the Alpha on PHP (Debian).

=20

I can't directly connect Debian on the Alpha : the specific ODBC driver
is only supported on RedHat...

For information, the ODBC driver is made by Attunity.

=20

=20

=20

Our ODBC driver himself works. We could use it on PHP pages as well.

=20

--> We defined our odbc.ini as follows :

[Tolas]

Driver =3D Connect to VMS

Driver =3D /usr/local/navroot/lib/navcli.so

=20

=20

Here is the config file

=20

------------------------------------------------------------------------
--

<?xml version=3D"1.0"?>

<instances>

            <instance id=3D"toto" port=3D"9000"
unixport=3D"/tmp/sqlrelay.socket" dbase=3D"odbc" connections=3D"1"
maxconnections=3D"1" maxqueuelength=3D"1" growby=3D"1" ttl=3D"60"
endofsession=3D"rollback" sessiontimeout=3D"60" runasuser=3D"root"
runasgroup=3D"root" cursors=3D"5" authtier=3D"listener" =
handoff=3D"reconnect"
deniedips=3D"" allowedips=3D"" debug=3D"none">

                        <users>

                                   <user user=3D"titi" =
password=3D"titi"/>

                        </users>

                        <connections>

                                   <connection connectionid=3D"Tolas"
string=3D"user=3D;password=3D;dsn=3DTolas;autocommit=3Dyes" =
metric=3D"1"/>

                        </connections>

            </instance>

</instances>

------------------------------------------------------------------------
--

=20

=20

=20

=20

The problem is that the configuration seems correct, but when we start
sqlrelay, we got the following message :

=20

=20

------------------------------------------------------------------------
--

Starting listener:

  sqlr-listener -id toto -config /root/toto

=20

Starting 1 connections to Tolas :

  sqlr-connection-odbc -id toto -connectionid Tolas -config /root/toto

Couldn't log into database.

=20

sqlr-connection-odbc failed to start.

=20

=20

Thanks to MP3.com for sponsoring:

        Clustered/Replicated database support.

        Perl API.

------------------------------------------------------------------------
--

=20

=20

Thank you for any suggestion.

[Sqlrelay-discussion] Orphaned Oracle Cursor

[Himsl, J. <jas...@mc...>]

Hello,

I have encounter an interesting error with sqlrelay, oracle and cursor
resources. I'm executing a chunk of code which calls an Oracle Procedure
and binds to an Oracle output cursor. In my code I use two sqlrelay
cursors. An initial cursor to create the call and an output bound cursor
to retrieve the record set. The code executes normally and returns the
correct record set, but after a large number of calls the code begins to
fail. I invested and it appears there is a third cursor which is created
when I call the 'sqlrcur_executeQuery' function. After I finish
executing my code, the two cursors I created are cleared but the third
one is left open and can still be seen in the 'v$_cursor' oracle view.
This cursor resource is never cleared and after I get a little over 900
of these orphaned cursors my database calls break. I spent some time
investigating this issue since it is causing a stability issue. I double
checked to make sure that the orphaned cursor is not one of mine. I was
able to prove this by not freeing up my cursor resources. When I don't
free my two cursors I see three orphaned cursors in the database instead
of the normal one. So the cursor created by the call to
'sqlrcur_executeQuery' does not appear to deallocate properly. Also,
when I use the OCI classes with PHP and the same Oracle Procedure call,
I do not get the orphaned cursor.  Is there anything I can do to clear
this cursor explicitly?


My development environment consists of the following:
1) A Debian server running Apache with PHP 4.3.4
2) The same Debian server running sqlrelay (4 pooled connections to the
DB)
3) A Solaris server running Oracle 9.2

Thank you for your time,

~ J D H

Re: [Sqlrelay-discussion] PostgreSQL, bytea and stored procedures

[Firstworks/4access <dav...@fi...>]

On Wed, 2005-06-29 at 16:24 +0100, Ville Silventoinen wrote:
> Hi,
>=20
> I've been trying to use PostgreSQL bytea column type with SQL Relay.
> I think PostgreSQL functions support bytea input parameters and return
> type:
>=20
> http://www.postgresql.org/docs/8.0/static/plpgsql.html#PLPGSQL-ARGS-RES=
ULTS
>=20
> (Although it doesn't clearly say if bytea is supported or not.)
>=20
> I did notice that SQL Relay documentation says:
> "Blob/Clob bind variables are only supported in Oracle 8i or higher."
>=20
> Why are they not supported with PostgreSQL?

I just haven't gotten around to implementing it yet.  I need to support
Postgresql 8 binds too.

>=20
> I tried anyway! :) I installed "psycopg" module so I could quote the
> binary data, which is generated with cPickle.dumps(data, 2). This is
> how blobs are used in Python Cookbook, 2nd ed. chapter 7.11.
>=20
> I tried to call the PostgreSQL function like this (slightly simplified)=
:
>=20
> cur.prepareQuery("select * from sesmgr.create_session(:sid,:sdata,:suse=
r)")
> cur.inputBind('sid', '7e5578193932c84fb66c8d848d150d33')
> cur.inputBind('sdata', psycopg.Binary(cPickle.dumps(data, 2)))
> cur.inputBind('suser', 'vsi')
>=20
> This returns an error (cur.getErrorMessage):
>=20
> ERROR:  syntax error at or near ":" at character 72
>=20
> I *think* this is because SQL Relay doesn't actually bind the "sdata"
> input variable. If I turn connection debug on, I can see in the log fil=
e:
>=20
> 06/29/2005 16:02:04 BST connection [26828] : select * from sesmgr.creat=
e_session(:sid,:sdata,:suser)
> 06/29/2005 16:02:04 BST connection [26828] :            getting query s=
ucceeded
> 06/29/2005 16:02:04 BST connection [26828] :            getting input b=
inds...
> 06/29/2005 16:02:04 BST connection [26828] :                           =
 :sid
> 06/29/2005 16:02:04 BST connection [26828] :                           =
 STRING
> 06/29/2005 16:02:04 BST connection [26828] :                           =
 7e5578193932c84fb66c8d848d150d33
> 06/29/2005 16:02:04 BST connection [26828] :                           =
 :suser
> 06/29/2005 16:02:04 BST connection [26828] :                           =
 STRING
> 06/29/2005 16:02:04 BST connection [26828] :                           =
 vsi
> 06/29/2005 16:02:04 BST connection [26828] :            done getting in=
put binds


Yeah, it looks like the bind variable isn't getting sent at all.
inputBind() expects to send a string, there's a compiled-in maximum
string length, and strlen() of the pickled data is probably returning a
value longer than that value, so it's getting ignored.

>=20
>=20
> If I remove the psycopg.Binary call and pass the cPickle.dumps value
> as it is, SQL Relay binds the "sdata" as a STRING:
>=20
> 06/29/2005 15:58:50 BST connection [26828] : select * from sesmgr.creat=
e_session(:sid,:sdata,:suser)
> 06/29/2005 15:58:50 BST connection [26828] :            getting query s=
ucceeded
> 06/29/2005 15:58:50 BST connection [26828] :            getting input b=
inds...
> 06/29/2005 15:58:50 BST connection [26828] :                           =
 :sid
> 06/29/2005 15:58:50 BST connection [26828] :                           =
 STRING
> 06/29/2005 15:58:50 BST connection [26828] :                           =
 c0db86a43be8f9c6ae47af9dadb980a1
> 06/29/2005 15:58:50 BST connection [26828] :                           =
 :sdata
> 06/29/2005 15:58:50 BST connection [26828] :                           =
 STRING
> 06/29/2005 15:58:50 BST connection [26828] :                           =
 =C3=83=C2=AF=C3=82=C2=BF=C3=82=C2=BD}qUmydataq]q(U1U2U3es.
> 06/29/2005 15:58:50 BST connection [26828] :                           =
 :suser
> 06/29/2005 15:58:50 BST connection [26828] :                           =
 STRING
> 06/29/2005 15:58:50 BST connection [26828] :                           =
 vsi
> 06/29/2005 15:58:50 BST connection [26828] :            done getting in=
put binds
>=20
> But then cPickle.loads cannot read the value anymore:
>=20
> cPickle.UnpicklingError: invalid load key, '\'.

I suspect that the unpickled data is getting truncated because there's a
NULL somewhere in it.

>=20
>=20
> Is there a way to use bytea datatype with SQL Relay and PostgreSQL
> functions? Or do I have to use normal INSERT statements instead?

I think for now you'll have to use normal insert statements, but I'll
see if I can get it working in the next release (0.37).

>=20
> Thanks for any advice,
>=20
> Ville

Dave
dav...@fi...

[Sqlrelay-discussion] sqlrelay-0.36.4pre1

[Firstworks/4access <dav...@fi...>]

I just uploaded a prerelease to 
http://www.firstworks.com/sqlrelay-0.36.4pre1.tar.gz

This pre-release hopefully fixes a few bugs in 0.36.3:

* PHP pear db driver failed to bind NULL's
* Python driver could crash if getNullsAsNone was used
* fetching really large CLOB/LOBs was really slow

If you've had one of these problems, give it a try, let me know how it
works.  Unless someone reports trouble, I'll be making the official
0.36.4 release soon.

Dave
dav...@fi...

[Sqlrelay-discussion] PostgreSQL, bytea and stored procedures

[Ville S. <vs...@eb...>]

Hi,

I've been trying to use PostgreSQL bytea column type with SQL Relay.
I think PostgreSQL functions support bytea input parameters and return
type:

http://www.postgresql.org/docs/8.0/static/plpgsql.html#PLPGSQL-ARGS-RESULTS

(Although it doesn't clearly say if bytea is supported or not.)

I did notice that SQL Relay documentation says:
"Blob/Clob bind variables are only supported in Oracle 8i or higher."

Why are they not supported with PostgreSQL?

I tried anyway! :) I installed "psycopg" module so I could quote the
binary data, which is generated with cPickle.dumps(data, 2). This is
how blobs are used in Python Cookbook, 2nd ed. chapter 7.11.

I tried to call the PostgreSQL function like this (slightly simplified):

cur.prepareQuery("select * from sesmgr.create_session(:sid,:sdata,:suser)")
cur.inputBind('sid', '7e5578193932c84fb66c8d848d150d33')
cur.inputBind('sdata', psycopg.Binary(cPickle.dumps(data, 2)))
cur.inputBind('suser', 'vsi')

This returns an error (cur.getErrorMessage):

ERROR:  syntax error at or near ":" at character 72

I *think* this is because SQL Relay doesn't actually bind the "sdata"
input variable. If I turn connection debug on, I can see in the log file:

06/29/2005 16:02:04 BST connection [26828] : select * from sesmgr.create_se=
ssion(:sid,:sdata,:suser)
06/29/2005 16:02:04 BST connection [26828] :            getting query succe=
eded
06/29/2005 16:02:04 BST connection [26828] :            getting input binds=
=2E..
06/29/2005 16:02:04 BST connection [26828] :                            :si=
d
06/29/2005 16:02:04 BST connection [26828] :                            STR=
ING
06/29/2005 16:02:04 BST connection [26828] :                            7e5=
578193932c84fb66c8d848d150d33
06/29/2005 16:02:04 BST connection [26828] :                            :su=
ser
06/29/2005 16:02:04 BST connection [26828] :                            STR=
ING
06/29/2005 16:02:04 BST connection [26828] :                            vsi
06/29/2005 16:02:04 BST connection [26828] :            done getting input =
binds


If I remove the psycopg.Binary call and pass the cPickle.dumps value
as it is, SQL Relay binds the "sdata" as a STRING:

06/29/2005 15:58:50 BST connection [26828] : select * from sesmgr.create_se=
ssion(:sid,:sdata,:suser)
06/29/2005 15:58:50 BST connection [26828] :            getting query succe=
eded
06/29/2005 15:58:50 BST connection [26828] :            getting input binds=
=2E..
06/29/2005 15:58:50 BST connection [26828] :                            :si=
d
06/29/2005 15:58:50 BST connection [26828] :                            STR=
ING
06/29/2005 15:58:50 BST connection [26828] :                            c0d=
b86a43be8f9c6ae47af9dadb980a1
06/29/2005 15:58:50 BST connection [26828] :                            :sd=
ata
06/29/2005 15:58:50 BST connection [26828] :                            STR=
ING
06/29/2005 15:58:50 BST connection [26828] :                            =C3=
=AF=C2=BF=C2=BD}qUmydataq]q(U1U2U3es.
06/29/2005 15:58:50 BST connection [26828] :                            :su=
ser
06/29/2005 15:58:50 BST connection [26828] :                            STR=
ING
06/29/2005 15:58:50 BST connection [26828] :                            vsi
06/29/2005 15:58:50 BST connection [26828] :            done getting input =
binds

But then cPickle.loads cannot read the value anymore:

cPickle.UnpicklingError: invalid load key, '\'.


Is there a way to use bytea datatype with SQL Relay and PostgreSQL
functions? Or do I have to use normal INSERT statements instead?

Thanks for any advice,

Ville

Re: [Sqlrelay-discussion] Strange problem with sqlrsh

[Ville S. <vs...@eb...>]

Hi Dave,

I ran strace and it looks like the problem is that the user
doesn't have a home directory in the machine:

open("/homes/swlib/.sqlrelay.conf", O_RDONLY) = -1 ENOENT (No such file or directory)
open("/swlib/v1/etc/sqlrsh", O_RDONLY)  = -1 ENOENT (No such file or directory)
open("/homes/swlib/.sqlrshrc", O_RDONLY) = -1 ENOENT (No such file or directory)
open("/homes/swlib/.sqlrsh_history", O_WRONLY|O_APPEND|O_CREAT, 0666) = -1 ENOENT (No such file or directory)
--- SIGSEGV (Segmentation fault) @ 0 (0) ---
+++ killed by SIGSEGV +++


I will create a fake home directory for this user.
Thanks very much for the help!


Ville


On Tue, 28 Jun 2005, Firstworks/4access wrote:

> Could you run "strace sqlrsh -id web1-sysdev" and send me the output.
> It'll help me track down the crash.
>
> Dave
> dav...@fi...
>
> On Thu, 2005-06-23 at 17:23 +0100, Ville Silventoinen wrote:
>> Hi again,
>>
>> I forgot to mention that I get the usage text from sqlrsh successfully, so
>> I haven't completely botched the compilation. ;-)
>>
>> $ sqlrsh
>> usage: sqlrsh  host port socket user password [script]
>>    or   sqlrsh  [-config configfile] -id id [script]
>> $ sqlrsh -id web1-sysdev
>> Segmentation fault
>> $ sqlrsh -config /swlib/v1/etc/sqlrelay.conf -id web1-sysdev
>> Segmentation fault
>>
>> Cheers,
>> Ville
>>
>>
>>
>> On Thu, 23 Jun 2005, Ville Silventoinen wrote:
>>
>>> Hi,
>>>
>>> I'm using SQL Relay 0.36 with rudiments-0.28.2 on three different Linux
>>> machines (all web servers). I just noticed today that sqlrsh dumps core
>>> (Segmentation fault) on one of the machines. I can use PySQLRClient without
>>> any problems, it's just a problem with sqlrsh.
>>>
>>> The machine is:
>>> Red Hat Linux release 9 (Shrike)
>>> Linux 2.4.20-8rlx4smp #1 SMP
>>>
>>> Is there something I can do to find out what goes wrong? I turned debug
>>> logging on, but the problem seems to be just with sqlrsh client, not the
>>> listener or connections. I also restarted SQL Relay, but it didn't help.
>>>
>>> sqlrsh works ok in the other machines:
>>>
>>> CentOS release 3.4 (final)
>>> Linux 2.4.21-27.0.1.ELsmp #1 SMP
>>>
>>> I've compiled SQL Relay on the first machine and the installation
>>> is shared between all the machines over NFS. (If I compile it on
>>> CentOS, the C libraries differ too much and it doesn't work on the
>>> older Linux OS.)
>>>
>>>
>>> Thanks,
>>> Ville
>>>
>>>
>>>
>>> -------------------------------------------------------
>>> SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
>>> from IBM. Find simple to follow Roadmaps, straightforward articles,
>>> informative Webcasts and more! Get everything you need to get up to
>>> speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
>>> _______________________________________________
>>> Sqlrelay-discussion mailing list
>>> Sql...@li...
>>> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>>>
>>
>>
>> -------------------------------------------------------
>> SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
>> from IBM. Find simple to follow Roadmaps, straightforward articles,
>> informative Webcasts and more! Get everything you need to get up to
>> speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
>> _______________________________________________
>> Sqlrelay-discussion mailing list
>> Sql...@li...
>> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>>
>
>
>
> -------------------------------------------------------
> SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
> from IBM. Find simple to follow Roadmaps, straightforward articles,
> informative Webcasts and more! Get everything you need to get up to
> speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>

Re: [Sqlrelay-discussion] Strange problem with sqlrsh

[Firstworks/4access <dav...@fi...>]

Could you run "strace sqlrsh -id web1-sysdev" and send me the output.
It'll help me track down the crash.

Dave
dav...@fi...

On Thu, 2005-06-23 at 17:23 +0100, Ville Silventoinen wrote:
> Hi again,
> 
> I forgot to mention that I get the usage text from sqlrsh successfully, so 
> I haven't completely botched the compilation. ;-)
> 
> $ sqlrsh
> usage: sqlrsh  host port socket user password [script]
>    or   sqlrsh  [-config configfile] -id id [script]
> $ sqlrsh -id web1-sysdev
> Segmentation fault
> $ sqlrsh -config /swlib/v1/etc/sqlrelay.conf -id web1-sysdev
> Segmentation fault
> 
> Cheers,
> Ville
> 
> 
> 
> On Thu, 23 Jun 2005, Ville Silventoinen wrote:
> 
> > Hi,
> >
> > I'm using SQL Relay 0.36 with rudiments-0.28.2 on three different Linux 
> > machines (all web servers). I just noticed today that sqlrsh dumps core 
> > (Segmentation fault) on one of the machines. I can use PySQLRClient without 
> > any problems, it's just a problem with sqlrsh.
> >
> > The machine is:
> > Red Hat Linux release 9 (Shrike)
> > Linux 2.4.20-8rlx4smp #1 SMP
> >
> > Is there something I can do to find out what goes wrong? I turned debug 
> > logging on, but the problem seems to be just with sqlrsh client, not the
> > listener or connections. I also restarted SQL Relay, but it didn't help.
> >
> > sqlrsh works ok in the other machines:
> >
> > CentOS release 3.4 (final)
> > Linux 2.4.21-27.0.1.ELsmp #1 SMP
> >
> > I've compiled SQL Relay on the first machine and the installation
> > is shared between all the machines over NFS. (If I compile it on
> > CentOS, the C libraries differ too much and it doesn't work on the
> > older Linux OS.)
> >
> >
> > Thanks,
> > Ville
> >
> >
> >
> > -------------------------------------------------------
> > SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
> > from IBM. Find simple to follow Roadmaps, straightforward articles,
> > informative Webcasts and more! Get everything you need to get up to
> > speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
> > _______________________________________________
> > Sqlrelay-discussion mailing list
> > Sql...@li...
> > https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
> >
> 
> 
> -------------------------------------------------------
> SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
> from IBM. Find simple to follow Roadmaps, straightforward articles,
> informative Webcasts and more! Get everything you need to get up to
> speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
> 

Re: [Sqlrelay-discussion] number of binding variables problem

[Laurent C. <lau...@sm...>]

Firstworks/4access wrote:
> If you modify the file src/common/defines.h, changing the 256 in the
> line:
> 
> #define MAXVAR 256
> 
> to a larger number and rebuild, then you will be able to use as many
> bind variables as you change the number to.

thank you i'll try this.

> How many bind variables are you using?

the number of keys to update, between 10 and 1000+. most of the time ~100.


-- 
Laurent Corbes - lau...@sm...
+33 (0)1 4996 6319
Smartjog SA - http://www.smartjog.com/

Re: [Sqlrelay-discussion] number of binding variables problem

[Firstworks/4access <dav...@fi...>]

If you modify the file src/common/defines.h, changing the 256 in the
line:

#define MAXVAR 256

to a larger number and rebuild, then you will be able to use as many
bind variables as you change the number to.

How many bind variables are you using?

Dave
dav...@fi...

On Fri, 2005-06-24 at 17:12 +0200, Laurent Corbes wrote:
> Hi,
> I'm using sqlrelay 0.35-10 with an oracle8 database (i recompile the
> debian package with oracle8 support) and have a problem with queries
> that use a lot a binding variables (more than 256). The program is in
> python and use the debian python2.3-sqlrelay 0.35-10 package.
> 
> i must make an update of a list of keys (between 10 and 1000+) and build
> a query like this:
> 
> UPDATE table SET datevalue=sysdate WHERE key IN ( :value_1, value_2,
> value_3, .... , value_x)
> 
> The i do an inputBind for each value_x.
> 
> The query is run successfully when i have less than 256 bindings but
> failed with an oracle exception: 'ORA-01008: not all variables
> bound\n\nAttempted Query:\nUPDA....
> 
> It works perfectly with cxoracle previously used.
> Is this a known limitation (i didn't find anything about this) ? I
> didn't see anything related to this in the 0.36.x changelog so i don't
> check with this version.
> 
> I can bypass the limitation with a lot of rebindings but this is 2 times
> slower with only 256 bindings.
> 
> Thanks.

[Sqlrelay-discussion] number of binding variables problem

[Laurent C. <lau...@sm...>]

Hi,
I'm using sqlrelay 0.35-10 with an oracle8 database (i recompile the
debian package with oracle8 support) and have a problem with queries
that use a lot a binding variables (more than 256). The program is in
python and use the debian python2.3-sqlrelay 0.35-10 package.

i must make an update of a list of keys (between 10 and 1000+) and build
a query like this:

UPDATE table SET datevalue=sysdate WHERE key IN ( :value_1, value_2,
value_3, .... , value_x)

The i do an inputBind for each value_x.

The query is run successfully when i have less than 256 bindings but
failed with an oracle exception: 'ORA-01008: not all variables
bound\n\nAttempted Query:\nUPDA....

It works perfectly with cxoracle previously used.
Is this a known limitation (i didn't find anything about this) ? I
didn't see anything related to this in the 0.36.x changelog so i don't
check with this version.

I can bypass the limitation with a lot of rebindings but this is 2 times
slower with only 256 bindings.

Thanks.
-- 
Laurent Corbes - lau...@sm...
+33 (0)1 4996 6319
Smartjog SA - http://www.smartjog.com/

Re: [Sqlrelay-discussion] Special chars

[Robert (J. M. <ja...@te...>]

David Muse wrote:
> Unicode takes care of all of this.  In unicode, each characer code is
> 16 bits long, so there are enough codes to represent all special
> characters in a single character set.  But, it takes twice as much
> space to store the text.

It's not quite as simple as that. Unicode 2 can have up to 31 bits
(currently!), but there is a way to compress and encode it so that all
the characters below 0-127 appear as normal ascii, and other characters
are stored with more bytes as neccesary. This is called UTF-8. The rule
is that for every binary 1 at the start, you add a byte on the end & you
must always use the shortest encoding. So you can have 0bbbbbbb for
characters up to 7 bits (i.e. 0-127 or normal ascii) or 10bbbbbb
bbbbbbbb for characters up to 12 bits (128-4095) or 110bbbbb bbbbbbbb
bbbbbbbb for characters up to 21 bits long (4096-2097151 - more than
defined in Unicode 1) etc.

Unless there's an overwhelming reason not to use it, I use UTF-8
wherever I can. The only downside is that if I look at UTF-8 in
something that is not UTF-8 aware, the accents get broken and end up as
2 or 3 characters. I don't have much that is not UTF-8 aware, so this
doesn't actually happen all that often.

Robert Munro

Re: [Sqlrelay-discussion] SQLRelay, Python and getNullsAsNone problem

[David M. <dav...@fi...>]

I'm looking into this bug right now and I can't reproduce it with SQL
Relay 0.36.3.  Is that the version you're using?  Could you send me a
sample table definition and data that exhibits the problem?

Thanks,

Dave
dav...@fi...

On Tue, 2005-06-21 at 10:50 +0200, Maciej Wisniowski wrote:
> Hi
> 
> I've already described my problem in previous post but there were no 
> answer, so I'll try again,
> more precisely.
> 
> I have simple script:
> 
> from SQLRelay import PySQLRClient
> from SQLRelay import PySQLRDB
> 
> con=PySQLRClient.sqlrconnection('localhost',9000,'','test','test',0,1)
> cur=PySQLRClient.sqlrcursor(con)
> cur.getNullsAsNone()
> 
> for a in range(100):
>     res = cur.sendQuery('select * from CKK_KLIENCI')  
>     print 'cursor:', cur.rowCount()
>     print 'run_no:', a
>     for row in range(cur.rowCount()):
>         print 'row:', row
>         for col in range(cur.colCount()):
>             cur.getField(row,col)
> con.endSession()
> print 'end'
> 
> 
> The result is:
> 
> cursor: 329
> run_no: 0
> row: 0
> row: 1
> row: 2
> row: 3
> row: 4
> row: 5
> row: 6
> row: 7
> row: 8
> row: 9
> row: 10
> row: 11
> row: 12
> row: 13
> Fatal Python error: deallocating None
> 
> The problem disappears when I remove getNullsAsNone or when I change 
> select to:
> select column1 from CKK_KLIENCI.
> So the problem is with getNullsAsNone and with select that returns more 
> than two
> columns (maybe this is not the rule with column count).
> 
> Is this only my problem, or maybe somebody can confirm that this error 
> appears??
> 

Re: [Sqlrelay-discussion] Strange problem with sqlrsh

[Ville S. <vs...@eb...>]

Hi again,

I forgot to mention that I get the usage text from sqlrsh successfully, so 
I haven't completely botched the compilation. ;-)

$ sqlrsh
usage: sqlrsh  host port socket user password [script]
   or   sqlrsh  [-config configfile] -id id [script]
$ sqlrsh -id web1-sysdev
Segmentation fault
$ sqlrsh -config /swlib/v1/etc/sqlrelay.conf -id web1-sysdev
Segmentation fault

Cheers,
Ville



On Thu, 23 Jun 2005, Ville Silventoinen wrote:

> Hi,
>
> I'm using SQL Relay 0.36 with rudiments-0.28.2 on three different Linux 
> machines (all web servers). I just noticed today that sqlrsh dumps core 
> (Segmentation fault) on one of the machines. I can use PySQLRClient without 
> any problems, it's just a problem with sqlrsh.
>
> The machine is:
> Red Hat Linux release 9 (Shrike)
> Linux 2.4.20-8rlx4smp #1 SMP
>
> Is there something I can do to find out what goes wrong? I turned debug 
> logging on, but the problem seems to be just with sqlrsh client, not the
> listener or connections. I also restarted SQL Relay, but it didn't help.
>
> sqlrsh works ok in the other machines:
>
> CentOS release 3.4 (final)
> Linux 2.4.21-27.0.1.ELsmp #1 SMP
>
> I've compiled SQL Relay on the first machine and the installation
> is shared between all the machines over NFS. (If I compile it on
> CentOS, the C libraries differ too much and it doesn't work on the
> older Linux OS.)
>
>
> Thanks,
> Ville
>
>
>
> -------------------------------------------------------
> SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
> from IBM. Find simple to follow Roadmaps, straightforward articles,
> informative Webcasts and more! Get everything you need to get up to
> speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>

[Sqlrelay-discussion] Strange problem with sqlrsh

[Ville S. <vs...@eb...>]

Hi,

I'm using SQL Relay 0.36 with rudiments-0.28.2 on three different Linux 
machines (all web servers). I just noticed today that sqlrsh dumps core 
(Segmentation fault) on one of the machines. I can use PySQLRClient 
without any problems, it's just a problem with sqlrsh.

The machine is:
Red Hat Linux release 9 (Shrike)
Linux 2.4.20-8rlx4smp #1 SMP

Is there something I can do to find out what goes wrong? I turned debug 
logging on, but the problem seems to be just with sqlrsh client, not the
listener or connections. I also restarted SQL Relay, but it didn't help.

sqlrsh works ok in the other machines:

CentOS release 3.4 (final)
Linux 2.4.21-27.0.1.ELsmp #1 SMP

I've compiled SQL Relay on the first machine and the installation
is shared between all the machines over NFS. (If I compile it on
CentOS, the C libraries differ too much and it doesn't work on the
older Linux OS.)


Thanks,
Ville

Re: [Sqlrelay-discussion] SQLRelay, Python and getNullsAsNone problem

[David M. <dav...@fi...>]

I'll look into this, looks like there's a bug.  It shouldn't be too hard
to track down.  I have a hunch where it is and what to do to fix it.
I've got to make a 0.36.4 maintenance release soon, I should be able to
fix this in it too.

Dave
dav...@fi...

On Tue, 2005-06-21 at 10:50 +0200, Maciej Wisniowski wrote:
> Hi
> 
> I've already described my problem in previous post but there were no 
> answer, so I'll try again,
> more precisely.
> 
> I have simple script:
> 
> from SQLRelay import PySQLRClient
> from SQLRelay import PySQLRDB
> 
> con=PySQLRClient.sqlrconnection('localhost',9000,'','test','test',0,1)
> cur=PySQLRClient.sqlrcursor(con)
> cur.getNullsAsNone()
> 
> for a in range(100):
>     res = cur.sendQuery('select * from CKK_KLIENCI')  
>     print 'cursor:', cur.rowCount()
>     print 'run_no:', a
>     for row in range(cur.rowCount()):
>         print 'row:', row
>         for col in range(cur.colCount()):
>             cur.getField(row,col)
> con.endSession()
> print 'end'
> 
> 
> The result is:
> 
> cursor: 329
> run_no: 0
> row: 0
> row: 1
> row: 2
> row: 3
> row: 4
> row: 5
> row: 6
> row: 7
> row: 8
> row: 9
> row: 10
> row: 11
> row: 12
> row: 13
> Fatal Python error: deallocating None
> 
> The problem disappears when I remove getNullsAsNone or when I change 
> select to:
> select column1 from CKK_KLIENCI.
> So the problem is with getNullsAsNone and with select that returns more 
> than two
> columns (maybe this is not the rule with column count).
> 
> Is this only my problem, or maybe somebody can confirm that this error 
> appears??
> 

Re: [Sqlrelay-discussion] bind pear php module

[David M. <dav...@fi...>]

Looks like there's a stray debug command left over from testing.

Edit sqlrelay.php, look for sqlrcon_debugOn() and remove the line.

I think also I put the identity variable in the wrong class.  Remove it
from the DB_sqlrelay_cursor class (around line 33) and add it to the
DB_sqlrelay class (around line 52) and see if that doesn't fix it.

Dave
dav...@fi...

On Fri, 2005-06-10 at 14:42 +0200, david forums wrote:
> Hello
> 
> After my last email, I recompile the rpm myself, even if I meet some  
> difficulties, I could get a new "sql_relay.so", with works fine.
> 
> But using "sqlrelay.php" pear module, I get the following output :
> Result Set Buffer Size: 100 Connecting to listener... Unix socket:  
> /tmp/MyPostgres.socket Authenticating with listener : pgTest:test No  
> authentication error. Must Not Reconnect. Authenticating with connection :  
> pgTest:test No authentication error. Sending Query: select  
> nextval('public.mySequence') Length: 35 Requesting a new cursor. Sending 0  
> Input Bind Variables: Sending Output Bind Variables: Send Column Info: yes  
> Skipping and Fetching row to get: 99 Skipping 0 rows Fetching 100 rows  
> Checking For An Error... none. Getting Cursor ID... Cursor ID: 0 Previous  
> result set was not suspended. Parsing Column Info Actual row count: 1  
> Affected row count: 0 Column count: 1 "nextval","20", 8 (0,0) NOT NULL  
> Receiving Output Bind Values: Parsing Data "4", Got end of result set.
> Warning: Missing argument 2 for DB_sqlrelay_cursor::DB_sqlrelay_cursor()  
> in /usr/share/pear/DB/sqlrelay.php on line 36
> Aborting Result Set For Cursor: 0 Deallocated cursor
> Notice: Undefined variable: identity in /usr/share/pear/DB/sqlrelay.php on  
> line 278
> Identifying... postgresql
> Notice: Undefined variable: identity in /usr/share/pear/DB/sqlrelay.php on  
> line 297
> 
> Notice: Undefined variable: identity in /usr/share/pear/DB/sqlrelay.php on  
> line 297
> 
> Notice: Undefined variable: identity in /usr/share/pear/DB/sqlrelay.php on  
> line 297
> 
> Notice: Undefined variable: identity in /usr/share/pear/DB/sqlrelay.php on  
> line 297
> 
> Notice: Undefined variable: identity in /usr/share/pear/DB/sqlrelay.php on  
> line 297
> 
> Notice: Undefined variable: identity in /usr/share/pear/DB/sqlrelay.php on  
> line 297
> Result Set Buffer Size: 100
> Notice: Undefined variable: connection in /usr/share/pear/DB/sqlrelay.php  
> on line 204
> Sending Query: INSERT INTO myTable (id,libelle,lid) VALUES  
> (:id,:libelle,:lid) Length: 74 Requesting a new cursor. Sending 3 Input  
> Bind Variables: id(4:STRING)=1(1) libelle(9:STRING)=Monsieur(8)  
> lid(4:STRING)=4(1) Sending Output Bind Variables: Send Column Info: yes  
> Skipping and Fetching row to get: 99 Skipping 0 rows Fetching 100 rows  
> Checking For An Error...
> Aborting Result Set For Cursor: 0 Deallocated cursor
> Fatal error: Uncaught exception 'Exception' with message 'DB Error:  
> unknown error' in my.class.php5:251 Stack trace: #0 my.class.php5(251):  
> my::save() #1 other.php5(23): other->save() #2 {main} thrown in  
> my.class.php5 on line 251
> 
> Ending Session Deallocated connection
> 
> 
> 
> 
> Is it normal ???
> 
> Could you help please
> 
> david
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: NEC IT Guy Games.  How far can you shotput
> a projector? How fast can you ride your desk chair down the office luge track?
> If you want to score the big prize, get to know the little guy.  
> Play to win an NEC 61" plasma display: http://www.necitguy.com/?r=20
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
> 

Re: [Sqlrelay-discussion] bind pear php module

[David M. <dav...@fi...>]

I think I know what's causing that.  I need to fix that and another bug
and make a 0.36.4 release.  I should be able to get that done this week.

Dave
dav...@fi...

On Tue, 2005-06-14 at 15:46 +0200, david forums wrote:
> Hi
> 
> Concerning the last update you made in the version 0.36.3.
> 
> I detecte the following error concerning php pear module.
> 
> When I try to use autoExecute and I send NULL value, it is replying :
> syntax error on or close to «:»
> 
> I try the same sentence using straight database access and it works well.
> 
> Could you help to solve it, please.
> 
> Regards
> 
> David
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: NEC IT Guy Games.  How far can you shotput
> a projector? How fast can you ride your desk chair down the office luge track?
> If you want to score the big prize, get to know the little guy.  
> Play to win an NEC 61" plasma display: http://www.necitguy.com/?r=20
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>