<html> <head> <base href="https://bugs.documentfoundation.org/"> </head> <body> <div> <a class="bz_bug_link bz_status_NEW " title="NEW - calc: storing of rounded values and recalculation on load harming display after reload | was: calc: save-load-cycle: some cells shown with wrong value until recalc (F9)" href="https://bugs.documentfoundation.org/show_bug.cgi?id=140722#c10">Comment # 10</a> on <a class="bz_bug_link bz_status_NEW " title="NEW - calc: storing of rounded values and recalculation on load harming display after reload | was: calc: save-load-cycle: some cells shown with wrong value until recalc (F9)" href="https://bugs.documentfoundation.org/show_bug.cgi?id=140722">bug 140722</a> from <a class="email" href="mailto:newbie-02@gmx.de" title="b. <newbie-02@gmx.de>"> b.</a> <pre>Created <a href="https://bugs.documentfoundation.org/attachment.cgi?id=173078" name="attach_173078" title="file_showing_another_flavour_of_the_same_problem">attachment 173078</a> <a href="https://bugs.documentfoundation.org/attachment.cgi?id=173078&action=edit" title="file_showing_another_flavour_of_the_same_problem">[details]</a> file_showing_another_flavour_of_the_same_problem this file is loaded / shown with seemingly correct content, but the value? loaded for B6 is now a real 0,30000000000000000000 instead of the correct calculation result of 0,30000000000000004~~~ resulting effects: recalc of C6 results in '0', and changes back to 5,55E-17 after any recalc of B6. B6 is stored in the content.xml file as: 'table:formula="of:= [.B4] + [.B5]" office:value-type="float" office:value="0.3" calcext:value-type="float"><text:p>0,30000000000000000000</text:p>', changing that to: 'table:formula="of:= [.B4] + [.B5]" office:value-type="float" office:value="0.30000000000000004" calcext:value-type="float"><text:p>0,30000000000000004000</text:p>' resulted in a little delay on recalc for C6, but the irritating effect is gone. thus evaluation of precise values in content.xml works, it's just the storing which harms? above scheme might be root cause for other save - load - recalc issues. injecting inaccuracies and ambiguities already on file load complicates understanding for users and debugging for supporters and developers. The referential integrity is undermined. The effect occuring rare and unpredictable makes that even worse. proposal: don't use the 'prettyfied' display values to save in file, but a string which uniquely identifies the fp double value valid for that cell, this sample needs 17 digits. (values in the 'weak precision ranges' ([0,5 .. 1[, [8 .. 10[ etc.) would be sufficiently identified with less digits, but not harmed by too many.) proposal II: do the same - storing the 'correct string' - for the 'text' of the cell, it will avoid irritations and we might need it in very near future when doing decimal correct calculations. if nobody want's to work on it, but somebody could give me a code pointer i'll give it a try ...</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the assignee for the bug.</li> </ul> </body> </html>