In case of translate5 whitespace protection, the original whitespace should be recreated before sending it to OpenTM2 (as long as it is in line with xliff spec).

When OpenTM2 returns the the match, translate5 should

• in case of 100%-matches convert back the whitespace
• in case of Fuzzy-Matches: convert back the whitespace and check, if all contained Whitespace-Protection-Tags are identical with the source tags. Where this is not the case, the whitespace-Protection tag in the Fuzzy-Match must be deleted. This way it is ensured, that no whitespace is eliminated by the Fuzzy-Match, because if a tag misses, translate5 has to add it to the end of the segment anyway on taking over the match
  • dealing with additional tags created by the whitespace parser is already implemented through T5DEV-95.