Apache OpenOffice (AOO) Bugzilla – Issue 18826
Merge process trashes StringArray ItemList resources
Last modified: 2013-08-07 15:00:08 UTC
I have found that the Hungarian locale is marked as Iceland in some non-english OOo localazations. I have found it after I install the Hungarian dictionary and Tools->Options->Language Settings->Writing Aids shows that that the Iceland dictionary was added. Note: It is correct in the English localization. But it is broken for example in the Czech localization. I have searched sources and the bug probably is in svx/source/dialog/langtab.src I think that nearly all translations are shifted between "Hebrew" and "korea (Johab)" For example the Dutch translation: < "Grieks" ; GREEK ; > ; < "Hebreeuws" ; GUJARATI ; > ; < "Hindi" ; HEBREW ; > ; < "Hongaars" ; HINDI ; > ; < "Ijslands" ; HUNGARIAN ; > ; < "Indonesisch" ; ICELANDIC ; > ; < "Italiaans (Italië)" ; INDONESIAN ; > ; < "Italiaans (Zwitserland)" ; ITALIAN ; > ; < "Japans" ; ITALIAN_SWISS ; > ; < "Kannada" ; JAPANESE ; > ; < "Kasjmir" ; KANNADA ; > ; < "Kasjmir (India)" ; KASHMIRI ; > ; < "Kazachstaans" ; KASHMIRI_INDIA ; > ; < "Konkani" ; KAZAK ; > ; < "Koreaans" ; KONKANI ; > ; < "Koreaans (Johab)" ; KOREAN ; > ; < "Lets" ; LATVIAN ; > ;
Hi, Unfortunately, this problem is not under lingucomponent. I am shifting this to the sd/graphics project since I belive they own the module svx in the hopes they will know what is going or and can redirect to someone else. Kevin
Hi, I think graphics/drawing owns svx but I am not sure. Kevin
Czech localization is buggy only in CVS. Our current version is OK. Petr don't you want to again use our GSI file?
Reassigned to Christian. Please have a look what can be done here. Thanks.
Thomas, is this your file?
TL->NF: to you.
This is one of those issues we are not able to finally solve. The problem is within the current mechanism of extracting and merging strings within string lists. The items are numbered within the gsi files and once a item is inserted or removed, the order of translations is missalligned with the source languages German and English/US. Fixing this depends on merging up to date gsi files. As we permanentally do so when we receive updated translations, I would like to keep this issue as a reminder to change the numbering mechanism of the transex tool. Any objections? If yes, please let me know.
Ivo, can you please keep this issue. Once you have no other issues on your list (will this ever happen ;-)), we can discuss another numbering theme for string lists.
This is a severe problem. As soon as language builds other than English or German are involved, the pairings of the string and the underlying value of an item are messed up. You mostly even don't notice at UI level because you only get a list of strings presented, but if the user selects a language the value of a different language is passsed to the application instead. The same effect holds true for other StringArray ItemList resources, e.g. txenctab.src The situation now got even worse, because in the localize.sdf you don't see the paired value anymore and there's just the offset number where the string will be wrongly inserted, so this isn't even fixable anymore after a database merge. This gotta be fixed for OOo2.0 with high priority => changing target and prio, adapting summary, and component to l10n.
fixed in ivo06
I've tested the behaviour, looks good
.