How to handle preprocess2 errors "preprocess2.topic-fragment"

[stefan-jung]

Hi, we have recently upgraded our XSL:FO/FOP publishing to map-first processing aka preprocess2. This has fixed the issues we had with a combination of branch filtering, key scopes and content key references. However, this also has introduced completely new errors. One of those errors is this one:

[preprocess2.topic-fragment] Failed to process XML filter: 
  Failed to transform file:/tmp/temp20241009160559798/foo-bar/72f172eac52dd18e675027a700188d7f77936470.ditamap: 
  /tmp/temp20241009160559798/foo-bar/72f172eac52dd18e675027a700188d7f77936470.ditamap (No such file or directory)

This is undocumented and does not have a proper DITA-OT error code. Has anyone seen this before and can help me to understand and fix this problem?

[jelovirt]

Does running OT with --verbose or --debug log any additional information? Are you able to share the source that's causing the error?

[stefan-jung]

@jelovirt , I've analyzed this further. I'm (mis-)using this discussion to log my findings. This error occurs five times in this particular publication.

[preprocess2.topic-fragment] Failed to process XML filter: Failed to transform 
C:\Users\eike\workspace\dometic-docs\AC and Ventilation\Marine Cabin Control\VMD2.5\temp\topic-pdf\oxygen_dita_temp\0d178ef43bf9cf1995263137851f38cb894156d0.ditamap (The system cannot find the file specified)
C:\Users\eike\workspace\dometic-docs\AC and Ventilation\Marine Cabin Control\VMD2.5\temp\topic-pdf\oxygen_dita_temp\3870121184c251457b4a4a018147e2dcdae6e8a5.ditamap (The system cannot find the file specified)
C:\Users\eike\workspace\dometic-docs\AC and Ventilation\Marine Cabin Control\VMD2.5\temp\topic-pdf\oxygen_dita_temp\519b48ff0c011f09ee99081e8899b5c025e5737f.ditamap (The system cannot find the file specified)
C:\Users\eike\workspace\dometic-docs\AC and Ventilation\Marine Cabin Control\VMD2.5\temp\topic-pdf\oxygen_dita_temp\b24a939607e56f16c4608141e8fcbdd1ebd5a76b.ditamap (The system cannot find the file specified)
C:\Users\eike\workspace\dometic-docs\AC and Ventilation\Marine Cabin Control\VMD2.5\temp\topic-pdf\oxygen_dita_temp\b3b31cbfffda712bd9586eb8ac831c1d43b728c4.ditamap (The system cannot find the file specified)

When checking the .job.xml, I see that this is always the same file, our Subject Scheme Map. The document contains 5 languages (1 keyscope per language). The Subject Scheme Map is referenced 6 times. Once from the root map, and once for every language submap to have the key definitions available on submap level. This is common practice and needed for Oxygen XML and Fluenta. Another factor which might be relevant here is, that the Subject Scheme Map is not in the same folder structure as the bookmap.

The bookmap is here:

file:/C:/Users/eike/workspace/dometic-docs/AC%20and%20Ventilation/Marine%20Cabin%20Control/VMD2.5/DometicVMD2-5_SIO_4445104332_GLOBAL5.ditamap

And the Subject Scheme Map is here:

file:/C:/Users/eike/workspace/dometic-docs/_profiling/SBJ_General.ditamap

This is common practice and is intentionally.

The problem now is, that the DITA-OT incorrectly calculates the @result attributes for the Subject Scheme Map.

<file src="file:/C:/Users/eike/workspace/dometic-docs/_profiling/SBJ_General.ditamap"
            uri="0d178ef43bf9cf1995263137851f38cb894156d0.ditamap"
            path="0d178ef43bf9cf1995263137851f38cb894156d0.ditamap"
            result="file:/C:/Users/eike/workspace/dometic-docs/_profiling/SBJ_General-5.ditamap"
            format="ditamap" has-keyref="true" subjectscheme="true"/>

<file src="file:/C:/Users/eike/workspace/dometic-docs/_profiling/SBJ_General.ditamap"
            uri="3870121184c251457b4a4a018147e2dcdae6e8a5.ditamap"
            path="3870121184c251457b4a4a018147e2dcdae6e8a5.ditamap"
            result="file:/C:/Users/eike/workspace/dometic-docs/_profiling/SBJ_General-3.ditamap"
            format="ditamap" has-keyref="true" subjectscheme="true"/>

<file src="file:/C:/Users/eike/workspace/dometic-docs/_profiling/SBJ_General.ditamap"
            uri="519b48ff0c011f09ee99081e8899b5c025e5737f.ditamap"
            path="519b48ff0c011f09ee99081e8899b5c025e5737f.ditamap"
            result="file:/C:/Users/eike/workspace/dometic-docs/_profiling/SBJ_General-4.ditamap"
            format="ditamap" has-keyref="true" subjectscheme="true"/>

<file src="file:/C:/Users/eike/workspace/dometic-docs/_profiling/SBJ_General.ditamap"
            uri="b24a939607e56f16c4608141e8fcbdd1ebd5a76b.ditamap"
            path="b24a939607e56f16c4608141e8fcbdd1ebd5a76b.ditamap"
            result="file:/C:/Users/eike/workspace/dometic-docs/_profiling/SBJ_General-1.ditamap"
            format="ditamap" has-keyref="true" subjectscheme="true"/>

<file src="file:/C:/Users/eike/workspace/dometic-docs/_profiling/SBJ_General.ditamap"
            uri="b3b31cbfffda712bd9586eb8ac831c1d43b728c4.ditamap"
            path="b3b31cbfffda712bd9586eb8ac831c1d43b728c4.ditamap"
            result="file:/C:/Users/eike/workspace/dometic-docs/_profiling/SBJ_General-2.ditamap"
            format="ditamap" has-keyref="true" subjectscheme="true"/>

My rough guess is, that the DITA-OT might have problems collecting ressources which are outside of the bookmap directory.