More supported syntax

[Juergen-aus-Zuendorf]

Hi,

great plugin !!!

I have tested it today the first time and found some syntax elements that are not supported:

Normal tables:
| //Basismaschine://\\ siehe Spielplatz | wird zu | //Basismaschine://\\ See playground |
--> "Basismaschine" isn't translated

imagecaption plugin::
Hier ist ein Bild mit Unterschrift:\\ <imgcaption image1|Unterschrift zum Bild>{{bild.jpg|}}</imgcaption>
wird zu:
Here is a picture with a signature:\\ <imgcaption image1|Unterschrift zum Bild>{{playground:bild.jpg|}}</imgcaption>
--> The text "Unterschrift zum Bild" isn't translated

Best regards
Juergen

[NetaliDev]

I wil try to fix this but it is always difficult to exclude or include specific syntax things from specific plugins.

[NetaliDev]

But I know the reason for your table issue:

If you write | //Basismaschine://\\ siehe Spielplatz | it will think that Basismaschine:// is a hyperlink that will be excluded from the translation. But you can write // Basismaschine: // and it will work (it is important that there is whitespace between : and //.

[Juergen-aus-Zuendorf]

Ok, I understand the problem. But why is Basismaschine:// interpreted as Link? In Dokuwiki-Syntax there is nothing defined like that. And on the other hand, when I am in the Dokuwiki editor and want to define text as italic by menu this text is automatically included in slashes – without whitespace.

[NetaliDev]

Because you can also paste plain links in your WikiText and I need something to detect them. So I have used the :// because it is included in every link like http://example.com or https://example.com or even ftp://example.com.

[NetaliDev]

But you can also write //Basismaschine:// \\ because it only interprets it as a link if there are multiple non-whitespace characters in front of and behind :// and in this case there would be whitespace behind it and so it is not detected as a link by the plugin.

[splitbrain]

Instead of using regexes or string replacements to exclude parts of the wiki text from translation, the DokuWiki parser should be used. That parser knows what is a link and what isn't. deepl's API supports an XML mode AFAIK. A wiki text could be parsed and rendered to a custom XML representation, sent to DeepL and the resulting XML could be converted back to wiki syntax without any ambiguities.

[NetaliDev]

You already suggested that to me, but I didn't find an easy way to convert the output of the parser back into DokuWiki syntax (and I now work in full time and don't really have the time to refactor the whole escaping mechanism).