Hi Kendall,

Kendall Shaw said:

But if I create output that is a zip file, the zip contains files from a different publication.

A classic question is, what has changed recently. Furthermore if only the "ZIP" output format is affected, then I wonder if something got stuck there (hardcoded for debugging, always pointing to a faulty server, etc)

If none of the other OutputFormats have this, then breaking up the problem in smaller pieces is the way to go.

-Dave

Thanks, Elizabeth and Dave,

This is for a throw away publication used to test styles. The baseline contains a single map and a single topic. It existed for about 45 seconds before the problem was exhibited.

The publication was create after another throw away publication exhibited this behavior. That was another single map single topic, publication. But, the content was copied from a topic in another publication, with GUIDs changed. That topic did reference images from the publication the content was copied from. The second had all of those references removed.

A suspicious part, maybe, is that the content contained elements, which were being processed by the transtype in the style processor and failing internally, but both the DITA-OT build and the CMS publish would succeed and the event log showed the publish as succeeding.

Now the part that was failing internally, no longer fails. But, the result is the same for those publications.

I recreated several publications to see if new map and topic would change the result, but it did not.

Thanks, Elizabeth and Dave,

This is for a throw away publication used to test styles. The baseline contains a single map and a single topic. It existed for about 45 seconds before the problem was exhibited.

The publication was create after another throw away publication exhibited this behavior. That was another single map single topic, publication. But, the content was copied from a topic in another publication, with GUIDs changed. That topic did reference images from the publication the content was copied from. The second had all of those references removed.

A suspicious part, maybe, is that the content contained elements, which were being processed by the transtype in the style processor and failing internally, but both the DITA-OT build and the CMS publish would succeed and the event log showed the publish as succeeding.

Now the part that was failing internally, no longer fails. But, the result is the same for those publications.

I recreated several publications to see if new map and topic would change the result, but it did not.