calc autocalculate fundamentally broken? need help how to handle

hi @all,

i’m new to ‘online bug hunting’, pls. be so kind and give some help,

one thing first, i like libreoffice, especially calc, imho it’s much better than commercial competitors, i had massive problem with ex_el and i know why i changed, calc gave so much to me and my life, i ow it gratitude and respect!

i criticise only! to get faults out and keep it’s standard high.

imho calc autocalculate has one or plenty massive bugs in it, which exclude some cells from being recalculated after changes on them, on copies of them, in ranges they point to, and similar.

reports about this show up at least since 2013!, and since version 4.x, plenty of them mentioning ‘obstructive’, ‘randomly’, ‘hard to reproduce’.

i know!!! that working on such bugs is very!!! hard and difficult

but it’s neccessary, reliable autocalculation is one of the fundamental functions of a spreadsheet,

‘F9’ - ‘recalculate’ doesn’t help (in most cases), acc. to what i’d read while ‘autocalculate’ is on F9 recalculates only! intentionally changing cells with functions like ‘random’ and ‘now’ and their dependents, thus ‘normal’ cells and functions are not touched and not recalculated.

(irritating behaviour for users who use their private silly understanding of the word ‘recalculate’, difficult to understand without rtfm, but ok, it’s written somewhere)

(it’s not read by plenty persons, thus hints like ‘Hast Du schon mal im Menü Daten>Berechnen>Neu berechnen (F9) angestoßen, …’ fill the scene)

‘strg-shift-F9’ - ‘hard recalc’, does! correct wrong results in most cases (my limited lifetime blocks me from checking ‘all’), it is suggested as ‘help’ to irritated users … imho that’s ‘eyewash’:

  • acc. to it’s ‘definition’ in the online wiki ‘autocalculate’ should! recalculate every cell after any change in the sheet.

  • there is no help to users which cells will fail and when they have to do manual recalculation,

  • AND!!! … strg-shift-F9 corrects the actually shown results in mostly when you ‘fire’ it, BUT LEAVES THE AFFECTED CELLS EXCLUDED FROM AUTOCALCULATE!!! thus you have to fire it again, and again, and again …

  • how shall a user produce reliable results with a spreadsheet where most of the cells are autocalculated while sometimes some of them fail???

i’m adressing this problem for some days now, and it starts that i get hints like ‘you already filed bugs about it’, ‘that’s not the same bug’, ‘why do you repeat known bugs?’ ‘use hard recalc’, ‘this bug is fixed’, ‘may be open-gl related’, and so on, while it’s evident to me that the fundamental weaknesses is still present,

everytime i play around and look for errors like these i find new ones, not really new, but other situations where errors show up.

my two cents:

  1. somebody introduced something like excluding some cells from autocalculation when they are not subject to changes, may be his intention was ‘performance’,

  2. he strayed in decision about how to decide which cells may be excluded, especially when ‘groups’ of ‘shared formulas’ are in the game,

third cent:

  • since then plenty developers plastered single symptoms of that weakness, leaving the fundamental bug untouched,

it’s just my idea about the situation, i may be fundamentally wrong …

i’m not! wrong in analysis that there is! at least one fundamental bug producing errors, that it is! still present in ver. 6.3.0.0.alpha0, and that it should be obliterated!

so what shall i do???

file a new bug for every ‘new’ misbehaviour? e.g. there was a fault for ‘=SUM(of a range)’, hen a new one for there is a fault for ‘=PRODUCT(of a range)’ too?

file a new bug for every new place in a sheet i find wrong behaviour? e.g. there is a bug report for moving H7:G8 failing in one sheet, i shall write a new one for moving B1 to B2 produces a similar error in another sheet?

reopen bug reports where i find the error still occurring and get them ‘re-closed’ implying i’m incapable to understand the subject that’s written as headline of the bug report?

i think i’m able to understand ‘Cells not recalculated after moving data within ranges pointed to by a shared formula group’

write a makro to test sheets and produce bug reports automatically?

no - that can’t be the solution,

but i hope there could be a solution, how can i / can we find someone with the skills, time and mood to address this problem in detail?

imho it’s urgent, if the bug is left unfixed i have to change to another software and recommend the same to others :frowning:

any help appreciated,

reg.

newbie-02

P.S. for those who need a detailed test case, use search or google for ‘calc’, ‘autocalculate’, ‘recalculate’, ‘wrong’, ‘fail’, ‘error’, ‘excluded’ and similar keywords, you’ll find enough of them

Are you on a crusade or something? Doesn’t that question suffice?

This is not even a question, it’s a rant.

‘duplicate’? - no, it’s a try to get things together.
‘that question’ - in the view of plenty people describes a different bug.
‘crusade’? - too small military potential with me, but if a crusade ist startet against that bug i’d follow.
‘rant’? - i admit being a little ‘pissed’ by that bug and the handling of it, but i try to keep calm and provide positive help. if it’s neccessary i’m in the mood to awake people, the problem is too widespread and too dangerous and too many complaints faded out without real fixes :frowning:
‘that question’ is a very complex network of “if’s”, and very hard to resolve while 1+1=3 resides in the system.
proof?: A1: 1, B1: 2, C1: =sum(A1:B1), copy A1:C1 to A2:C2, move B1 to B2, enter 1 in B2, observe result in C2. with 6.3.0.0.alpha0 you’ll see 1 + 1 = 3 in the second row
‘autocalculate is broken in many respects around shared formulas!’

hello @all,

i apologize for having been a little ‘too cruel’ in wording,

@erack:

i kindly ask to reopen this thread, imho there are! two questions of public interest,

  • how can we get rid of the problem with ‘changes in the area of shared formulas block autocalculate’?

  • what to do if similar things happen again?

my solution for the moment: i’m back to 4.1.6.2, no irritations yet, it makes problems with a 4k screen, icons and dialogs much too small, the rest works fine - afaisy -

i’d bee very thankful if someone can give me a hint once these problems are wiped out

regards

b.

Bugs are submitted and being worked on. You are the submitter of the bugs, you will be notified.

as far as I can see these problems have been worked on for version 6.4. I’m not able to check if everything is working correctly now, and I’m not qualified to tell if it’s been thoroughly cleaned up or just plastered, but it works much better now. I thank all the people who contributed to it and I’m happy if they continue to take old bugs seriously and fix them, it results in a more reliable program and saves administrative work for duplicates of error messages.

thank you very very much, i know you are on a difficult job, and excuse me that I had to get a little rough to get this thing moving.

b.