Description: hello all, libreoffice calculates 1 plus 1 to be 3, it's not! a matter of rounding, i couldn't stand waiting if and when someone will fix tdf#123714, thus played and 'stressed' some sheets to narrow down where errors evolve, i found one effect that i consider more than critical: reproducible in plenty situations the following happens: after making a copy of a cell with a formula, and editing that copy, the originating cell is excluded from autocalculate, no matter what you do with the copy, you may even delete it from the sheet, the 'broken' source cell will always need 'F9' or 'crtl-shift-F9' to be recalculated after changes in the values of the cells referenced in the formula, thus you can see: 1 + 1 = 3 for the sum function '=SUM(A1:B1)' in C1 with '1' in A1 and changing B1 from '2' to '1'. formula, formatting, everything else is correct, even calculating works correct *when* it's initiated by F9, the only 'fault' is: the cell is excluded from autocalculate! you can make a new copy of the cell with the 'first formula', and it - the source - will be included in autocalculate again, until you edit the new copy, then autocalculating for the source cell will be broken again, (i don't know if it's affecting only distinct formulas and cells, i have the impression formulas calculating ranges fail ('=SUM(Xi:Yj)', '=MAX(A1:A15)') while formulas calculating distinct cells work ('=A1+B2' or similar) just play with it, below is one general description and under 'steps' one failing sample with concrete values) general: - check that autocalculate is on, - put a formula in one cell, - make a copy of that cell, - put some data in the sheet to be calculated by he formulas, - check the results, - change something in the area / cells used in the formula, - check that the results change correctly, - edit something in the copied formula while leaving the cell with the 'source formula' untouched, - watch that from now on changes in the area / cells used in the formulas were not! anymore reflected in correct results in the 'source-formula-cell', while the 'copied-formula-cell' still calculates correctly, that's definitively *not* how a reliable spreadsheet should work, changes producing erroneous results in cells not even touched by the change. if the above description works correct in your case pls. check the concrete sample in 'steps to reproduce'. my Version: 5.4.7.2 (x64) Steps to Reproduce: 1. check that autocalculate is on, 2. put '=SUM(A1:B1)' in C1, 3. copy C1 with 'ctrl-c', 4. insert that copy in C2 with 'ctrl-v', 5. put '1' in A1 and A2 and '2' in B1 and B2, 6. check that C1 and C2 calculate '3', 7. edit C2 from '=SUM(A2:B2)' to '=SUM(A2:B3)', do not! touch C1, 8. change B1 from '2' to '1', 9. check the value of C1, in my case it still shows '3' while calculating 1+1, 10. watch that from now on changes in the area / cells used in the formulas are not! anymore reflected in correct results in the 'source-formula-cell' C1, while the 'copied-formula-cell' C2 still calculates correctly, Actual Results: C1 isn't updated on changes of A1 or B1 Expected Results: C1 should be updated by autocalculate in the same manner as C2 Reproducible: Always User Profile Reset: No Additional Info: maybee that's the source of plenty similar bugs, maybee it can be triggered late after 'construction' of these 'shared formulas' ('landmines' in sheets causing destructions in areas far away?), maybee it's not the real fundamental fault but only an effect in a shell around it ... but, not negotiable, it should be checked and fixed as fast as possible ... Version: 5.4.7.2 (x64) Build ID: c838ef25c16710f8838b1faec480ebba495259d0 CPU threads: 8; OS: Windows 6.1; UI render: default; Locale: de-DE (de_DE); Calc: group i had to choose an 'earliest affected' version, could only tell about my actual install, i'd bet it's in since 4.2
The bugs can be rectified with the help of https://errorcode0x.com/resolve-steam-error-code-80/ which is very useful.
>>> can someone - *please* - remove the Comment #1 from perryg and stop perryg from spamming around here ... thks <<<
Created attachment 149635 [details] demo recalc sum after fomula change steps to reproduce: - open attached spreadsheet - 7. edit C2 from '=SUM(A2:B2)' to '=SUM(A2:B3)', do not! touch C1, - 8. change B1 from '2' to '1', -> C1 will not update -> a hard recalc is necessary
and to make it work without a hard recalc a save & reload cycle is necessary. this issue is not reproducible with: Version: 4.4.7.2 Build-ID: f3153a8b245191196a4b6b9abd1d0da16eead600 Gebietsschema: de_DE
narrowing down: fails looks dependent on placing the copy above or below the original, (maybe that's the situation where 'shared formulas' are created) fails looks dependent on formulas calculating a range, ('=SUM(A1:B1)' and '=MAX(A1:B1)' will fail, while '=A1+B1' and '=SUM(A1;B1)' will work), there must be some sort of 'property' with a cell that can lock it from being 'autocalculated'?, or some sort of label attached to its 'place in the sheet' saying 'autocalculate unneccessary'?, this property is hidden, out of view and access by the user, it violates the statement what autocalculate should do ('All cells are recalculated after a sheet cell has been modified.') i'd like to yell 'alarm!!!' in capitals but don't know if that's appropriate in the community, it's 'nice' that this error disappears after save - load, other similar problems don't!, this fault is trapping users and waisting time of programmers and debuggers, the problem is still present in Version: 6.2.2.0.0+ (x64) Build ID: 5f9104ef6f42d9d42ce3ec564affcba88889e76c CPU threads: 8; OS: Windows 6.1; UI render: default; VCL: win; TinderBox: Win-x86_64@42, Branch:libreoffice-6-2, Time: 2019-02-27_17:10:55 Locale: de-DE (de_DE); UI-Language: en-US Calc: threaded
please ignore Comment 4 issues *is* reproducible with lo 4.4.7.2 but not with: Version 3.6.7.2 (Build ID: e183d5b)
acc. my experiments with old versions something got definitively broken between 4.1.6.2 and 4.2.0.1, just i didn't test this fault while having these versions installed.
bibisected, but result looks a bit suspious to me (OpenGL releated issue?): /cygdrive/d/sources/bibisect/bibisect-win32-4.4 $ git bisect bad bfc26f557731dd3abb6b5cba0cb23ddf7215c03f is the first bad commit commit bfc26f557731dd3abb6b5cba0cb23ddf7215c03f Author: buildslave <buildslave@tb51.libreoffice.org> Date: Sun Apr 24 07:29:46 2016 +0200 source-hash-a55f740045d6d30e1ef5522889d2dc359ca5784a commit a55f740045d6d30e1ef5522889d2dc359ca5784a Author: Markus Mohrhard <markus.mohrhard@collabora.co.uk> AuthorDate: Thu Nov 20 09:52:03 2014 +0100 Commit: Markus Mohrhard <markus.mohrhard@collabora.co.uk> CommitDate: Thu Nov 20 13:50:51 2014 +0100 include the unx part for getting OpenGL driver & device information desktop/unx/source/glxtest.cxx is taken directly from the Mozilla project. THe whole concept is taken from Mozilla and is based on starting an early process that creates an OpenGL context. This prevents crashing drivers to crash Libreoffice. We read the information from the pipe as soon as we create the first vcl Window. In that place we then decide if the device/driver combination is blacklisted. Change-Id: I2624d4ce06d503281a4459cf3174f57cf1f7b733 :100644 100644 02019dad467cdde684ab8a5a99f68ac6764e8836 6fcd2c25d2b8a72b78d2c15904093210ff9f7ea6 M autogen.log :100644 100644 7a976663fd4b97c65156f3984293adc6a1182b4a be9630c306c3c4981acd4217f62d1a27349a9188 M commitmsg :040000 040000 35c9ddb6c58f760e9c582a64cc526c08389156b4 c4573bf85192acfbdb744df7d71582e1b6f064a7 M instdir :100644 100644 c6491b465af20406ef5df3527d5388fda9742300 c7c52f5fb0273247e2c71810116bbf32bfa332da M make.log /cygdrive/d/sources/bibisect/bibisect-win32-4.4 $ git bisect log # bad: [489ffd1df414698b6cea9ab03bf6f663b8b5af7e] source-hash-3f94c9e9ddfd807b449f3bb9b232cf2041fa12d2 # good: [8aa9fc9a0c92172593d6cd97662116a965db229d] source-hash-dea4a3b9d7182700abeb4dc756a24a9e8dea8474 git bisect start 'master' 'oldest' # good: [897913acd244cb6a5d2f4c2da1d625d9b978edb6] source-hash-ac57362b23859591c088e36b7218f4a606dcf3bb git bisect good 897913acd244cb6a5d2f4c2da1d625d9b978edb6 # good: [860f6172be84c3d801e1e533ab3d12a28815cb4c] source-hash-80c11d648e88f3195f7e4fb465cd6902bcf46867 git bisect good 860f6172be84c3d801e1e533ab3d12a28815cb4c # good: [fe0abc50dc1e45fd73635394f8df41e0dfc3eec8] source-hash-4e223fab04279c3583689e69fa1342966e81de36 git bisect good fe0abc50dc1e45fd73635394f8df41e0dfc3eec8 # good: [247d0249eb9794ccedec0ca2e3309b8d970593cc] source-hash-1535f39388223de53e4b923c6f7bb71ee32c1858 git bisect good 247d0249eb9794ccedec0ca2e3309b8d970593cc # good: [a6dcec797b8de713ce989d1d5cf732a7cf054d09] source-hash-3adad743e7702a5403927c89d2d19534509c7085 git bisect good a6dcec797b8de713ce989d1d5cf732a7cf054d09 # good: [d60dfe6cafa2df5b2720a421a17368500771d491] source-hash-e2f4289ff208ab067ab89bcd7aea8a75c3227be3 git bisect good d60dfe6cafa2df5b2720a421a17368500771d491 # good: [bf8cefdc50f9ede61040be693c101114ed019aba] source-hash-2d92a84a6aac37e34d1699fdebe0270468b4f746 git bisect good bf8cefdc50f9ede61040be693c101114ed019aba # good: [b2361aaa06315e31d2c698bcc1cf01f73e7a633f] source-hash-725150f4ad8bcbe4c9706eea264238c1e3159915 git bisect good b2361aaa06315e31d2c698bcc1cf01f73e7a633f # bad: [47c1fc249ac7e316f79f139c6438b657c9c574a1] source-hash-3b0b76462aa54f60bc9823d9c6ef64e5e35c1208 git bisect bad 47c1fc249ac7e316f79f139c6438b657c9c574a1 # bad: [bfc26f557731dd3abb6b5cba0cb23ddf7215c03f] source-hash-a55f740045d6d30e1ef5522889d2dc359ca5784a git bisect bad bfc26f557731dd3abb6b5cba0cb23ddf7215c03f # first bad commit: [bfc26f557731dd3abb6b5cba0cb23ddf7215c03f] source-hash-a55f740045d6d30e1ef5522889d2dc359ca5784a
imho not 'OpenGL related' same effect with anything 'OpenGL-related' switched off in tools - options - view reg. b.
The bibisected commit is totally unrelated. However, this is related to bug 123714.
Eike Rathke committed a patch related to this issue. It has been pushed to "master": https://git.libreoffice.org/core/+/7fdc5df36f5b50e0629405a47ff3d5765fcfeb93%5E%21 Resolves: tdf#123714 tdf#123736 all split formula groups; tdf#120013 related It will be available in 6.3.0. The patch should be included in the daily builds available at https://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: https://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
(In reply to Commit Notification from comment #11) > Resolves: tdf#123714 tdf#123736 all split formula groups; tdf#120013 related this issue is still *reproducible* > steps to reproduce: > - open attached spreadsheet > - 7. edit C2 from '=SUM(A2:B2)' to '=SUM(A2:B3)', do not! touch C1, > - 8. change B1 from '2' to '1', > -> C1 will not update > -> a hard recalc is necessary with my developer build: https://gerrit.libreoffice.org/plugins/gitiles/core/+log/a283db9b6553b232a29560dcc427329e5246f0ca Version: 6.3.0.0.alpha0+ (x64) Build ID: a283db9b6553b232a29560dcc427329e5246f0ca CPU threads: 4; OS: Windows 10.0; UI render: default; VCL: win; Locale: de-DE (de_DE); UI-Language: en-US Calc: threaded
the originally filed error was about "cells of a 'shared formula group' being excluded from autocalculate after one of the copies is edited" or similar (on my tests i'd change the last one in a column). that error is still present in ver.: Version: 6.3.0.0.alpha0+ (x64) Build ID: fe632c86aa250bb355a59ce6acf4dd75eae7afe0 CPU threads: 8; OS: Windows 6.1; UI render: default; VCL: win; TinderBox: Win-x86_64@42, Branch:master, Time: 2019-03-09_03:34:13 Locale: de-DE (de_DE); UI-Language: en-US Calc: threaded i re-tested and edited the third of four formulas in a column, after that all unchanged copies, above as well as below the edited cell, are excluded from autocalculate, they are calculated correctly by forced recalc, but persistent over that excluded from autocalculate ... b.
My bad, this shouldn't had been marked as resolved in the commit summary but related instead. Also confused things in the summary, adjusting.
Eike Rathke committed a patch related to this issue. It has been pushed to "master": https://git.libreoffice.org/core/+/1634a6d926c6cfd8fe92be1f3ae6083d2fb80f5e%5E%21 Resolves: tdf#123736 re-establish listeners for unshared formula groups It will be available in 6.3.0. The patch should be included in the daily builds available at https://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: https://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Eike Rathke committed a patch related to this issue. It has been pushed to "master": https://git.libreoffice.org/core/+/6160025b27e97841321be29863bb1efd8c194a5f%5E%21 In case of sc::NoListening only SetNeedsListeningGroup(), tdf#123736 follow-up It will be available in 6.3.0. The patch should be included in the daily builds available at https://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: https://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
(In reply to Commit Notification from comment #16) > Affected users are encouraged to test the fix and report feedback. this issue (-> comment #3) is no longer reproducible with: Version: 6.3.0.0.alpha0+ (x64) Build ID: 3140194a85fe4a6ac69c8cddc4d3b019430cd6e8 CPU threads: 4; OS: Windows 10.0; UI render: default; VCL: win; Locale: de-DE (de_DE); UI-Language: en-US Calc: threaded
Eike Rathke committed a patch related to this issue. It has been pushed to "master": https://git.libreoffice.org/core/+/564d0d145cf9c164ea9c717b4b2113fd971fa0af%5E%21 Related: tdf#123736 re-establish listeners also for vector unsharing It will be available in 6.3.0. The patch should be included in the daily builds available at https://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: https://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Eike Rathke committed a patch related to this issue. It has been pushed to "master": https://git.libreoffice.org/core/+/e5de84e73ffbaa1a45ab787750f5997582bbfa49%5E%21 Reget position_type as the block or type may have changed, tdf#123736 related It will be available in 6.3.0. The patch should be included in the daily builds available at https://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: https://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Eike Rathke committed a patch related to this issue. It has been pushed to "master": https://git.libreoffice.org/core/+/de024e572dd7a588f82b84c68daa2051ec6b20e9%5E%21 Listening when grouping in ScColumn::AttachNewFormulaCells(), (tdf#123736) It will be available in 6.3.0. The patch should be included in the daily builds available at https://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: https://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
the 'late_redraw_issue' from comment #22 and #24 of tdf#123714 is still present in ver: Version: 6.3.0.0.alpha0+ (x64) Build ID: 218916eb47e35fd14832d8f52225bf1d4d9c80a6 CPU threads: 8; OS: Windows 6.1; UI render: default; VCL: win; TinderBox: Win-x86_64@42, Branch:master, Time: 2019-03-17_04:56:25 Locale: de-DE (de_DE); UI-Language: en-US Calc: threaded dated 2019-03-17 04:56:25, the rest looks fine for me, at least much better than before! (not yet tested all variations, but me feeling is that the fundamental bug is out) (will continue testing once versions with latest patches are available) thks @Eike!
Eike Rathke committed a patch related to this issue. It has been pushed to "master": https://git.libreoffice.org/core/+/8925bd2efca8dd07a3ac044ecbc6505bb7bfa0a6%5E%21 Add unit tests for tdf#123736 and related unshared/regrouped listening It will be available in 6.3.0. The patch should be included in the daily builds available at https://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: https://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
*** Bug 122264 has been marked as a duplicate of this bug. ***
Fwiw, I created bug 124266 for the repaint failure.
Pending review https://gerrit.libreoffice.org/69554 for 6-2
more links info: https://forum.rhymezone.com/forum/get-help-with-writing-and-homework/449199-why-it-is-important-for-designer-to-know-about-trends-prevailing-in-the-market https://forum.rhymezone.com/forum/get-help-with-writing-and-homework/450339-i-need-help-with-my-text https://forum.rhymezone.com/forum/share-riddles-teasers-and-contests/450664-nba-2k19-aims-as-always-to-get-the-title-of-best-sports-video-game-year https://forum.rhymezone.com/forum/rhymezone-2016-poetry-prize-entries/371409-tethered-to-the-land https://forum.rhymezone.com/forum/rhymezone-2016-poetry-prize-entries/374738-unknown-address https://www.60millions-mag.com/forum/transport-voyage-tourisme-f63/promovacances-karavel-t11160-10.html https://subscribe.ru/group/historias-adultos/ https://www.plerb.com/tocnaza https://phish.net/user/tocnaza https://www.coactivate.org/projects/relatos-xxx/project-home http://flocks.dunhakdis.com/classic-minimal/groups-2/relatos-xxx/ https://bookme.name/relatosx https://forum.netcup.de/user/14573-tocnaza/ http://cs.athletic.net/members/836796/default.aspx https://www.esportsearnings.com/tips/15625 https://www.esportsearnings.com/tips/15626 https://www.esportsearnings.com/tips/15627 https://www.esportsearnings.com/tips/15628 https://www.esportsearnings.com/tips/15629 https://www.esportsearnings.com/tips/15630 https://www.esportsearnings.com/tips/15631
spam? here? poor people, pls. delete this comment, previous comment (#26), and the account of juancax
Setting as VERIFIED FIXED based on comment 17 and comment 21. ( @Oliver, @b., thanks for verifying ). If you find any related issue, please report it in a follow-up bug. @Eike, thanks for fixing this issue!!
Eike Rathke committed a patch related to this issue. It has been pushed to "master": https://git.libreoffice.org/core/+/8b616f94fbd040581e9a61d48066d59009ed6ae3%5E%21 Add unit tests for tdf#123736 and related unshared/regrouped listening It will be available in 6.2.3. The patch should be included in the daily builds available at https://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: https://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
However that was copied, the summary of that commit is Resolves: tdf#120013 tdf#123714 tdf#123736 shared formula group split
yep, my bad!
Eike Rathke committed a patch related to this issue. It has been pushed to "libreoffice-6-2": https://git.libreoffice.org/core/+/a4bb79de7fa046aa540cc984e43baceff9304a11%5E%21 Unit tests for tdf#121002 tdf#120013 tdf#123714 tdf#123736 It will be available in 6.2.3. The patch should be included in the daily builds available at https://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: https://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Created attachment 156009 [details] screenshot of miscalculated sheet would you mind checking if this error: https://ask.libreoffice.org/en/question/12165/calc-auto-recalc-does-not-work/ (still being in ver 6.4.0.0.a1-lin-x64 when working in 'threaded' calculation mode) might be a leftover of an incomplete fix? the behaviour is similar to that of this bug, hard-recalc corrects the results, but following the steps described the link produces obviously wrong results as visible in attached screenshot. reg. b.
Created attachment 156010 [details] better screenshot, here you can see the formula of one of the miscalculated cells better screenshot, here you can see the formula of one of the miscalculated cells
(In reply to b. from comment #33) > would you mind checking if this error: > > https://ask.libreoffice.org/en/question/12165/calc-auto-recalc-does-not-work/ Please open a new bug report on that.