Apache OpenOffice (AOO) Bugzilla – Issue 101973
Collate print option defaulting to "on" causes problems
Last modified: 2017-05-20 09:26:20 UTC
The behavior in OO 3.0 & 3.1 was changed... Now the default in OO is to always have "collate on" in the print dialog, with no default choice allowed, without remembering the previous setting, and with no system-wide settings option. This is a now MAJOR problem for us. Under Linux/CUPS with network PS printers, the printer handles collation automatically. You just send a multipaged job to the printer, and the number of copies (with collate in the application OFF). The printer will repeat the job in it's memory quickly and efficiently producing a collated output. (It is extremely rare anyone would want uncollated, anyway). Now that the default in OO has changed, the users don't know that they are NOT supposed to check "Collate". When one checks that box, OpenOffice will generate the output N number of times. When a user requests 100 copies of a 50MB document, and collation is "ON" in OO, it sends 5,000MB of data to the printer. It causes a tremendous waste of CPU (we are centralized with thin clients) and tremendous waste of network bandwidth. It also slows down the printing process. All because that default was changed. I can't find a way to change the default to OFF like it used to be before OO 3.0. We need an environment variable or a way to hack the config files in the installation directory. A better long-term fix would be to at least have a default setting for collate in tools->options. Have a three-way setting: "always on", "always off", and "remember last setting".
duplicate *** This issue has been marked as a duplicate of 59506 ***
closed
Adding updated info from followup on issue 59506 here... Some people will want collate on as the default... just for convenience. Many people have asked for it, and unfortunately it was given to them in OO 3.0 without any real regard to the consequences. Many people will have severe problems with the default being "on", and our business is one of them (as I explained in a previous comment, above). There might be some type of different options under other OS's (MSWin/MacOS), but under Linux it is very problematic using the collate function on. The suggestion about storing the collate setting per printer is probably not practical, since there is no user interface in OO that addresses saved options on a per-printer basis (unless they want to resurrect the spadmin tool, which is now pretty much dead). Storing such a setting in a document doesn't really make any sense. So the only best remaining way to handle the problem is to make an option in tools->printing that can hold a user's preference for collation: 1) Remember previously used state (begin with setting off) 2) Always default to off 3) Always default to on 4) Always remain off (and don't show option in print dialog) With the default setting for new installations set to #1 (since that is most logical). The above change would solve the problem for most people once and for all. Although there should also be a way that a system administrator can set the default for new users, to make the problem absolutely solvable on systems where printers do their own collation all the time (which is most modern laser printers and copiers).
pardon me, but a listbox on an option page that nobody finds that configures the on/off behavior of a single checkbox seems a tad overengineered to me. Especially since it does not satisfy you requirement that a site administrator can force collation to off. How about you get the configuration file entry for the site administrator that supports "Collate persistent (as last state)" and "Never collate" ? Please be also advised, that we're talking only the print dialog here, not whatever any script will do if printing a document.
I agree that my UI suggestion be overdoing it a tad...as an administrator, I tend to like lots of options :) As an alternative, I suggest that the print dialog should be changed to remember the previously selected state for collate, for all users, and start off with collate unchecked (OFF) for new installs, since that it is the safest setting for all situations. Everyone will get what they want and need. Those irritated with the old behavior of it always reverting to uncollated will only have to check it once. Those who don't know the difference, don't care, or won't mess with it, will have a default that does not cause problems for page printers. Then for further refinement or lockdown, give options in the configuration file entry for the site administrator that could turn it off and/or hide it. I don't think having only a config file without changing the print dialog default behavior is good enough, since that presupposes that sites (and even end users) know there is the problem to begin with. For example, I didn't discover the problem until after implementation was already completed and OO 3.0/1 rolled out to users because I never thought to test multiple copies of large documents. It never occurred to me that OO would send hundreds of identical 50MB print jobs when collation used to work fine with collate "off". After the network became saturated and server and end-user performance suffered, only then did I discover it... (and had absolutely no way to fix it). Whatcha think? A reasonable compromise?
It would be useful to know if this issue only affects Linux or also other platforms as well. For example, if you set collate "ON" in MacOS version and send 2 copies of a 10MB multipage document to an intelligent page printer, does it send 20MB (2 x 10) to the printer or just 10MB? (It is easiest to determine this by including large bitmap graphics in the test document, then check the CUPS logs to see the size for a 1 qty print job, vs a 2 qty print job) Do note, however, that the proposed solution would not adversely affect other OS versions.
Considering that the current behavior, if it may be a real problem, is not an "error" in the meaning of "Ooops something is broken" but the result of an intended enhancement, this issue should be qualified as enhancement as well. @crxssi: Please don't set "Target milestone" by yourself. Let DEV/QA people do it.
Sorry, didn't know I wasn't supposed to set targets... perhaps "wishful thinking?" :) You imply it might not be a real problem....trust me, it is for sites like mine. We can't afford the network and CPU load that sending zillions of unnecessary copies of print job data places on us, especially when it is completely avoidable (since *all* our printers will always automatically collate). Anyway, I do agree with your reasoning about the issue type. It isn't necessarily an "error" ("defect")... the inherent behavior of sending multiple redundant copies in the print job has always been there when collate was "on", even before the defaults were changed. It is just that many of us wouldn't know about the problems (or have issues with it) until the option was forced as default "on" every time the print dialog appears. However, it is somewhat more that just an "enhancement", because it is attempting to fix latent problems/issues that arose from the previous "enhancement". Since there aren't all that many "issue types" in the QA system, I suppose "enhancement" would best fit... but with more of an urgency than a typical "enhancement" would justify. Fortunately, I believe the minimal "fix" (start OFF, then remember previous state across dialogs and sessions) is probably easy to define, understand, test, and implement, and with no negative consequences; all while retaining the original intent of the first "enhancement": making it so nobody will have to constantly change the collate option flag.
AFAIR the reason why the same content is sent to printer multiple times in case of "Collate" is selected was that many printer drivers didn't handle collating correctly, so it was decided not to use the "Collate" function of them but instead of that emulate it by sending multiple copies explicitly. Maybe it is time to reconsider that? Or do we still fear/know that some printer drivers screw it up?
wrt. to whether "Collate" should be the default: if we hadn't changed it recently, I just would do it as most users know it from other applications. OTOH (changing something back and forth is also unfortunate, so an acceptable compromise that keeps the new default would be nice. IMHO the idea to keep the setting permanently until it is changed again is good and should be this acceptable compromise for everybody as with this enhancement the default setting after installation doesn't matter so much anymore.
@mba: I think you lost me :-) Does that mean you think persistence of the collate box is enough ?
Yes. :-) If the setting in the print dialog would be remembered the question what should be the default probably is moot. But we shouldn't forget the original problem of the reporter: can we hand over handling the "Collate" property to the printer drivers or are there any reasons to still implement it in our code?
@mba: Oh, I understand WHY OO emulates collation- non-page-printers (cheap inkjet, for example) can't collate on their own. And before OO started really using CUPS, there probably weren't a whole lot of options to interact with printers correctly. And OO is not the only application under Linux that does this (it just affects us the most). It would be best left up to CUPS to handle the whole collate process. I don't know enough about CUPS to know if it has a readable property to determine if the printer can collate on its own. I also don't know if you can tell CUPS to do the collation, and CUPS will decide if the printer is smart or stupid and handle it properly, itself. This is, of course, the root cause of our issues. Perhaps this CUPS interaction concept should be yet another issue number, assigned to PRINT? @pl/mba: Our (my) problem would be 75% solved with just adding persistence. However, if there is no way for me to get the starting default to be OFF, some of my problem remains, since I am forced to rely on education to let every user know they need to change it manually, at least once. And, then, every time there is an upgrade. I would still hope for a sysadmin setting/config file, or environment variable or something. But I will take whatever I can get as a major improvement over what we have right now. :)
So persistence plus an configuration entry without UI should do the trick. I added persistence yesterday to the new dialog, that leaves the config entry.
@pl: yes (assuming persistence will last across dialog pop-ups AND across sessions) Thanks so much... sounds like at least that part could be pushed out quickly. @EVERYONE: I need to correct something I said in my very first posting... I implied the issue is only a problem when printing more than one copy of multi-page documents with collate ON. This is not correct, it is worse... The problem occurs even when printing more than one copy of *single-page* documents with collate ON! So 100 copies of a single 50MB page also sends 5000MB to the printer. I do wish I could find some type of hack to immediately change the default so I can roll out 3.1.0 here (our compiling OO is not an option). The problem is severe enough for us that I had to halt our rollout upgrade from 2.3 to 3.1 that we spent several weeks configuring, testing, and documenting :( Isn't there some secret file I can hack?
sorry, that is as far as I can see hardcoded in svtools/source/dialogs/prndialog.cxx lines 136 and 137.
target
Hi! I just want to remember that the settings in the Tools -- Options usually set the default behavior for all documents. These settings can be altered per document via the usual print dialog options. Letting the print dialog remember the last state "breaks" this kind of behavior. So my proposal is to add this kind of option setting the default to either Tools -- Options -- OpenOffice.org -- Print (all modules) or for each of the modules (if required). I don't like to add another option there, but it respects the current behavior. And since many of our users (non-admins) fear too many options, I would like to remember that we will need a replacement for the overstuffed Tools -- Options dialog (URL on request *g*).
Why can't "Tools-Options" just take over the remembered value?
@christophnoack: No, it breaks nothing by having the print dialog remember the last collate setting used. The collate setting is not saved with documents (neither is the number of copies, for example). As long as it is not saved with the document, then remembering the last setting used will not break (interfere) with anything else. As far as using "tools->options": I first proposed that we use tools->options->print by adding a selection for collate. But that was shot down by "pl" (read the 5th message from the top). I see the advantage and disadvantage of both approaches. It seems the most logical place to address it would be in tools->options. But the *easiest* (and fastest) way to address the problem is by just having the print dialog remember the last used setting; storing that state in some config file. The latter approach requires zero UI changes/testing. More importantly, "pl" has supposedly already added code for it to remember the state in the print dialog, so it is probably moot now to worry about it. I will just be happy if this is fixed in the next release- whatever form the fix takes. ANYTHING is better than the current behavior in 3.0.0 - 3.1.0
FYI: The persistence is not just for the collate checkbox, but for all controls that are not stored in the document (copy count, N-Up, etc.); this was what came out of our discussions with the application people.
Yikes!!!!!!!!! You don't want persistence on number of copies! I can see that being a big problem! User opens document, prints 100 copies. Closes document. Next day, they open some other document and then print it. Users will not think about number of copies, because they assume 1 (like all other applications)... but they get 100! Even worse- next day, user opens another document, then prints using the quick-print icon. They get 100! If there is one thing I know really well, it is user behavior. Trust me, you don't copies to be persistent. Persistence with collate makes sense, because that is a decision that is likely to never change. But copies should always default to "1". It is much, much safer. BTW, Print->Options and Print->Properties are things stored in the document itself (I just checked); which makes sense. So there are only two other options- collate and copies. I don't see "N-up" or anything else...
N-Up is new in the revised print dialog. And I'll take into consideration making the copycount 1 each time; I can understand your reasoning.
@pl: great! Hey- N-up sounds like another really neat option! (If it is what I think it is- printing multiple pages on a single page side/face). But beware- I suspect that is another option that should probably always default to "1". At least there is far less "damage" (print resources) if that is accidentally left on, if you know what I mean.
Added a config setting in VCL.xcu in the "PrintDialog" section so that you can switch the Collate CheckBox to off always; administrators can do that in the file <office>/share/registry/data/org/openoffice/VCL.xcu by changing the value for key "CollateBox" from "Default" to "AlwaysOff" (actually case insensitive). In the default case you can also set the initial value for the collate box there by changing the value for the key "Collate" (which is currently "true"). This value however will get overwritten by the user checking the collate checkbox on/off. changes committed to CWS printerpullpages
@pl: that is great news, indeed. From what I am reading of what you programmed, that is exactly the type of behavior that will allow the collate problems to be fixed for *everyone*. I (and my users) can't thank you enough. Hopefully this can be made into documentation somewhere too, since others are bound to hit the same problem. Is there any expectation this will make it into release sooner than 3.2? As for the N-Up- I talked about that with some people at work, and they are quite excited about that option. Would be nice for "preview printing" and archival of large documents to paper. We all still think it would be best if that were also not made a persistent/remembered setting, although it would be far, far, far less "damaging" than the issue with collate or number of copies.
N-Up is not persistent as per your suggestion. The CWS will hopefully get integrated in a month or so (no promise though, things could come up), so a development milestone for testing should be available well before 3.2. At least that is the plan.
I am guessing the change in target is bad news and we won't be able to upgrade from OpenOffice 2.X for another year? :(
I consider the change in target "bad news", yes. OTOH I wouldn't want "almost" ready printer code to make 3.2 a bad version. Whether you can migrate or not only you can tell.
Well, I don't want untested or unstable code in a mainline release either. But I was at least hoping the proposed <office>/share/registry/data/org/openoffice/VCL.xcu setting would be available quickly. Since that doesn't affect the UI, it should be a safe place to start, wouldn't it? As for upgrading- the problem is so serious, there is no way I can roll out 3.X to my users if I cannot disable the collate option (or at least default it to "no").
Sometimes things don't happen as you want it. When we (mainly pl) started to work on the printing stuff in CWS printerpullpages, we all thought that we would finish in time for 3.2. As it doesn't make a lot of sense to have several work spaces changing the same code at the same time, it seemed to be appropriate to fix issues like this one in the CWS "printerpullpages" also. Unfortunately now this means that because the main feature in this CWS is not ready in time also the smaller fixes will have to wait also. You might complain about bad planning, but that's unfair: hindsight is easier than foresight. I think pl did a great job, not only in what he has achieved in this work, but also in taking the responsibility for product quality and not rushing things too much.
please verify in CWS printerpullpages
Verified with cws printerpullpages = The Checkbox for the collate is persistent now but this fix brings some side effects like that the first page of third, will be printet at the backside from third page by duplex printing. I'll file an follow up issue.
When will there be a patch? We have 60 people here and each of them is printing 30-40 times a day on some Cannon IRC2380i. WE NEED A PATCH!
Only 60? I have over 150 users and we have been waiting over a year. I had to completely halt the rollout of OpenOffice 3.0 because this problem was too severe for us. Waited through 3.0, 3.1, then 3.2. I am praying it is in 3.3 (I think it is). I, too, was hoping there would have at least been a back door fix or something until the new print system stuff was rolled out.
A whole Year! o0 I give my best wishes to 3.3 too... "if the issue is just marked as FIXED but not CLOSED, then this probably means that the latest milestone doesn't contain the fix yet..."
@ pooradmin: Just stop complaining: Try OOO330_m2 which will be OOo 3.3 and see if the fix will work for you.
Well, the collate print option seems to work correctly on Lenny 64 :) Just a few days more and i can install the German version. Thank you so much! You all save my live :D