PDF builds successfully but doesn't exist

[JBJ]

I'm a novice user. I am not using customizations.

On Friday, my DITA output to PDF worked fine. Monday morning, it still built successfully (no errors, says the PDF was created), but only a back-level version of the PDF opened. If I remove the back-level PDF from the Out folder. The transform is still successful but it does not create a new PDF (with changes). I also tried removing the entire Out folder and letting Oxygen rebuild it from scratch. On transform, it rebuilt the folder, but still no PDF is there.

It also started creating several other occurrences of the output directories in other places on my local. I removed all of those output directories and tried again. Again, build is successful, Out folder is created...but, no PDF.

If I take a back-level PDF and drop it into the Out folder, again, build is successful and the PDF opens, but it's the same back-level version. It won't overwrite.

Every other transform works fine (XHTML, WebHelp). This is only happening with PDF.

I also tried upgrading to 13.2 hoping that might jiggle it loose. But that didn't work either.

Oh, and I have a deadline tomorrow, naturally...

Any help is greatly appreciated!!

[Radu]

Hi,

Oxygen uses the DITA Open Toolkit to publish DITA content to various output sources.
What do you mean by back-level PDF?
After the DITA transformation is over, in the "DITA Transformation" console tab you have the logging output of the entire publishing stage.
Could you send us the contents of this console view to the technical support email address "support@oxygenxml.com"?

Also, in the DITA Maps Manager view there is a toolbar action called "Validate and check for completeness". You can invoke it to check for linking and other types of problems in your DITA Map.

Regards,
Radu

[JBJ]

Hi Radu,

The PDF that opens is an old version of the PDF. It does not contain any changes that have been made to the DITA map or associated files since the last successful transform on Friday. Transform to WebHelp yields updated content, so I know there is no problem with the actual map or files. A completeness check yields no errors and no warnings. I have refreshed references several times. I have closed the DITA map, restarted Oxygen, and rebooted several times.

Again, I changed no settings between Friday evening and Monday morning. It worked fine on Friday (would generate and open PDFs of the current file content) but would not generate an updated PDF on Monday.

I'm sending the transform reports for: 1) WebHelp transform (no errors/no warnings); 2) PDF transform (now it keeps generating random overwrite errors in the Transformation Problems dialog, but it still says the build was successful; and each time I run the PDF transform, it generates a different set of overwrite errors); 3) The Transformation Problems report

Many thanks!

[Radu]

Hi Julia,

Thanks for the additional details that you sent to the support email address.
I'm replying this post because maybe the conversation would be useful to other users as well.

So based on your console log when publishing to PDF you seem to encounter two types of problems:

1)

org.apache.fop.fo.ValidationException: The column-number or number of cells in the row overflows the number of fo:table-columns specified for the table. (See position 370:-1)

This problem usually indicates that one of your DITA tables contains more cells in a row than the number of table columns.
This makes the Apache FO processor break when building the PDF file from the XSL-FO.

Either you can look at all your tables in the DITA project and see if one of them has such a layout problem or:

If you edit the transformation scenario, in the "Parameters" list there is a parameter called "retain.topic.fo" which you can set to "yes" and also a parameter called "clean.temp" which you can set to "no".

Then you can try to transform to PDF and after the transformation fails you can look inside the C:\Documents and Settings\Julia McClung\My Documents\InContext\Projects_Clients\Pangen\CTBlazer_NC1006\Documentation\Users Guide\temp\topic.fo XSL-FO file after the transformation finishes at line 370 and somehow get more info about which table failed to be processed.

2)

[DOTJ009E] Cannot overwrite file 'C:\Documents and Settings\Julia McClung\My Documents\InContext\Projects_Clients\Pangen\CTBlazer_NC1006\Documentation\Users Guide\temp\Users Guide\EDC_SiteImport.xml' with file 'C:\Documents and Settings\Julia McClung\My Documents\InContext\Projects_Clients\Pangen\CTBlazer_NC1006\Documentation\Users Guide\temp\Users Guide\EDC_SiteImport.xml.temp'. The modified result may not be consumed by the following steps. Check to see whether the file is locked by some other application during the transformation process.

This indeed means that somehow the publishing process could not overwrite a certain topic in the temporary files directory at some stage. But without a sample DITA project to reproduce the problem on my side I cannot be certain of why this happens, this could be a bug in the DITA Open Toolkit. But this problem does not make the publishing fail to generate the PDF, it's more of a warning.

Regards,
Radu

[Radu]

For problem number (1) I managed to reproduce the situation by using a table like this:

Code: Select all

<table frame="all" id="table_3tc_qbh_hf">

                <title>TEST</title>

                <tgroup cols="2">

                    <colspec colname="c2" colnum="2" colwidth="1.0*"/>

                    <thead>

                        <row>

                            <entry/>

                            <entry/>

                        </row>

                    </thead>

                    <tbody>

                        <row>

                            <entry/>

                            <entry/>

                        </row>

                        <row>

                            <entry/>

                            <entry/>

                        </row>

                        <row>

                            <entry/>

                            <entry/>

                        </row>

                    </tbody>

                </tgroup>

            </table>

So as you see the table has 2 entries on each row but only one <colspec> defined. This might be similar to the problem that you are having.

[JBJ]

Thank you so much, Radu! Oddly, late last night, the PDF began building again. I will investigate the tables and see if correcting that clears things up permanently.

Yes, the overwrite errors come and go. Most of the time, there are no Transformation Problems listed. But when they do occur, it's almost never the same files referenced in the error. And they occur even when the PDF builds with no problem.