Skip to main content
New Idea

Localization XLIFF files truncates sentences

Related products:PX Engagements
  • May 29, 2025
  • 3 replies
  • 58 views
V

If styling is applied to a sentence in the engagement (bold text) text strings are truncated into different lines when exporting the language XLIFF file.


Truncating sentences is bad for localization as different languages require different sentence structure/order. For example, for some languages the bold text could be at the beginning of the sentence rather than in the middle.

At the moment we need to manually adjust the placement of bold and regular text in the downloaded XLIFF file, and it’s wasting a lot of our time and patience. 

3 replies

rschlette
Forum|alt.badge.img+2
  • rschlette
  • Expert ⭐️
  • May 30, 2025

Hi Veronica. That behavior should be unique to the ‘contextual’ XLIFF download. If the HTML XLIFF download would meet your needs, you should check that file format. I don’t think you’ll see those mid-sentence breaks for formatting in the HTML type.

download dropdown from Localization module

 

rschlette
Forum|alt.badge.img+2
  • rschlette
  • Expert ⭐️
  • May 30, 2025

btw, if you’re getting the XLIFF via the PX REST API, the ‘strategy’ URI param is the one to set for contextual vs html

M
  • mikethetexan
  • Contributor ⭐️
  • July 29, 2025

If using the HTML XLIFF flavor, you will need a TMS/CAT or MT tool which can apply an HTML cascading/subfilter, as the raw HTML content is copied into the translation unit of the XLIFF without any parsing.

For some customers, it will be possible to handle it that way, but for other customers, it won’t be.

For example, in Phrase TMS:
 

HTML subfilter in TMS




Is GainSight looking at a clean transformation of the HTML content into the XLIFF, with HTML tags turned into <g> or <bpt>/<ept> XLIFF representations, so that you do not need a system with an HTML cascading filter?

Thank you.

Terms & ConditionsCookie settingsAccessibility statement