Problem with SVG generation for anions (

Hi all,

There is a problem of generating SVG for anions - incorrect display for "minus" symbol.

There are source mol and generated svg file in attachement.

SVG genereted by jc_molconvert without any additional options for svg or with 'H_off'.

JChem Cartridge Environment:

 

Java VM vendor: Sun Microsystems Inc.

Java version: 1.6.0_23

Java VM version: 19.0-b09

JChem version: 5.6.0.0

JChem Index version: 5060000

JDBC driver version: 11.1.0.7.0-Production

on Oracle Database 10.2.0.5 x86_64

 

 

Thanks

Hi,

Please, could you execute on the JChem Server host in the jchem/bin directory a command similar to the following:

C:\Users\pkovacs\chemaxon\packages\jchem-5.6.0.0beta1\bin>molconvert svg C:\Users\pkovacs\Downloads\_misc\bad_anion_-1141224919.mol > C:\Users\pkovacs\b.svg

? (That is: molconvert svg <mol-file> .) Does the resulting svg image exhibit a similar problem?

Thanks

Peter

Hi, Peter.

I've executed:

jchem/bin/molconvert svg:H_off molfiles/bad_anion_-1141224919.mol > svgs/bad_anion_-1141224919-1.svg

and this SVG is correct (attached).

I could reproduce the problem. I believe this due to some character encoding discrepancies between the Oracle server and the JCC server, I have to exactly pin down the problem yet.

One potential work-around is to use jc_molconvertb which will avoid the character decoding step (from byte to char) in the Oracle server. You'll thus have a chance to use the decoding corresponding to the encoding (from char to byte) in the JCC server. This should work around the problem "automatically" in most environments, since the default non-Oracle decoding is likely to automatically match the equally non-Oracle encoding.

Peter

Hi, Peter

pkovacs wrote:

... I believe this due to some character encoding discrepancies between the Oracle server and the JCC server...

Yes, I thought about the same.

To reproduce this issue maybe this information will be useful:

JChem server start with LANG=C environment (also I tried use LANG=en_US.UTF-8 - no affect on result). Oracle sessions have NLS_LANG=AMERICAL_AMERICA.UFT8 parameter and Oracle Database have NLS_CHARACTERSET=AL32UTF8.

One potential work-around is to use jc_molconvertb which will avoid the character decoding step (from byte to char) in the Oracle server. You'll thus have a chance to use the decoding corresponding to the encoding (from char to byte) in the JCC server. This should work around the problem "automatically" in most environments, since the default non-Oracle decoding is likely to automatically match the equally non-Oracle encoding.

I'll try it soon and will feedback results.

Hi, Peter, again.

And your workaround is working (see result in attachement).

Thanks a lot!

Of course it will be more interesting using jc_molconvert directly to avoid additional converting clob to blob.

Hi,

I've found the culprit in the JChem Cartridge code which was responsible for the encoding problems. It will be fixed in JChem version 5.8.

Peter

pkovacs wrote:

Hi, I've found the culprit in the JChem Cartridge code which was responsible for the encoding problems. It will be fixed in JChem version 5.8. Peter

Thanks, Peter.

When are you plan to release 5.8 version?

There is no release date scheduled yet for 5.8, but based on the current commit dead-line (2011-11-11), I'd expect it to be sometime in December (or in the worst case early next year).

A version 5.7.1 in our release schedule is currently planned for 2011-10-21. If you need this fix urgently, I can include it in 5.7.1 as well.

Peter

pkovacs wrote:

There is no release date scheduled yet for 5.8, but based on the current commit dead-line (2011-11-11), I'd expect it to be sometime in December (or in the worst case early next year). A version 5.7.1 in our release schedule is currently planned for 2011-10-21. If you need this fix urgently, I can include it in 5.7.1 as well. Peter

Thanks Peter. It's very good news.

If it's possible, please, include this fix in 5.7.1. It will be the best help for us.

I am very appreciate to you for help and most importantly its efficiency. Now, unfortunately, it's rare to find a tech support with such qualities as yours.
Why do you thank you again.

Pavel,

The fix will be included in 5.8. (If 5.7.1 is released, itt will be a hot-fix release, for the most critical problems only.) A preview release (based on production JChem version 5.7.0) with the JChem Cartridge fixes and improvements planned for 5.8.0 is available here: https://shared.chemaxon.com/projects/jcc/jchem-jcc-5.8preview-on-5.7.zip . Please, have a look when/if you have time.

Peter

JChem 5.8.0 has been released with the fix.

Peter