generating canonical uri breaks on CV with titles having '-'

[cessda-bitbucket-importer]

Original report on BitBucket by Stefan Dlugolinsky (GitHub: Stifo).

---

getCanonicalUri method is using the removeLanguageInformation method, which counts on - character as an indicator of a language code followed by it. This does not work with CV titles having -. Moreover, this function has hard coded condition (< 20).

[cessda-bitbucket-importer]

Original comment by Martin Šeleng (GitHub: pakoselo).

---

@john-shepherdson @‌dolinarm If you look to the latest version of the VersionDTO.java: https://github.com/cessda/cessda.cvs.two/blob/master/src/main/java/eu/cessda/cvs/service/dto/VersionDTO.java it should be somehow solved. Stefan was doing some cleaning there because of the 3 digit versioning system. I will try to test it for while and it maybe broken on some other place. Is there any URL I can test to against it?

[cessda-bitbucket-importer]

Original comment by Martin Šeleng (GitHub: pakoselo).

---

@john-shepherdson I had test it for a while and didn’t see the described problem .

[cessda-bitbucket-importer]

Original comment by Martin Šeleng (GitHub: pakoselo).

---

@john-shepherdson @‌dolinarm This part of code is used mainly for generating pdf and html outputs. Should be also already solved by @Stifo . My question is as follows: There is automatically removed the last digit and replaced by 0. As the 0 means SL. Probably it still holds true for the 3 digit version system, but this time, when TL Admin adds new TL version, the last digit will be increased not only for TL, but for SL as well. So the question is, if we want to generate the URI (canonical, URN and so on), with the real patch number (third digit, means patch) or with zero as the main version?

[cessda-bitbucket-importer]

Original comment by Stefan Dlugolinsky (GitHub: Stifo).

---

related to #484