To avoid double spaces I can run the corresponding QA, I know. To make it faster, I replaced any double space with a single space with the Find dialogue. But this means that any space next to a tag is going to be deleted, no matter what the tag means.
Go to the chapter 1Introduction2,
Go to the chapter1Introduction2,
I can only guess that CT interpretes the tags as non-breaking spaces (but they aren't).
Some non-formatting tags carry a space to place the segment correctly in the translation memory. For example, when you join two segments separated by a paragraph or a table cell, CT joins such two segments placing such a tag at the joining point. Then, when you add the joined segment to a TM, CT removes the tag leaving a space at this point, otherwise the two segments would have been merged without any spacing in between.
I did not explain the issue correctly.
The example above does not refer to any TM-related problem. It is just that CT deletes the space before the tag (and in this example these tags are tags around a field inside Word pointing to a header with the value "Introduction" – they are not tags appearing after joining 2 segments) when replacing double tags by a single tag with the Find-command (instead of doing the corresponding QA and replace them manually). The tag itself is not deleted, of course. And the exported Word doc does not contain the space (it simply has "Go to the chapter1Introduction2"), so I assume the tag did not contain any space either.