Parameter problems with ZuckerReports 1.8

[srazvigorov]

First of all I want to thank ZuckerReports authors for the great product.

The problem I have is as following: I have designed (using iReport) several jasper reports which require direct input parameter. Reports are running fine in iReport and on JasperServer. When these same reports are run through ZuckerReports they fail to return correct data when the input parameter contains non latin characters (Cyrilic in may case). It does not matter whether the parameter is directly input through the keyboard or an sql query is used to obtain data from the database.
Enabling "debug" option in ZuckerReports showed me that the parameter is correctly collected as Unicode and passed to JasperBatchMain class. At the same time the MySQL log shows that the sql server receives a query with smashed parameter (all characters interpreted as cp1252). In this case obviously the server does not return correct data. I am using SugarCRM 4.5.1e+ZuckerReports1.8+MySQL 5.0.27.

Can anybody help to resolve the issue? I can see that ZuckerReports authors actively participate in these forums and hope they will give an advise.

[ftreml]

Hi,
can you please try to make the following patch and tell me if it helps (by PM or email to f.treml (at) go-mobile.at


File: modules/ZuckerReportTemplate/ReportTemplate.php
Line: 245
Replace with:

fwrite($f, "parameter.".utf8_decode($name)."=".utf8_decode($value)."\n");

[kazkawamura]

Hi,

I couldn't have a look at the source code of JasperBatchMain.
But, zucker makes properties file. So, it might use java.util.Properties.
Assuming zucker uses java.util.Properties, Properties.load only accepts iso-8859-1. If there are some characters which are out of iso-8859-1, java.util.Properties couldn't read them correctly. In this case, we need to unicode-escape these characters like \uxxxx.

I guess it needs to native2ascii the prop file before calling java program.

[srazvigorov]

Hi again,

After additional debug I found out that:

1. parameters directly input from ZuckerReports are collected correctly as utf8, but are later passed to the jdbc driver as if they are latin1 characters (they are in utf8).
2. if you configure jdbc driver to use utf8 character set with MySQL server the driver converts the input parameters to utf8, but again from latin1 besides the parameters are actually in utf8.
3. The above points are valid only for jrxml reports in Zucker. If you use directly input parameters for sql queries defined in Zucker everything is ok.

It sounds like a problem in the java part of ZuckerReports but without the source I can not be sure.