Pathauto 6.x-2.x-dev translates the title »Von und mit Tieren« (2010) into the URL path %3E%3Evon-und-mit-tieren%3C%3C-2010 which makes absolutely no sense.

6.x-2.x-dev utilizes the 'transliteration' module to convert certain characters into safe code, in this case e.g. >>von-und-mit-tieren<<-2010. However, the >> and << might be a valid transliteration of » and «, but none of it is wanted in an URL.

I would assume that ./admin/build/path/settings should allow to strip out these unwanted characters with the "Punctuation" settings ("Remove <" and "Remove >"), but for whatever reason this does not work, at least not as expected, and this I'd consider either a bug or a usability problem. If 'transliteration' outputs >>, "Remove >" should remove it IMHO.

(A workaround is to check "Reduce strings to letters and numbers", but I'm not sure if forcing this is really wanted)

Thanks & greetings, -asb

Comments

greggles’s picture

greggles
he/him
Primary language English
Location Denver, Colorado, USA
commented
Title: Entities in URL » Make transliteration first, punctuation cleanup second

I think this is probably an order of operations kind of problem. Transliteration runs right after punctuation is fixed.

It would be simple to reverse the order. I'm not sure if that would introduce new problems - it seems harmless enough to me.

dave reid’s picture

dave reid
he/him
Primary language English
Location Nebraska USA
commented
Status: Active » Closed (duplicate)

Duplicate issue of #408420: Transliterization should occur before punctuation cleanup