On a system with no Java Runtime Environment (JRE) installed you get an error message only once. This is obviously nice for repetitive error messages. But it is not good that a *user action* produces no feedback. E.g. If the user (with no Java) tries Wizard Agenda (which still depends on Java) and gets a "LOdev requires a Java runtime environment (JRE) to perform this task. Please install a JRE and restart LOdev." message, that is fine. However if on the same session s/he then tries Wizard Web Page, s/he gets no message at all. The user might reach the wrong conclusion that the Web Page wizard simply doesn't work (especially because others do :) ) IMO until all scripts are converted from Java, there should be a message for each *user action*. NOTE: This probably also applies to a system with Java disabled (haven't tested)
I vaguely remember this being discussed at some point and there was mixed opinions about if it's better to constantly bug the user about java or just to let them know once and then if they do more that requires java, they should remember the first annoying message telling them to instal. Bjoern, I cc'ed you just because I thought you might know/remember something more than I do. Was this discussed and purposely done this way?
@joel I think the reporter is right. Bug the user every time he runs LiBo would be wrong, but inform the user about the reason why a certain functionality can't be run is another story, needs to be done every time he tries. If he tries again in the same session means he has forgot can't be done, and for sure has forgot or has not understood why, so would be better give a more informative message. In addition I think: a) all the java dependencies must be removed b) in the meantime, windows installer should gently inform the user about what is missing and have a clickable link to the download page. Now it prompts the user like "ERROR ERROR" just scaring c) if java is not present, user has to be informed about why something does not work because of this and provide a link like in b). Don't just tell me that it does not work in your jargon language, also explain to me how to fix it!
Marking as : New (confirmed) Minor (not really a bug but if it leads to confusion to users, worth leaving at Minor) Low (default for Minor, I don't see a reason to escalate this) We'll see what happens with this one when developers take a look at it
Pedro: could you give a try to a newer LO version? (for example 4.1.2) See https://bugs.freedesktop.org/show_bug.cgi?id=49453#c8
(In reply to comment #4) > Pedro: could you give a try to a newer LO version? (for example 4.1.2) > See https://bugs.freedesktop.org/show_bug.cgi?id=49453#c8 Hi Julien You are right, this bug no longer makes sense. However even if all Wizards were ported to Python, they don't all work. See http://nabble.documentfoundation.org/Libreoffice-qa-ANN-LibreOffice-4-1-2-RC3-test-builds-available-tp4075690p4076740.html And no, I haven't filed a bug yet ;)