Oracle® Application Server Release Notes
10g Release 2 (10.1.2) for Solaris Operating System (SPARC) B14500-14 |
|
Previous |
Next |
This chapter describes issues with Oracle Reports. It includes the following topics:
See Also:
|
This section describes general issues and their workarounds for Oracle Reports. It includes the following topics:
Section 18.1.1, "Enabling Backward Compatibility with 9.0.4"
Section 18.1.2, "Using the New Auto Save Feature in Reports Builder"
Section 18.1.3, "OracleAS Reports Services Demos Do Not Work If JServ Is Configured"
Section 18.1.4, "JSP Report Does Not Run in Oracle JDeveloper on UNIX"
Section 18.1.5, "Error While Editing Batch Registered Report in OracleAS Portal"
Section 18.1.6, "Limitation on Linked Queries for DelimitedData Output"
Section 18.1.11, "PL/SQL Compilation Failure When Using SQL Constructs"
Section 18.1.12, "Reports Engine Fails When NLS_LANG is Specified As JA16SJISTILDE or JA16EUCTILDE"
Oracle Reports 10g Release 2 (10.1.2) replaces the use of Borland VisiBroker with Sun Microsystems industry-standard Java Developer's Kit Object Request Broker (JDK ORB), providing support for Reports Server requests from clients across subnets, and using the broadcast mechanism for dynamic Reports Server discovery, both within a subnet and across subnets. For information about the changes related to JDK ORB, refer to the Oracle Application Server Reports Services Publishing Reports to the Web manual.
For releases prior to 9.0.4.3, you must apply the patch issued for 9.0.4.2 to achieve ORB backward compatibility between Oracle Reports 10g Release 1 (9.0.4) client and 10g Release 2 (10.1.2) server, or vice versa. Beginning with 9.0.4.3, the patch is included in the installation.
Platform-specific details for this patch are provided in bug 4092150.
Oracle Reports 10g Release 2 (10.1.2) introduces the Auto Save feature, which recovers unsaved changes in the case of an unexpected event of Reports Builder or system fail. When Auto Save is enabled, Reports Builder can automatically recover unsaved changes in open reports.
To enable the Auto Save functionality, you must perform the following steps:
In Reports Builder, select Edit, and then select Preferences to display the Preferences dialog box.
In the Preferences dialog box, on the General tab, select Auto Save.
Alternatively, you can enable the Auto Save functionality by modifying the Reports.auto_save
property in the preferences file, cauprefs.ora
for Windows or prefs.ora
for UNIX platforms. Modify the file as follows:
Reports.auto_save = [YES|NO] DEFAULT: NO
Usage Notes
When Auto Save is enabled, any modifications done in the reports that are open are saved automatically in temporary recovery files at various events originating from menu, toolbar, tool palette, object navigator, property inspector, and other editor windows in Reports Builder.
For a new report definition, Auto Save is enabled only after the report is saved the first time.
After you save your report, Reports Builder deletes the current recovery file. Even if you just recovered unsaved changes, Reports Builder deletes the recovery file when you save the report.
The temporary recovery file is created and saved in .rdf
format in the same location as that of the original definition file. The format of recovery files is reportname_extension.rdf
. For example, the recovery file of an original definition file, emp.jsp
, would be emp_jsp.rdf
. The emp_jsp.rdf
file is saved in the same location as the emp.jsp
file.
If you reopen a previously unsaved report definition file for which a recovery file exists, then Reports Builder prompts you to save the changes. The recovery file contains the changes made until the last Auto Save event. After the changes are recovered, you must select the Save or Revert option. The Save option writes the unsaved previous changes to the original report definition file. The Revert option does not save the recovered changes. Reports Builder deletes the recovery file when you close the report.
If an error occurs when autosaving the recovery file, then Reports Builder prompts you to disable Auto Save or continue with the error.
Note: When Auto Save is enabled, Reports Builder assumes that the recovery file contains changes made since the last save of the report definition file. Before you choose to save the recovery file, you must ensure that the recovery file contains your latest changes. If the original report definition file is a later version than the recovery file, then you can choose to revert to the original report definition, so that you do not overwrite it with the changes read from the recovery file. |
To run OracleAS Reports Services demos successfully, do not configure Apache JServ on the computer where OracleAS Reports Services is installed.
To run a JSP-based Web report in Oracle JDeveloper on UNIX, perform either of the following steps:
For the in-process Reports Server, edit ORACLE_HOME
/jdev/bin/jdev
to include the path to lib32
. Append the path of lib32
to LD_LIBRARY_PATH
=ORACLE_HOME
/lib
, as shown in the following example:
LD_LIBRARY_PATH=ORACLE_HOME/lib:ORACLE_HOME/lib32
Save and invoke Oracle JDeveloper.
Use the standalone Reports Server to run JSP-based Web report in Oracle JDeveloper.
If you batch register a report in OracleAS Portal, then subsequently manually edit the report's registration in OracleAS Portal, you may encounter an unexpected Save As dialog box during the manual editing process.
To work around this issue, specify the Display Name property while editing the report's registration in OracleAS Portal.
Refer to the Oracle Application Server Reports Services Publishing Reports to the Web manual for information on how to register a report with OracleAS Portal.
DelimitedData output does not allow multiple queries to be linked to the main query, because sibling groups are not supported.
For example:
Case 1: You have a link between Q1
and Q2
, a link between Q2
and Q3
, and a standalone Q4
query.
Q1<-->Q2<-->Q3 Q4
In this case, DelimitedData output is generated correctly.
Case 2: You have a link between Q1
and Q2
, a link between Q2
and Q3
, and a link between Q4
and Q1
.
Q1<-->Q2<-->Q3 Q1<-->Q4
In this case, DelimitedData output is not generated correctly. Alternatively, you can use XML output.
If any summary column values are not correctly left-aligned in the output of JSP-based Web reports with a Group Above or Matrix style, then you can implement the following workaround to get proper alignment:
Open the JSP report in Reports Builder.
Select the Web Source view.
In the Web Source view, locate the section where the summary column is defined, and delete the following line to remove the extra space:
<th class="
summary_column_name
"> </th>
If OracleAS Web Cache is configured as SSL-enabled and Oracle HTTP Server is not, then the images in the HTML output of Oracle Reports is not displayed when viewed from Oracle Enterprise Manager 10g. This is because the image URLs mentioned in the HTML output points to OracleAS Web Cache ports, which are SSL-enabled, whereas the request for past job outputs of Oracle Reports from Oracle Enterprise Manager 10g is non-SSL.
To work around this issue, edit the Servlet property of the URL in the targets.xml
file for Reports Server type such that it points to the OracleAS Web Cache port instead of the Oracle HTTP Server port. The targets.xml
file is available at ORACLE_HOME
/sysman/emd
directory.
For example, if your OracleAS Web Cache is SSL-enabled and the listening port number is 443
, then the Servlet property of the URL for the Reports Server target in the targets.xml
file should be as shown in the following example:
<Target TYPE="oracle_repserv" …..> ...... <Property NAME="Servlet" VALUE="https://xyz.mycompany.com:443/reports/rwservlet"/> ...... </Target>
If you use Acrobat Reader 7.0 to view multibyte and unicode PDF reports that use enhanced font subsetting feature (default) and the report size is very large, then some of the characters displayed will not be aligned properly.
To work around this issue, you can do either of the following:
Set the environment variable REPORTS_ENHANCED_SUBSET
to NO
to revert to the Type 3 font subsetting implementation used in releases prior to Oracle Reports 10g Release 2 (10.1.2).
Use Acrobat Reader 6.0 or earlier.
If any of the following conditions are present, then the text in the Paper Design view of Reports Builder may display in an incorrect color:
A text object includes multiline text, which wraps beyond the first line.
The text object's Contains HTML Tags property is set to Yes.
More than one color is applied to different parts of the multiline text.
Note: This issue is shown in the Paper Design view only. The report output in all other output formats show correct colors. |
PL/SQL compilation may fail if you use SQL constructs in your reports and if the following conditions are present:
The SQL constructs were introduced in Oracle Database versions after 9.0.1, for example, ROW_NUMBER()
.
The SQL constructs are used in client side PL/SQL procedures.
To work around this issue, you can do any of the following:
Use Oracle Database Release 10.1.0.4 or later.
Move those SQL constructs to the server-side stored procedures.
Use the SQL constructs within SQL queries directly instead of procedures.
Executing the following command when the NLS_LANG
environment variable is set to JA16SJISTILDE
or JA16EUCTILDE
causes the Reports Engine to crash on the AIX platform:
ORACLE_HOME/bin/rwrun.sh report=test.rdf destype=file desformat=pdf desname=test.pdf
The following error message is displayed:
Error: Couldn't find the display information
The Japanese character sets JA16SJISTILDE
and JA16EUCTILDE
are supported on other platforms (such as Solaris, Linux, Windows, and HP-UX).
To resolve this issue on AIX platforms, set NLS_LANG
to JA16SJIS
or JA16EUC
instead of JA16SJISTILDE
or JA16EUCTILDE
. However, if this workaround is implemented, the Java application issue of FULLWIDTH TILDE cannot be solved. FULLWIDTH TILDE is displayed as WAVE DASH in Oracle Reports.
This section describes vendor-specific issues and their workarounds for Oracle Reports. It includes the following topics:
Section 18.2.1, "Underlines Do Not Display in HTMLCSS Output in Browser"
Section 18.2.2, "JSP-Based Web Report with Large Number of Columns Generates JSP Compilation Error"
Section 18.2.3, "Java AWT Windows Hangs with Specific Graphics Cards"
In HTMLCSS output, underlines may not display in the browser if you use a custom style for a field object and apply both background color and underline as part of the style, as shown in the following example:
.ReportLevel {text-decoration:underline;color:blue; background-color:yellow}
This is a browser limitation. The browser cannot display underlines, with absolute positioning, for the fields that have both underline and background color applied to them.
When you run a JSP-based Web report with more than 149 columns, you may encounter the oracle.jsp.provider.JspCompileException
error, as shown in the following example:
500 Internal Server Error OracleJSP: oracle.jsp.provider.JspCompileException: Errors ... 6764 code too large for try statement catch( Throwable e) {18 code too large public void _jspService(HttpServletRequest request, HttpServletResponse response) throws java.io.IOException, ServletException}
This occurs due to a limitation in the Java language that does not enable compilation of Java files with large methods.
When you start Reports Server with rwserver.exe
, the UI may not display and Reports Server may hang on computers with specific graphics cards (for example, ATI Rage XL PCI card). This occurs due to a problem with the UI mechanism used in Java.
As a workaround, start Reports Server with the JVMOPTIONS
command line keyword set as shown in the following example:
rwserver SERVER=test JVMOPTIONS="-Dsun.java2d.noddraw=true"
Note: It is recommended that you start Reports Server from OPMN. |
This section describes Globalization Support issues and their workarounds for Oracle Reports. It includes the following topics:
Section 18.3.1, "Images Do Not Display Correctly in Turkish Environments"
Section 18.3.2, "MS Mincho Font in PostScript Font Mapping Does Not Work When NLS_LANG is JA16EUC"
Section 18.3.3, "Multibyte Reports Containing Oracle Object Types Hang Reports Builder"
Section 18.3.4, "Japanese and Korean Translated UI Strings Are Truncated in the Report Wizard"
Section 18.3.6, "ENVID Ignored While Detecting Duplicate Job"
If you are running reports that include images and use advanced imaging in Turkish environments (when NLS_LANG
is set to TURKISH_TURKEY.TR8MSWIN125
or TURKISH_TURKEY.WE8ISO8859P9
), then set the environment variable REPORTS_OUTPUTIMAGEFORMAT=[PNG|GIF]
.
Alternatively, you can control the value in the reports request using the command line keyword OUTPUTIMAGEFORMAT=[PNG|GIF]
.
When you deploy reports created with Reports Builder on Windows to a Solaris or LINUX Reports Server that has NLS_LANG
set to JA16EUC
, the PDF font mappings will not work. This happens if any one of the following fonts is used in the PostScript font mapping entry:
"<MS Mincho in JP>"..... "<MS PMincho in JP>".....
To work around this issue, you must perform one of the following tasks:
Set NLS_LANG
to JA16SJIS
.
Use MS Gothic
instead of MS Mincho
in PostScript font mappings when creating the reports.
Reports Builder may stop responding when you run a multibyte report containing an embedded Oracle Object type.
This is a known issue.
In the Report Wizard in Reports Builder, the UI strings Query Builder... and Import SQL Query... are truncated on the Solaris platform when used in a Japanese and Korean environment. To resolve this issue, you must edit the Tk2Motif*fontList
setting in ORACLE_HOME
/guicommon/tk/admin/JA/Tk2Motif.rgb
, as follows:
For Japanese, change:
Tk2Motif*fontList: -misc-fixed-medium-r-normal--14-130-75-75-*-*-*-*:
To
Tk2Motif*fontList: -dt-gothic-medium-r-normal--14-*-*-*-m-*-jisx0208.1983-0:
For Korean, change:
Tk2Motif*fontList: -*-*-medium-r-normal--16-*-*-*-*-*-*-*:
To
Tk2Motif*fontList: -*-*-medium-r-normal--14-*-*-*-*-*-*-*:
Oracle Reports status messages on the Job Status page in Application Server Control may contain invalid characters if the middle tier character set, NLS_CHARACTERSET
(the third field of NLS_LANG
) does not correspond to the encoding of the Reports Servlet response for the Servlet locale.
Note: The middle tier's default encoding can be different from the Servlet's encoding. For example, when you setLANG to ja or ja_JP.eucJP , the default middle tier system encoding is EUC . However, in this locale, Reports Servlet messages are encoded in Shift_JIS . This is because default encoding for ja_JP is Shift_JIS in the servlet.
|
Oracle Enterprise Manager 10g converts the messages from the middle tier's NLS_CHARACTERSET
to UTF8
resulting in invalid characters in the status messages on the Job Status page. For example, the Reports Servlet messages are encoded in Shift_JIS
when the middle tier runs with LANG
set to ja
and NLS_LANG
set to JAPANESE_JAPAN.JA16EUC
. However, Oracle Enterprise Manager 10g assumes that the original message is in JA16EUC
, and converts it to UTF8
, which results in the display of invalid characters in the status messages.
When detecting a new incoming request for a duplicate job where the job request includes the TOLERANCE
keyword, Reports Server ignores the value of the
ENVID
keyword. As a result, the job is marked duplicate of a previous job that was submitted with the same values for all keywords except ENVID
.
The following example illustrates this issue:
Report Request | Report | ENVID | TOLERANCE | Output |
---|---|---|---|---|
1 | Japanese.rdf
|
JA
|
10
|
As expected |
2 | Arabic.rdf
|
UTF
|
10
|
As expected |
3 | Arabic.rdf
|
AR
|
10
|
Same as request 2 (incorrect) |
4 | Japanese.rdf
|
JA
|
10
|
Same as request 1 (correct) |
In this example, request 4 is correctly marked as a duplicate of request 1 as both the requests use the same values for all keywords, including the ENVID
keyword. However, request 3 is incorrectly marked as duplicate of request 2 because both requests use the same values for all keywords, except the ENVID
keyword. As a result, the changed ENVID
will not be accepted in request 3.
To work around this issue, either do not specify TOLERANCE
on the command line or specify a dummy user parameter in the job request to differentiate the job requests in duplicate job detection.