Do you happen to know what does “Error 112: Name expected” error message mean?
Sometimes, when I want to run a QA check, Xbench 2.9 displays this error message during loading the files. I usually work with XLIFF files (extracted from Google Translator Toolkit) alongside with a tab-delimited text file used as glossary for key terms. If this happens, Xbench is able to process the file partially, until the faulty segment(?), I suppose, but the rest of the file is omitted. I tried to look at the XLIFF file many times, but I can see no difference between those segments that are sure to be good (because they are processed), and those that are not processed.
If you do not have access to Xbench 3.0, you can try to find the offending segment with a dihcotomic search, by dividing in 2 part the XLIFF file until you find the segment that is failing. However if you have a file with 1,000 segments it might take 10 divides in the worst case.
All right, it seems that I found out what the problem is. This answer might help those who encounter the same issue.
I have made a proofreading job today, and the translator provided his XLIFF files, which were created with a Chrome extension very-very similar to the one that I use (the only difference—at least I believed it is the only one—is that my extension is able to exclude segments based on their status: CM, 100%, fuzzy, etc.). I compared the two versions in Notepad++, and noticed that the standalone & characters are missing from the translator’s files. I can only assume that his extension removes this character for some reason, while mine does not do this. I can also only assume that Xbench treats the & character as some kind of parameter, and expects a value (name?) for them. If there’s no value, Xbench will fail.
After removing these standalone & characters, I was able to run a proper Xbench check. I’m not sure though how Xbench behaves with strings like & and >. It is yet to be confirmed whether these can also cause problems or not.
Per your description, it seems that such extension created an invalid XML file. XLIFF is an XML format so Xbench expects XLIFF files to be a valid XML files.
The XML specification requires that standalone ampersands (&) are escaped as the & entity.
Probably if you try to load these files with standalone ampersands in some other application that reads them, such as Chrome or Firefox, you would also get a validity error.
There are other characters that XML requires to escaped, such as lower than (<), which should be escaped as the < entity.
So it seems that Error 112: Name expected means that you are missing the name of the entity after the &.