To comment on the following update, log in, then open the issue: http://www.openoffice.org/issues/show_bug.cgi?id=113568 Issue #|113568 Summary|Silent corruption of user dictionaries Component|framework Version|DEV300m84 Platform|PC URL| OS/Version|All Status|UNCONFIRMED Status whiteboard| Keywords| Resolution| Issue type|DEFECT Priority|P3 Subcomponent|code Assigned to|tm Reported by|jurf
------- Additional comments from j...@openoffice.org Sun Aug 1 07:43:05 +0000 2010 ------- OOo-dev330m85 and probably OOO330m2 may corrupt .dic user dictionaries when user adds words. Buggy behaviour is not present in 3.2.1, hence regression. DESCRIPTION When adding words to a user dictionary (right-click menu), one of four things may happen: 1. word is added normally, speller behaves as expected; 2. word is added to *another* user dictionary (not the one selected, and not necessarily active); 3. only part of the word is added, appended at the end of the .dic (tested on UTF8-versioned dic files); 4. word is not added at all, but part of the last word in a .dic (not necessarily the one selected) is repeated as the new last entry. The buggy behaviour was first noticed on an updated install (using settings, including user dictionaries, created in older versions of OOo), but also occurred on a clean installation with default user settings. In that case, words I'd supposedly added to standard.dic were actually added to the ignore list. Having seen similar bugs in a variety of programs (garbage appearing at the end of incorrectly processed files, paragraphs or text selections, or even in other non-selected paragraphs or files), I'd guess the problem is caused by a text/bound parser that isn't counting straight when it encounters anything other than single-byte characters, which in turn causes miscalculated insert positions and data corruption while sorting fields etc. I'm quite confident about this guess as the same build I was testing (OOo-dev330m85) also has data-corrupting flaws in its new text-casing options, which also relies on calculating text bounds - see Issue 113558. For the user dictionary routine, I noticed the problem surfaced when multi-byte characters (such as a letter with unusual accents, for which there are no precomposed versions in Unicode) were present, either in the word being added, or - fatally - already included in a user dictionary (not necessarily the one selected to receive the word). I'm afraid I got fed up with M85 and reverted to 3.2.1 before completing testing. As such, I'm only guessing that the bug might also be triggered by the presence of ligatures, or by trailing periods (OOo's spell-checker normally ignores trailing periods, such that "A.D." is entered as "A.D", but this appears to have buggy for a while), or indeed by anything else that might trip up a mathematically-challenged parser. If you need a quick and dirty test word, copy/paste aaaṣbbb (I know, not a word, but it will at least appear at the top of your word lists to save you a scroll - the s has a dot underneath, by the way). --------------------------------------------------------------------- Please do not reply to this automatically generated notification from Issue Tracker. Please log onto the website and enter your comments. http://qa.openoffice.org/issue_handling/project_issues.html#notification --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@framework.openoffice.org For additional commands, e-mail: issues-h...@framework.openoffice.org --------------------------------------------------------------------- To unsubscribe, e-mail: allbugs-unsubscr...@openoffice.org For additional commands, e-mail: allbugs-h...@openoffice.org