Apache OpenOffice (AOO) Bugzilla – Issue 96133
i18npool: ambiguous && ||
Last modified: 2013-08-07 15:02:43 UTC
in i18npool/source/breakiterator/breakiteratorImpl.cxx we have in case WordType::WORD_COUNT: while (nPos > 0 && (u_isUWhiteSpace(ch = Text.iterateCodePoints(&pos, -1))) || isZWSP(ch)) nPos=pos; operator precedence of && before || gives this the meaning of while ((nPos > 0 && (u_isUWhiteSpace(ch = Text.iterateCodePoints(&pos, -1)))) || isZWSP(ch)) nPos=pos; (as attachment one does to make it explicit) but looking at the *other* case statements that looks a little suspicious, the other similar ones do... while (nPos > 0 && (u_isWhitespace(ch = Text.iterateCodePoints(&pos, -1)) || isZWSP(ch))) nPos=pos; so the intent *might* have been as in attachment 2 [details] to make the logic equivalent to the other types
Created attachment 57920 [details] keeps the logic as it is now, but makes it explicit
Created attachment 57921 [details] changes current behaviour, but might be what was intended
reassign
*** Issue 96830 has been marked as a duplicate of this issue. ***
Karl, please review and apply the appropriate solution to your next CWS. If you review and tell me before 10:00 UTC tomorrow (Thursday) I can as well add that to my CWS locales31.
It was my typo. It should be same like other case. Fixed in i18n49.
Test case can be found in i96133 (https://so-web.germany.sun.com/iBIS/servlet/edit.ControlPanel?tid=i96133)
Sorry, test case can be found in https://so-web.germany.sun.com/iBIS/servlet/edit.ControlPanel?tid=i96372
*** Issue 96372 has been marked as a duplicate of this issue. ***
ready for QA.
Verified in CWS i18n49.
seen in OOO310_m9 and DEV300_m45