Background of the issue:
I was trying to make small changes to some pages that have been translated into dozens of languages. I noticed that ATE had changed its interface and suggested retranslation of many text fragments due to different tag interpretation.
Symptoms:
ATE suggests retranslation of almost half the text because it can't correctly interpret tags in already translated texts.
Questions:
Why does ATE suggest retranslation of fragments with tags?
How can I prevent ATE from misinterpreting tags in already translated texts?
ATE requires not only a retranslation but also an indication of what should be included in the a-tag (in this example it is 'Application').
Another example, where the translation is lost, is as follows:
By <a href="<em><u>hidden link</u></em>...">judgment of 4 February 2020</a>, the ECHR found that the applicant had been tortured by the police and awarded him compensation of €39,700.</p>
<p><a href="<em><u>hidden link</u></em>...">Applicant's written observations</a><br>(Russian version)
Firstly, this fragment is merged, even though there are separate paragraphs here. Secondly, not only are the translations lost, but, as in the previous example, ATE asks to specify what should be included in the a-tag.
Translations of all fragments in the text with this arrangement are lost. This is true not only in the new version of ATE but also in the old version in which the translation was done.
The problem concerns old translations, i.e. situations where changes are made that do not alter the original text of such translations. If new text with the same layout is added to the same page, it will be processed correctly by ATE, including being segmented correctly, in the same way as the text was segmented in the original translation. In this example, this means that 'Application' will be a separate fragment, '(In Russian)' will be a separate fragment, everything before and after it will be separate, and ATE will not require any tagging.
The wpml_invalid_tag tag is placed by ATE where there is a <br> tag. I think this is important because one of the problems with translating Contact Form 7 contact forms involves exactly this tag. See my other ticket: https://wpml.org/forums/topic/wpcf7-translation/
Languages: English (English )Spanish (Español )Italian (Italiano )
Timezone: Europe/Rome (GMT+01:00)
Hello,
I guess when you mean "tag" you mean the HTML markers, right? you changed them and it basically retranslated them?
Also, you mention that they are merged. this leads me to believe that you were using "smaller segments" in the past. could you try to see if you can get them back in one of the pages where you had them?
To do that, access the translation editor on the interested page, click on the gear icon on the top right section of your site and disable "Bigger Segments".
Check if you like that mode better.
Bigger segments is good when your site needs more context to translate correctly. however if you translate manually, review constantly or overall modify your content a lot, maybe shorter segments is the way to go here.
If you change a marker, that will also make the content disappear for now, because the translation is not faithful to the new source, however we're working on a way to improve this specific scenario.
Let me know a page where that happened so I can also come and check a bit better on our end the entire situation.
No, I made a single minor change by moving a 'technical' element a few lines up, nothing else. I didn't change any tags, let alone the text. However, when I logged into ATE to apply this change in another language, I found that many of the translations had been removed, ATE requires the fragments to be retranslated. All the fragments that require translation contain some form of html tags, particularly a-tags and br-tags. For example, it could be text inside an a-tag. It could be text before or after a br-tag.
When I say they are merged, I mean that ATE now sees them in such a way that it is not possible to split them, i.e. ATE does not offer a split (no paperclip button). Previously, as far as I can remember, the parts of these fragments that it now sees as merged were separate. I don't usually combine texts into larger chunks for translation unless there is a compelling reason to do so. In this case, there could hardly have been one. So I think the pieces were separate.
I had previously disabled the ability to merge into larger chunks in the preferences.
I am afraid that almost all my translations (again, into dozens of languages) are now in this state, literally stuffed with various tags.
As for the br-tag, if I remember correctly, ATE used to see the text before it as a separate fragment and the text after it as a separate fragment. And that's how translations were done. Now ATE sees it as a single text with the br-tag in it, and as untranslated (requiring translation) text, and not only requiring translation, but also requiring an indication of where exactly the br-tag is located (of course, ATE does not require you to select the tag itself, which it does not display, but the text before and (or) after it, which determines where the tag is located). Also, of course, a feature of the br tag is that it can be adjacent to the text before and (or) after it, without spaces. This is probably important. Again, the layout of the texts was the same. And ATE processed and translated them correctly.
Yes indeed, if this feature is disabled before any changes are made to the original, or if more changes are made to the original later, everything will work as before.
I hope this feature doesn't become irreversible, at least not with so many bugs and problems with it.
