Created attachment 94334 [details] The attachment is an 'ods' file with a sample stack chart with 'Stock Chart 4' option. Besides the chart the file also includes two print screens for comparison. Problem description: When charting using Stack Chart 4 (LO 4.2.0.4) the candles are not formatted/displayed properly. It works just fine for the older versions (4.1.5). Steps to reproduce: 1. Open the attachment using version LO 4.2.0.4. 2. Open the attachment using version LO 4.1.5.1. 3. Compare. Current behavior: The Stack Chart 4 (LO 4.2.0.4) the candles are not formatted/displayed properly. Expected behavior: The Stack Chart 4 (LO 4.2.0.4) the candles should be formatted/displayed properly. Operating System: Mac OS X Version: 4.2.0.4 release Last worked in: 4.1.5.1 rc
Opened the .odt file with LibO 4.2.0.4 on Win7: - between the underlying table and the chart specification, "low" and "high" are interverted, - it results in the predictable wrong display (the rectangle top is "low" value and the rectangle height is "low"-"high", negative). The question remains on when the change occured: - is the interversion a wrong behaviour new to Lib0 4.2? - or was the interversion present in your .odt file already in 4.1, and LibO had an automatic feature to reaffect min value to "low" and max value to "hight", feature that had been abandonned for a reason I don't know about? Because you ensure that the SAME file is right when opened with LibO 4.2.0.4 and wrong when opened with LibO 4.1.5.1, we can conclude to the second hypothesis. The reason for no more reaffectation of "low" and "high" values when rectangle shows "negative height" seems legitim IMHO. So not a bug, but an error in your .odt that LibO doesn't correct anymore. Kept status to UNCONFIRMED to get other analysis.
Created attachment 94553 [details] The proper LOW-HIGH sequence .ods file. This .ods file has proper OPEN-LOW-HIGH-CLOSE sequence. The result is the same as the other .ods attached file with OPEN-HIGH-LOW-CLOSE sequence.
(In reply to comment #1) > Opened the .odt file with LibO 4.2.0.4 on Win7: > - between the underlying table and the chart specification, "low" and "high" > are interverted, > - it results in the predictable wrong display (the rectangle top is "low" > value and the rectangle height is "low"-"high", negative). > > The question remains on when the change occured: > - is the interversion a wrong behaviour new to Lib0 4.2? > - or was the interversion present in your .odt file already in 4.1, and LibO > had an automatic feature to reaffect min value to "low" and max value to > "hight", feature that had been abandonned for a reason I don't know about? > > Because you ensure that the SAME file is right when opened with LibO 4.2.0.4 > and wrong when opened with LibO 4.1.5.1, we can conclude to the second > hypothesis. The reason for no more reaffectation of "low" and "high" values > when rectangle shows "negative height" seems legitim IMHO. > > So not a bug, but an error in your .odt that LibO doesn't correct anymore. > > Kept status to UNCONFIRMED to get other analysis. Sorry, I should have mentioned that I tried both ways HIGH-LOW and LOW-HIGH. The result is the same: both of them work just fine when I opened them using LO 4.1.5 and both of them do not work when I opened them using LO 4.2.0. Please see the other attachment. The behavior is identical. So, answering your questions: - is the interversion a wrong behaviour new to Lib0 4.2? ANSWER: Yes, it is a new behavior to LO 4.2. - or was the interversion present in your .odt file already in 4.1, and LibO had an automatic feature to reaffect min value to "low" and max value to "hight", feature that had been abandonned for a reason I don't know about? ANSWER: the "interversion" was present in some of my .ods files, but it does not make any difference. Whether I have the "interversion" or not the result is the same: wrong. Please see the second attachment where I reversed the "interversion", which did not make any difference. The result is still identical. Proper for 4.1.5 and wrong for 4.2.0.
Thanks to your answer. OK, I reproduce the bug after interverting Low and High (LibO 4.2.1.1 on Win7) So status set to NEW
*** This bug has been marked as a duplicate of bug 75075 ***
Migrating Whiteboard tags to Keywords: (possibleRegression) [NinjaEdit]