Page MenuHomePhabricator

Punctuation mark leaks from Lang-* template when translated with Apertium
Open, MediumPublic


When translating the Kleisoura (Byzantine district) from English to Spanish, the Lang-el template on the first paragraph is adapted when added to the translation (i.e., the equivalent template in Spanish is added with the information). However, an additional ":" character is added after the template in the target language:

Screenshot 2020-04-24 at 13.18.37.png (92×845 px, 33 KB)

This extra ":" character seems to come from the rendering of the source template. The way a template is rendered should not interfere with the contents, so it is unclear how the ":" leaked from there into the target contents.

This happens with Apertium (and possibly with other plain text MT systems).

This quick link allows to test the issue based on this example page.

Event Timeline

Pginer-WMF renamed this task from Punctuation mark leaks from template when translated with Apertium to Punctuation mark leaks from Lang-* template when translated with Apertium.Apr 24 2020, 11:26 AM
Pginer-WMF triaged this task as Medium priority.
Pginer-WMF created this task.
Pginer-WMF moved this task from Needs Triage to Bugs on the ContentTranslation board.