nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

ReportEngine genreport Run and Render Causes Extra Blank Pages

gsgsoln

I'm hoping this is not a bug and is instead a mistake that I'm making that someone can help me resolve.
 
My environment:
<ul><li>Windows 7 Enterprise, SP1</li>
<li>Birt 422 (birt-runtime-4_2_2
 (Eclipse Juno build 20130225-0426, although not related to issue)
</li>
<li>
Upgrading from 2.5.2, where report works as expected
</li>
</ul> 
My issue:
<ul><li>When running genreport.bat (or .sh on AIX), using the -m run AND render mode (two step), generating a PDF file, I am getting extra blank page(s) on the PDF that is being caused by a label at the bottom of the report design where the label's text is a long string (set in onCreate) BUT the label's visibility is set to hide it from PDF</li>
</ul> 
What I've tested/experienced:
<ul><li>Issue does not occur when using the runrender mode (single step) 
We produce multiple output types from a single rptdesign so for performance reasons, using multiple runrender is not feasible</li>
<li>Issue does not occur when using "view report as PDF" in Eclipse (tested there just to be sure)</li>
<li>Issue does not occur if the label's text is reduced/shortened so that it does not span a "page"</li>
<li>Number of extra pages seems to correspond with the amount of "data" that is associated with the label</li>
<li>Issue does not occur if the label's visibility is set to hide for all outputs</li>
<li>Issue does not occur if I put the label between two tables that have data</li>
<li>Issue DOES occur if I put another element (label, text, dyn text) below it regardless if it has data or not or how it's visibility is set</li>
<li>Tried changing the label's this.text dynamically within it's onRender script, but no luck</li>
<li>Tried dropping the label from within the beforeRender script, but no luck</li>
<li>Changing the label to text or dynamic text produced the same results</li>
</ul> 
How you can recreate:
<ul><li>I've attached a sample report design and a copy of the genreport.bat that I'm using</li>
<li>There is no data source, no data set, no parameters, no masterpage, just a simple design with two grids (full visibility) containing labels and the label (hidden for PDF)</li>
<li>Download and unzip birt-runtime-4_2_2.zip (eclipse.org/downloads/download.php?file=/birt/downloads/drops/R-R1-4_2_2-201302161152/birt-runtime-4_2_2.zip)</li>
<li>Execute using
<ul><li>genreport -m run blank_page_genreport.rptdesign</li>
<li>genreport -m render -f PDF -o c:\temp\blank_page_genreport.pdf blank_page_genreport.rptdocument</li>
</ul></li>
<li>In the onCreate for the label with text "Grid label 1 - text set in onCreate script", remove the last occurrence of strData, which shortens the data to fit on a "page", thus generating no blank page. Expanding the data will produce more blank pages.</li>
<li>Alternatively, setting the label's visibility to hidden for all outputs will produce no blank page(s)</li>
</ul><div> </div>
Any assistance that can be provided will be GREATLY appreciated. Thanks in advance.

Find more posts tagged with

Comments

gsgsoln

Tried the below in the onRender event for the label but they did *not* resolve the issue.
 
<div>if (reportContext.getRenderOption().getOptions().get("format") == "pdf") {</div>
<div style="margin-left:40px;">this.text = "";</div>
<div>}</div>
<div> </div>
<div>if (outputFormat == Packages.org.eclipse.birt.report.engine.api.IRenderOption.OUTPUT_FORMAT_PDF) {</div>
<div style="margin-left:40px;">this.text = "";</div>
<div>}</div>

mwilliams

Try adding this to your beforeFactory method:
 
<pre class="_prettyXprint _linenums:0">if(reportContext.getHttpServletRequest().getAttribute("attributeBean").format == "pdf"){
reportContext.getDesignHandle().findElement("lbl_TextData").drop();
}
</pre>

gsgsoln

Thanks for the quick response Michael. Applied code to the beforeFactory but still got extra blank page.
 
To maximize performance, something our clients have come to expect, our process from using BIRT 252 was to run once and render twice (PDF and CSV). From what I've been able to determine, when the run task is executed by itself using "genreport -m run" the output format defaults to "pdf" even though none has been specified. This you can get from using reportContext.getOutputFormat() in initialize or beforeFactory.
 
If we drop the element at run time, then there won't be anything there when I go to render the CSV output (or any other type other than PDF). This is essentially what I'm struggling to resolve. I need to element, and it's data, there for the second render phase (genreport -m render -f CSV) but I don't want it to stay hidden when I render for the PDF (genreport -m render -f PDF). 
 
The PDF rendering is rendering pagination of an element that's hidden for PDF, but it's not rendering the data. That seems like a bug for it to be able to create pages for a type-specific hidden element but without the data.
 
Any other thoughts?

mwilliams

It's definitely a bug. I just tested with the designer as it was showing the same issue. It's possible that genReport doesn't recognize the code I used above. Is there any way you could download BIRT 4.4 and test with it to see if you get the same issue? If you do, please file a bug for this at eclipse.org/birt/community. Have you tried passing the format in genreport -m run -f CSV to see if that helps with the automatically passing pdf issue?

gsgsoln

Thanks again Michael. I've definitely tried using "genreport -m run -f CSV" and it did nothing. Doesn't appear that the -f parameter is recognized on the single run phase.
 
I'll download BIRT 4.4, test it out, and let you know the results.

mwilliams

Thanks. I'll let you know if I figure something else out.

gsgsoln

Downloaded birt-runtime-4_4_0-20140611.zip, configured the genreport.bat file, tested, and still got the extra blank page.
 
As we've been working internally to brainstorm a solution, a colleague of mine found this bug report: <a data-ipb='nomediaparse' href='https://bugs.eclipse.org/bugs/show_bug.cgi?id=331541'>https://bugs.eclipse.org/bugs/show_bug.cgi?id=331541</a>
 
It looks to be essentially the same issue, where the resolution was RESOLVED WONTFIX. 
 
Do you agree that it's a similar issue? If so, would the next step be an enhancement request?
Just trying to determine if I should stay on the path of trying to find a solution or move towards our worst case scenario contingency plan.

mwilliams

It's a similar issue, but I think there is still merit to you filing a bug report for the point that you said the format is PDF for the run task, no matter what. If this were resolved, it would seem that you'd be able to properly drop the element for PDF only and have it available for CSV.

mwilliams

Are you tied to using separate run and render for some reason? Your issue doesn't seem to happen with runAndRender.

gsgsoln

Again, thank you Michael. The worst case scenario contingency plan that I mentioned earlier would be to switch to using the single runAndRender task, but that would effectively double our processing time since all output files, both PDF and CSV in my case, would require re-execution of the entire report which would be a re-execution of all the report's queries. We generate over 4000 output files every month and at already takes a few days to complete. Doubling that isn't ideal since it then interferes with other time sensitive processing.
 
I will log a bug and hopefully get a better resolution than "WONTFIX."
 
Many, Many thanks for you insight and you quick responses Michael. Greatly appreciated.

mwilliams

In the meantime another option could be to duplicate the report without the text box and call that design when going to PDF format. Not a great workaround, but might help you to save the time.

gsgsoln

That's effectively what we're going to end up doing for all 200+ reports that we have.
 
Thanks again for your continued insight.

mwilliams

You're welcome. Sorry there wasn't a better solution. I'll let you know if I think of something else.

gsgsoln

Bug report created. Please vote for it if you find this issue a problem for you.
 
id=444142
 
<a data-ipb='nomediaparse' href='https://bugs.eclipse.org/bugs/show_bug.cgi'>https://bugs.eclipse.org/bugs/show_bug.cgi</a> ? id=444142

mwilliams

Thanks for the update!