Question Marks and Missing pieces in XML documents

[Guest]

Hi there,

I am using FM 10.  I have set up a conversion table to apply DITA structure to documents, with the intent of saving them as XML docs.  Unfortunately, after I do this, the XML docs are coming out a little strange...with ? added, with text and tables missing, etc.

Here is what the structure looks like before I save as XML:

Then, I save as XML, reopen in FM and get the following:

I can look at the XML directly and see that the culprit is a little box character that must not be interpreted correctly when FM reopens the XML doc:

All I don't understand is why that character is appearing in the XML to begin with.  I don't have anything - even a space - in the structured doc in that location. Any ideas as to why this is occuring?  This isn't a huge deal in the example XML file, but other XML files I've saved via this process are littered with ? or sometimes altogether missing entire elements (like the whole Title element has disappeared).

Hannah

[Michael_Müller-Hillebrand · Accepted solution]

Hannah,

Such characters can be the result of pasting text from application with CR+LF line-ends. FrameMaker internally uses only the LF character and so spurious CR can be left over.

To test if this is the case, you can

• search in your original documents for \x0d  (backslash, x, zero, d: the hex code for Carriage Return) and if found replace them with nothing.

• or, do a MIF wash before saving as XML: Save the document as MIF, then open this MIF document and proceed as you originally intended.

But maybe it is something different,

- Michael

[Guest]

Great, thank you.  I did a search on the characters "\x0d" in the original and did find them all over the place, and replaced them with nothing.  Unfortunately, the issue remained when I saved as XML.  However, a MIF wash worked perfectly.  Now, to figure out how to mass-save my documents as MIF, since I anticipate having to break up this guide into 50-60 file pieces and would love to not have to save them 100-120 times.

[Guest]

Ah...of course.  Doing the MIF wash on the original, unchunked files structred files works just as well as doing it on the individual chunk.    So, that cuts it down to 8-9 times, a more manageable process.  Thanks again, Michael!