Bug #16854
closedSpellchecker split words at some special chars (e.g. german umlauts)
0%
Description
If you try to spell-check the string "Salatköpfe", the spellchecker-popup regards the text as two individual words "Salat" and "köpfe" (but displayed as one word). See additional information and patch for further information.
Here some information about the system where the error occurs:
Fedora Core release 6 (Zod)
php-5.1.6-3.3.fc6
typo3 4.0.4 with rtehtmlarea 1.3.8
aspell-0.60.3-7.1
I tracked down the problem to the xml_set_character_data_handler php-function. If there is an umlaut the xml-parser return the string up to the umlaut and then calls the character-data-handler a second time with the rest of the string. So it looks we can not assume the handler returns always character-data in one chunk.
If you apply the patch the character-data-handler only collects the character chunks. The real spellCheckHandler is called from startHandler and endHandler, if there is already collected text available. So the chunks are glued together before the spell-check starts.
(issue imported from #M4777)
Files