Apache OpenOffice (AOO) Bugzilla – Issue 7689
Odd behaviors of the spellchecker while using 2 dictionaries and a single affix file
Last modified: 2012-02-20 02:24:23 UTC
Recently, the Italian dictionary is released with an additional dictionary called comuni.dic (attached to this issue) that includes the names of all cities of Italy. It is suggested to register it under the same locale of the main Italian dictionary with: DICT it IT comuni During the spell checking phase, if a word of comuni.dic is misspelled, OOo underlines it as such, but when you right click on the word, the context menu suggests words included in the main Italian dictionary only. When comuni.it is used as main dictionary, OOo suggests the correct words included in it. This fact should show that when 2 dictionaries are used with the same locale/affix file, OOo is able to handle the words of the second dictionary, but not the automatic suggestions for the corrections of a misspelled word.
It seems there is some problems with IZ and I'm not able to add attachments... However you can find comuni.dic included in the package you find here: http://prdownloads.sourceforge.net/ooodocs/it_IT_1.6.zip?download
Hi, Two things. 1. Every dictionary must have its own properly named affix file. Differnt dictionaries even for the same locale may have or use differnt encoding, use differnt TRY strings, etc. So if you have a comuni.dic you need a comuni.aff file with the proper settings. 2. Order is important when it comes to suggestions. Put the main dictioanry first and any other additional dictionaries second. If a word is mispelled in both, it tries to generate suggestions from the main (first) dictionary. If suggestions are found and they exceed some threashold (I have to see the code to remember the threshhold) suggestions from secondary dictionaries are not requested since the most common suggestions from the main dictioanry are most often right. If fewer than the threshold suggestions are found after looking through the main dictionary, then suggestions are requested for the additional or sub dictionaries. I can change the threshold to something higher but it will slow down spellchecking for systems with lots of dictioanries. The threshhold might be as low as 1 now but I could easily raise that to 2 or 5 if that makes sense. Kevin You can control the order
Hi, I've tested the threshold and it seems to be set to 1 because only the second dictionary words that have no suggestions in the first one are suggested correctly in the context menu. So, you should really raise the threshold to 5 (?) if possible.
Hi, Okay, I will change the code to raise the threshhold. I will post here when the change is done. Kevin
Hi, I decided to skip the threshhold completely sinc eit is arbitrary. I will now return all suggestions from all dictionaries registered for that locale. The implications for the dictionary maker is to make sure if they have multiple dictionaries registered for that locale that the intersection of the words in those dictionaries is 0 (i.e. they do not have any words in common). This should not be an issue since there is no reason to have two dictionaries with the exact same words (it wastes memory, time, cycles, etc). In the worst case, if the dictioanries do overlap, you may get duplicate suggestions (the same suggestion twice (once from each dictionary it appears in). This is of course not fatal! ;-) I will test this change and then commit it to the tree. It should appear in the next release after 643. Thanks, Kevin
this is fixed in the latest OO643C lngucomponet. Thanks, Kevin
The Issue you raised has been marked as 'Resolved' and not updated within the last 1 year+. I am therefore setting this issue to 'Verified' as the first step towards Closing it. If you feel this is incorrect, please re-open the issue and add any comments. Many thanks, Andrew Cleaning-up and Closing old Issues ~ The Grand Bug Squash, pre v3 ~ http://marketing.openoffice.org/3.0/announcementbeta.html
As per previous posting: Verified -> Closed. A Closed Issue is a Happy Issue (TM). Regards, Andrew