Docx4J.pdfViaDocuments4jLocal() searches for unknown class

[Lars]

Hello.

I'm running version 8.2.8 and tried the local conversion by following the getting started guide.

I added the the dependencies:

Code: Select all

  // Docx4J
  implementation 'org.docx4j:docx4j-core:8.2.8'
  implementation 'org.docx4j:docx4j-openxml-objects:8.2.8'
  implementation 'org.docx4j:docx4j-openxml-objects-pml:8.2.8'
  implementation 'org.docx4j:docx4j-openxml-objects-sml:8.2.8'
  // MOXy docx-JAXB Impl
  implementation 'org.docx4j:docx4j-JAXB-MOXy:8.2.8'
  implementation 'org.docx4j:docx4j-documents4j-local:8.2.8'


Docx4j itself works fine.. but PDF converting not at all.

The method Docx4J.pdfViaDocuments4jLocal(); internally looks for the class "org.docx4j.convert.out.documents4j.local.Documents4jLocalExporter" but this class is nowhere to be found in the class path. I also did not find another dependency on maven that is missing.

What dependency is missing? Where can i find it?

Kind regards
Lars

[jason]

Hi Lars

Thanks for the pickup. In the meantime, you can invoke Documents4jLocalServices directly; see the examples at https://github.com/plutext/docx4j/tree/ ... ts4j/local

cheers .. Jason

[Lars]

Hi Jason.
Thank you for you replay.

Unfortunately this does not work eigther.

I'm doing this now:

Code: Select all

      Documents4jLocalServices exporter = new Documents4jLocalServices();
      exporter.export(wordprocessingMlPackage, targetStream);


This produces this Exception when invoking the new Documents4jLocalServices() C'tor:

Code: Select all

java.lang.IllegalStateException: class com.documents4j.conversion.msoffice.MicrosoftWordBridge could not be created by a (File, long, TimeUnit) constructor
with cause:
java.lang.reflect.InvocationTargetException
class com.documents4j.conversion.msoffice.MicrosoftWordBridge could not be created by a (File, long, TimeUnit) constructor

java.lang.NoSuchMethodError: com.google.common.base.Preconditions.checkArgument(ZLjava/lang/String;Ljava/lang/Object;Ljava/lang/Object;)V


For me it looks like the code is using:
com.google.common.base.Preconditions
but may be it should use:
org.docx4j.com.google.common.base.Preconditions

Or do I just have the wrong version?

the import is in this class:

Code: Select all

package com.documents4j.job;
[..]
import static com.google.common.base.Preconditions.checkState;
[..]
public abstract class ConverterAdapter implements IConverter {
[..]
    private static File makeTemporaryFolder(File baseFolder) {
        File tempFileFolder = new File(baseFolder, UUID.randomUUID().toString());
        checkState(tempFileFolder.mkdir(), String.format("Cannot create folder: %s", tempFileFolder));
        return tempFileFolder;
    }
[..]
}


And indeed, in google-collections 1.0 there is no methode checkArgument(String,String,Object);

Is this the problem? Which version of google collection/guava should i use?


Can you please help?

kind regards
Lars

[jason]

mvn dependency:tree reports:

Code: Select all

[INFO] |  +- com.documents4j:documents4j-local:jar:1.1.3:compile
[INFO] |  |  +- com.documents4j:documents4j-api:jar:1.1.3:compile
[INFO] |  |  +- com.documents4j:documents4j-transformer:jar:1.1.3:compile
[INFO] |  |  +- com.documents4j:documents4j-util-conversion:jar:1.1.3:compile
[INFO] |  |  |  \- javax.activation:activation:jar:1.1.1:compile
[INFO] |  |  +- com.google.guava:guava:jar:28.1-jre:compile
[INFO] |  |  |  +- com.google.guava:failureaccess:jar:1.0.1:compile
[INFO] |  |  |  +- com.google.guava:listenablefuture:jar:9999.0-empty-to-avoid-conflict-with-guava:compile
[INFO] |  |  |  +- com.google.code.findbugs:jsr305:jar:3.0.2:compile
[INFO] |  |  |  +- com.google.j2objc:j2objc-annotations:jar:1.3:compile
[INFO] |  |  |  \- org.codehaus.mojo:animal-sniffer-annotations:jar:1.18:compile
[INFO] |  |  \- org.zeroturnaround:zt-exec:jar:1.11:compile
[INFO] |  \- com.documents4j:documents4j-transformer-msoffice-word:jar:1.1.3:compile
[INFO] |     +- com.documents4j:documents4j-transformer-msoffice-base:jar:1.1.3:compile
[INFO] |     +- com.documents4j:documents4j-transformer-api:jar:1.1.3:compile
[INFO] |     |  \- com.documents4j:documents4j-util-all:jar:1.1.3:compile
[INFO] |     \- com.documents4j:documents4j-util-transformer-process:jar:1.1.3:compile


Do you have those jars?

[Lars]

Hi Jason.

Yes, we got all of this.

The problem is deeper in the class path.

We have dependenciey that have google-collections as a dependency. Docx4J need a similar class but in a newer version provided by guava.

The problem is that in runtime the wildfly chooses the first match it finds in the class path and that is Preconditions class from google-collections not from google-guava. I do not know why wildfly is working like that because it brings guava on board so from the specification it should load guava first. I already tried to specify the dependency in the ear, but also did not work.

Do you thing it is possible to provide a version of docx4j where this check could be optional.

I see no other way to fix this in our environment

kind regards
Lars

[jason]

Hi Lars

package com.documents4j.job is part of documents4j, not docx4j

https://github.com/documents4j/documents4j

I see they have just release v1.1.7, I guess that if anything, this uses a still newer guava.

Can your other dependency be made to use more recent google stuff?

Before you go to too much effort here, I take it you have satisfied yourself that documents4j does what you want (ie in a dev environment without the other jars causing issues).

[Lars]

Hi Jason.

The Unit-Test results are looking great, it just not runs on the server.
We definetly want to use it like this.

>>> Can your other dependency be made to use more recent google stuff?
No. There are also 3rd Party, we can not update them without a much bigger effort beacause we are talking about "old" core features like sqlserver connectors and those things.

Are you suggesting to download and change the documents4j code?

kind regards
Lars

[jason]

Altering the documents4j code is one approach; you could then make a PR.

But first, consider https://github.com/plutext/docx4j/tree/ ... s4j-remote

With this, you could run documents4j server in a separate webapp in wildfly, and hopefully avoid the jar hell you are experiencing (I haven't checked).

Sample code is at https://github.com/plutext/docx4j/tree/ ... s4j-remote