The Word documents that were error-ing out as reported in this support thread now are converting correctly after WORDSCLOUD-631 was closed.
We are, however, still getting Word-to-PDF 400 conversion errors today. These documents view correctly within Word itself. The error Aspose returns is slightly different this time:
{
"Message": "Error while loading file 'd7324fac-ee8c-45ac-8da2-1d0ca3cbbbd6.doc' from storage: The given key was not present in the dictionary."
}
As in the previous case, these documents return the same new error we got in production when I submitted them manually today on Aspose.Words Cloud - API References
We are able to reproduce the issue on our end. The issue has been logged in our internal tracking software as WORDSCLOUD-641 and we are trying hard to fix it at the earliest.
The documents you shared have relative URLs to images. For example, the document 2df23327-1481-4655-9424-5180d21b6594.doc is containing e2fe112b-a776-465e-8a1d-f54ea79787bc_files/image002.gif relative URL.
In order for the API to work correctly, you need to upload these referenced images to the Cloud Storage as well. Please ensure the referenced images are located at the same level as the input document.
We don’t have full control of the Word documents being sent to us. They do not include the image files referenced.
However this conversion failure is seemingly new behavior for us.
Previously, documents would convert through Aspose if the images were missing. Is there a flag we can set in the API call to have it revert to the previous behavior of ignoring missing image references in Word docs?