Apache OpenOffice (AOO) Bugzilla – Issue 68812
Simultaneous modification of several equations by editing only one
Last modified: 2009-07-12 19:28:27 UTC
Below please find a bug report for OO-Impress & Math. Attached please find also a snippet of the presentation in which the bug occurred. The snippet is stripped down to show only the bug. OO version in which this appeared: ---------------------------------- Version : 2.0.0 Release : 1.2 Vendor: SUSE LINUX 10.0 Source RPM: OpenOffice_org-2.0.0-1.2.src.rpm Severity: --------- I consider the bug to be severe. It ruins your whole presentation. Brief description: ------------------ Upon editing *one* of several disjoint equations on a slide, *other* equations on that same slide will also change 'miraculously' more or less according to the changes made in the actual equation you are editing. Detailed description: --------------------- Please open the attached Impress file. It contains only one example of the many places in a 30-slide presentation in which this bug occurs. You see two 'equations': | psi_1 > = | psi_1 > = Select the right of those two equations and open it in the equation editor. Change the index '1' into '2' and close the equation editor, such as to get: | psi_1 > = | psi_2 > = However, on my screen, both, in the normal editing mode of the presentation, as well as in the slide show, the slide now reads: | psi_2 > = | psi_2 > = I.e., the left equation has 'somehow absorbed' the changes made to the right one. You can convince yourself of that by opening the left equation in the equation editor. Indeed it will have an index '2' by now. This can be made worse. Open the right equation for a 2nd time an replace the whole equation simply by the letter, say 'a' and close the equation editor. Now, on my screen and both, in the normal editing mode of the presentation, as well as in the slide show, the slide will not read | psi_2 > = a but a a where the left 'a' is strangely rescaled. Deleting, say, just the right equation and rewriting it looks like a cure at first sight. However, after saving and reopening the document also the left equation has 'disappeared'. The last resort seems to delete all equations from the presentation and write them all anew :-( In summary, the presentation behaves as if disjoint equations are 'somehow' linked to each other ??? Unfortunately I can not give a description of the action which may have induced this behavior. It just 'crept into' the presentation at some point that I can not pin down. Regards, 123ooofree
Created attachment 38663 [details] Simultaneous modification of several equations by editing only one
I checked with "2.0.2 German version WIN XP: [680m5(Build9011)]" - 'test.odp' - own testfile WFM @123ooofree My OOo told me that 'test.odp' is damaged. Pls. contribute a step by step instruction how to create an own presentation showing that bug.
Do you still know how you inserted these formulas? Did you use insert/OLE/Formula, typed in the text and then inserted the next formula? Would it be possible to get a more current version from the OOo site? Thanks in advance.
@rainerbielefeld For me my OOo surely does not tell me that 'test.odp' is damaged. I also re-checked, that the file 'test.odp' which I have uploaded with this issue still shows the bug and is not corrupted relative to my local version by downloading it again. @wg 1st: Can you confirm the behavior. Please let me know. 2nd: Usually I generate equations by exactly the stanza you quote, i.e.: insert->object->formula (although I have a shortcut for that in my toolbar just to click on it) Now, it does happen, that I just copy a formula, if they are sufficiently similar, in order to modify one of them later. In no previous case has this ever produced any problem like that, and I don't think it's forbidden to do that. @rainerbielefeld and @wg: I cannot possibly retrace the action that produced the issue. I was just working on a presentation, when I suddenly realized, that there was this 'coupling' between otherwise disjoint equations. Is there no way to 'look into' the *.odp file, to see why there is this behavior? Regards, 123ooofree
I can reproduce the problem with your file, yes. But if I start to design a similar file in a current version it works for me. Was your OO version part of your Suse installation? If so I would recommend to download the latest stable from the official OOo site and have a try with this one. You can open an .odp file with a packer program. Contained there are several xml files and graphics, but I doubt that this might be help for us in this case...
@wg: I saw the same. 1 time I succeeded in opening the attached 'test.odp', and I was able to reproduce the reported behaviour, but no Idea how to create an own document to quest for the reasons. There was a Method how to create something similar to heading or footer in OOo 1.1, may be this is just that, but accidental? I do not remember how that worked in 1.1.
@wg and @rainerbielefeld > I can reproduce the problem with your file, yes. > I saw the same. > But if I start to design a > similar file in a current version it works for me. ok., so the issue is 'real'. Does the issue hve to remain 'unconfirmed'? If not, then who can change it to 'confirmed' or whatever ... @wg Regarding the rewriting of all of the equations anew ... that's as I wrote in the description of the issue ".....The last resort seems to delete all equations from the presentation and write them all anew ...." Obviously, this is no problem if you have two simple equations like on the 'test.odp' slide. For me, however this happened while performing some final editing on an already more or less finished 30 page presentation with seveal equations 'somewhat' more complicated ... and while on conference travel ... and with 24 hours left to fix the issue and the presentation somehow ... So it would really be great if there would be any way to find the reason for this behavior from the example file. I just don't know what steps specifically produced the behavior. > Was your OO version part of your Suse installation? Yes. > If so I would recommend to download the latest stable > from the official OOo site and have a try with this one. I will do so once I'm back from travel - I definitely won't change a running version while having to give several talks abroad. Anyway since the issue has not been reported before, I doubt it will change in a newer version. I still hope there may be a clue solely from the file ... Thanks and regards, 123ooofree
@wg > You can open an .odp file with a packer program. Contained there are several > xml files and graphics, but I doubt that this might be help for us in this > case... Ok., so I peeked arround for a while in the unpacked content of the 'test.odp' file and also compared it to the content of a file created anew with the 'same' two equations, called 'nest.odp' It may be obvious to you, but at least for me it was interesting to see, that the unpacked *.odp files contains directories of name 'Object n', where 'n' seems to be a number counting the objects. 1) Now, in 'test.odp' there is only *one* directory of type 'Object n', while in 'nest.odp' there are *two*. So indeed the two equations one sees on the screen in 'test.odp' are actually only a single object (?). 2) In each 'Object n' there are two *.xml files. Looking at them with the Konqueror, one sees that one of them (content.xml) contains the actual formula. Now, in 'test.odp', content.xml shows the subscript '1' for the letter psi in the equation ... although on the screen the equation shows up twice for 'test.odp' ... and although it appears with two different subscripts, i.e. '1' and '2' So now, hopefully somebody of you knows how to make a file - all alone by working from the user interface - in wich there is only a single object, which however shows up twice and while doing so, changes its properties ... like the subscript 1 -> 2 in the present case. Regards, 123ooofree
Yes, I can confirm what you have seen in the unpacked files. I will set the issue to confirmed and reassign it. But I doubt if we can do something if we still do not know how this has happened... but lets see. Thanks for your helping anyway, that was great.
*** Issue 72158 has been marked as a duplicate of this issue. ***
Created attachment 49481 [details] Step by step description to generate issue 68812
Hi, more than a year ago we had some email exchange regarding issue 68812, which I posted. If this issue occurs - and it does so frequently - it will screw up completely any Impress presentation containing equations. I consider it very severe. The status of this error has remained 'as is', since there was no step-by-step procedure to reproduce it, although it its existence has been granted. Recently and within a multi-page presentation (which, as so many times was killed by issue 68812) I was finally able to track back, and isolate a reproducible sequence of steps to produce the error. This sequence may seem awkward, but actually represents a typical work-flow while editing heavily on a presentation. Attached is a tar-file 'issue.tar', containing file 'of1a.odp', which is the file to begin with, and the results of the steps outlined below. I very much hope this will help to fix the issue. 1) First: I'm running linux, SuSE 10.0, kernel 2.6.13-15.16-default and I'm currently using ooo v 2.2.1 - but this error has occurred to me with any other linux and any other version of ooo I've been using. So, please don't wait to check all of this until I have installed ooo version x.y.z, since I doubt the error is pertinent to only a particular version of ooo. 2) In the PWD of 'of1a.odp' I generate two directories: '1a' and '2a'. Into '1a' I unzip the xml-content of 'of1a.odp' This is also contained in the *.tar file attached. Please note, that you will find *two* object directories in '1a', i.e. 'Object1' and 'Object2' containing the xml data corresponding to the *two* equations displayed in in 'of1a.odp'. 3) Now we do the steps to generate the error a) open 'of1a.odp'. You see one line of text and two equations. These entities are not grouped. b) Next, group *only* the line of text and the *second* (lower) equation. c) Now, *enter* this group of two entities and double-click on equation in the group to edit the equation. Change the %chi_{%my%ny}(t,t')= in this equation to %chi_{%my%ny}(t,t')`=` (i.e. add some space the = sign) Click on the the viewgraph's window to close the equation editor. d) Now, *exit* all groups. e) Next, *ungroup* the line of text and the equation. You should be back with three ungrouped entities. f) Now, group *all* entities, i.e. the line of text and both equations. g) Save this presentation as 'of2a.odp'. This is also contained in the issue.tar file. h) exit/terminate ooo 4) Unpack the xml-content of 'of2a.odp' into PWD/2a. This is also contained in the issue.tar file. Now(!!), to begin, note that in '2a', and all of a sudden you will find only *one* xml-object representing a single equation, even though our presentation should still have two equations. If you now fire up ooo again and open the presentation, you will indeed find, that the content of one of the equations has 'magically' been replaced with the content of the other equation, leaving only one equation, however displayed two times. 5) If you now, would start editing one these two equation to 'correct' the error, bizarre things will happen, because the contents of the equations are linked. These dependencies usually will be visible to you only after saving and closing the presentation and exiting ooo, and then restarting ooo and opening up the presentation again. I'm not aware if this is the only way to produce this error. However, it seems to manifest itself more or less always in an identical way, i.e. by the fact, that equations seem to be 'linked' and change simultaneous under editing of only one of them. Along this line I have been screwing up over and over again many involved presentations, which usually have far more complicated grouped objects of equations and other things. Once the error is in the presentation it is hard to pin down the origin and usually it means: 'start from scratch'. Please fix this! Thank you! 123ooofree
The previous two comments contain a step-by-step instruction to create issue #68812 and a corresponding file-attachment. Please consider this. Is there any chance that this info will induce a fix of this P2 bug? Moreover, do the 'Keywords' of this issue still need to be set to 'needmoreinfo' ?
I am a bit confused with the status of this issue since the status is still "NEW". Anyhow, I did manage to reproduce the bug from your file under - SUSE LINUX 10.0 --> OOo 2.0 - Red Hat Enterprise Linux 5 --> OOo 2.3 and - Ubuntu 8.10 --> OOo 2.4.1 I found that, it did not matter whether I changed the left or the right equation, the other one was affected. For the scaling problem, I found that if I replace the right equation by the letter i.e. 'a' it is the right side, not the left equation, that gets strangely rescaled although both of them got changed to 'a'. As it was found by others, the problem can be fixed by starting the whole equation/presentation from the scratch. But I was wondering if anyone out there managed to fix this bug without it from the scratch!! Regards, Rafiq
> I am a bit confused with the status of > this issue since the status is still "NEW" ... Working with Ooo-Impress for quite a while, I've come to realize that this part of the Ooo-Suite is probably the one which gets the least attention by the developers. Impress contains quite a number of very distressing 'features' which just don't seem to change from version to version. Take a peek at http://www.fkt.tu-bs.de/brenig/OpenOffice.shtml For eg. issue 48179 "Cannot step back slideshow animations" has been around for *many* *years* by now, with a potential 'workaround (?)' available only/maybe in the next version - let's see. The problem here is the system of weighing the relevance of an issue by 'vote count'. With quite a large number of users of Ooo-Writer, it is obvious that any kind of request for enhancements of Ooo-Writer will immediately get a huge amount of votes and thereby attract the interest of the developers. However, with the number of users of Ooo-Impress remaining relatively small as compared to those of the competitors product (sad as it is), even *severe* *bugs* like issue 68812 just don't get enough votes and therefore remain as they are 'NEW (?)' - even after years. So your best chance to get a fix on this issue is to increase the vote count to, say 50 ... or maybe you have a way to fund a dedicated developer ... Anyway, I fear that issue 68812 will remain 'NEW' for the rest of my active career :-( Regards, 123ooofree
so much motivational speech made me miss that we already have a step by step description to reproduce the issue, retargeting
yehyyy ! I can retire :-) ... well, I just decided to wait and see if the issue may even get 'RESOLVED' Regards, 123ooofree
Just for a followup on this. It seems the bug is fixed for OOO300m15(Build:9379) on all OS. Well, at least I could not reproduce it. Can anyone confirm on this. Regards, Rafiq
Rafiq wrote: > It seems the bug is fixed for OOO300m15(Build:9379) on all OS I don't know if this bug has been fixed on all OS ? At least there has been no sign of activity on this issues here. Also, as you noted yourself a couple of days ago, its status was "NEW" by then. Moreover you also noted that: > I did manage to reproduce the bug from your file under > - SUSE LINUX 10.0 --> OOo 2.0 > - Red Hat Enterprise Linux 5 --> OOo 2.3 and > - Ubuntu 8.10 --> OOo 2.4.1 For more than a year I did not check if the 'example file' attached with this issue still 'works'. Since I don't have access to 'all OS' but only to Linux and also since I don't keep all previous versions of Ooo to play with, the only thing I can say is: 1) The 'example file' still produces the bug under Ooo 2.2.1 and 2.4.1 on SuSE 10.0. Moreover I remember that it also produced the bug on all previous Ooo versions I had - all runnning on various SuSE Linuxes. 2) On Ooo 3.0.0 OOO300m9(Build:9358) runnnig on SuSE 11.1 Linux the 'example file' does *not* (!) produce the bug. If that implies that anything has been fixed w.r.t. issue 68812, I cannot tell ? Maybe there will be other situations in which issue 68812 shows up, and it is just the 'example file' which one has to change ? Maybe issue 68812 depends on other components of Ooo which have been modified ? Maybe ... maybe ... it would be nice I someone more official w.r.t. the Ooo project than me 'user average' would comment on all this. Anyway, lets keep using Impress and see if issue 68812 has vanished in the haze miraculously :-) Regards, 123ooofree I really don't know if this bug has been fixed on all OS ?? Also I don't have access to So What I can confirm ishave at hand are three Lin Regards,
changed http://www.fkt.tu-bs.de/brenig/OpenOffice.shtml according to Rafiq's findings. Regards, 123ooofree
I also tried to reproduce this on OOo 3.0 and current dev300 m44 and had no look with the step by step instructions. Could have been fixed by work in other areas. I'm closing this issue for now, If someone manages to reproduce it again, please reopen
worksforme -> closed