empty mrk nodes found in translated xlf from Trados 2011

[jennlong]

We are using Flare 9 and Lingo 7. Our translator uses SDL Trados 2011, they have all the latest updates, and the Madcap filetypes (http://www.translationzone.com/en/opene ... ?appid=345 and http://www.translationzone.com/en/opene ... ?appid=342

We cannot merge the files back into Lingo. During the reimport merge, in Lingo we see the error message:

The XLIFF in nameoffile.htm.xlf is no longer valid, so you will need to contact the translator to repair it: Empty <mrk> nodes found.

The problem in encountered in all of the files.

Does anyone know specifically how to fix those files in Trados, or how to export them from Trados so that I can get them back into Lingo?

Thank you!
Jenn

[alt_jennifer]

Hi Jenn,

Well, I've never used those specific filtetypes before, because we've made our own custom filters for Trados Studio 2011 here in my office. And I haven't received that exact error, but when working with Lingo 5, I did have problems with malformed XLIFF not importing into Lingo (with translation done in Trados Studio 2011). Here's what happened.

The file was a glossary file. The phrase had a colon in it, like "See Also: Conversion Type". In Lingo this was all one segment but in Trados it was broken down into 2 segments, at the colon. The <mrk> tags were visible in the first segment, and the translator had put the closing </mrk> tag before the colon. Which left the colon out on its own, outside of any <mrk> tags.

When the XLIFF was exported from Trados, there was a second, identical set of <mrk> tags placed around the second phrase "Conversion Type". It had the same mid number as the first mrk tags. I assume it was the duplicate mid number and not the colon outside of the tags that caused the problem, but I don't know for sure. When I fixed that error in Trados, moved the colon inside the </mrk> tag, all was resolved.

To find this problem, though, it took awhile. I used the binary chop approach, which involves removing half of the translated glossary file (being careful not to cause any new tag errors) and trying to import it into Lingo. If it has no errors, then you know that the problem is in the half you deleted (so you would reverse it then -- keep only that bad half in the file and delete the tested, good half). If you still have an error when importing that first half, then you know the error is there. I kept doing that, narrowing the file down by half until I found the right glossary term that was causing the problem.

I recommend that you (1) do a general search for </mrk>: to see if you find a similar error (2) also potentially search for genuinely empty mrk tags, with no content in between <mrk> and </mrk> and (3) if still nothing, try the binary chop approach on a smaller file that's causing problems.

Hopefully this helps! Let me know if you need anything else.

[kwag_myers]

I've just encountered this same issue and thought I'd post what I found. Each XLIFF file contains three versions of each segment. These are the <source>, <seg-source>, and <target> (which is the translation). The latter two versions have the <mrk> tags alt_jennifer mentioned. In my case, there was also text outside the <mrk> tags. Here's how I found and corrected each:

1. Open the XLF file in NotePad.
2. Ctrl+F and Find the <mrk> tag with a space before it: " <mrk". The tag is suppose to next to <target>, e.g., <target state="needs-translation"><mrk...

XLIFF_Invalid3.png

In this case, "Perguntas" (Portuguese for "question") and the space that follows must be cut and pasted after the </bpt> tag of the same line.