#1014 Tag soup in XML/HTML files

[msoutopico]

Steps to reproduce

I created two OmegaT projects to translate the XSLT file attached (inside the packages).

In Okapi: On the one hand, I have created the okf_xml filter custom configuration in Rainbow and then I have created the OmegaT project directly from Rainbow. Project: xslt_rfe_okpkit_omtprj

In OmegaT: On the other hand I have created the project in OmegaT and added the same filter custom configuration to the Okapi XML filter (file okf_xml@xslt.fprm inside the omegat/ folder). Project: xslt_rfe_omtprj

Expected results

The source text displayed to the user for translation is the same in both projects. The text contains no leading/trailing tags that do not need to be translated and all segments contain at least some translatable text.

Actual results

In both projects extraction of translatable text and protection of inline codes look correct, however the OmegaT project created in OmegaT has many unnecessary tags whereas the OmegaT project created in Okapi Rainbow does not.

Some segments only include one or many tags, so those segments should not exist at all (because they don't need to be translated and exposing the markup entails a risk to the integrity of the file). Also, node opening/closing markup should not produce leading and trailing tags, which don't need to change position in the target language.

You can see a screenshot (Okapi on the left, OmegaT on the right):

Or here: https://imgur.com/a/CokitXr

Additional info

This issue reminds what happens with other file types (for example, when using the OOXML filter in OmegaT, the translatable segments include leading and trailing tags (corresponding to beginning and end of paragraph, I think) should don't need to be there and are a pain to translate, whereas they are not included when using Okapi's OOXML filter.

Files

Both project packages attached. They can be unpacked from OmegaT if you have the OMT plugin installed.

[Aaron Madlon-Kay]

The xslt_rfe_okpkit_omtprj project uses OmegaT's built-in XLIFF filter.

The xslt_rfe_omtprj project uses the Okapi plugin's XLIFF filter.

If you don't like the results of the latter, that's an issue for the Okapi plugin, not for OmegaT.

[Briac Pilpré]

reduced image size

[msoutopico]

Indeed it seems I got this wrong, the results are not that different between the two projects: node's leading/trailing tags are extracted in both bases (my settings were hiding them in project xslt_rfe_okpkit_omtprj -- duh!). I'll report this to the Okapi team, this ticket can be closed. Thank you for looking into it so quickly.