5 ms·
Request for Referee: QT Bug Report QTBUG-14664
After a recent article in Linux Journal, I tried using qdbusviewer to drive Rhythmbox and learn about DBUS. When I tried to skip to a given spot in the currently playing song, it did not work and no error message was displayed. I went to IIRC for help and it developed that if you used d-feet rather than qdbusviewer, it worked. Also, if you ran dbus-monitor you could see that qdbusviewer was setting up the call improperly and not reporting the subsequent error. Wanting to be a good Open Source citizen, I wrote a bug report thinking that Nokia would both be grateful for having the problem brought to their attention and willing to investigate and try to fix the problem. It didn't turn out that way. What seemed to happen, from my point of view, is that the bug report got slammed into the "Need More Info" state every time it hit the maintenance engineer's in-box with no attempt to investigate the problem from a technical prospective. It even reached the point where (and I honestly don't think I am exaggerating here) the "missing information" could not have been more than a few feet or few minutes away from the maintenance engineer. Nokia does not provide a means whereby one can request more senior review of a bug report "resolution". So, I would like some people with experience in this sort of thing to take a look and render an opinion regarding whether or not the bug report contained a reasonable level of detail and whether or not the maintenance engineer could have "taken it from there" with a reasonable expectation of resolving the issue with a fix.
I thank you in advance.
http://bugreports.qt.nokia.com/browse/QTBUG-14664?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
- dkersten 16y agoI only glanced at it but it looks like you were saying that the problem is caused when using qdbusviewer together with rythembox, but neither of the nokia guys have rythembox and requested a Qt-only project to use instead of rythembox which demonstrates your bug, which you didn't provide them with. They also said that, from the log you provided, the problem appears to be with rythembox and not with Qt - ie, it is not a Qt bug - and asked you to demonstrate that the bug is on Qt's side with a Qt-only project. They did ask you for a Qt-only demonstration of the bug a number of times and each time you didn't provide one. Their actions seem quite reasonable to me. I guess you could spell it out for them by posting the d-feet and dbus-monitor log and showing exactly where qdbusviewer gets it wrong (and d-feet gets it right). Better yet: do as they ask and reproduce the problem with a Qt-only project, instead of rythembox.
- jda2000 16y agoUntil very recently, if you had Ubuntu, you had Rhythmbox.
- dkersten 16y agoCounter argument: I'm a heavy linux user, yet I don't have rythembox. Seriously, who knows why they won't just install it. Perhaps its because of a nokia program-installation policy. Who knows. I don't think we should expect that they will have it or are able to install it.
- jda2000 16y agoHowever, time permitting. I guess I will send them the dbus-monitor trace of the d-feet call.
- dkersten 16y agoHopefully that will be enough to convince them.
- jda2000 16y agoLet's hope so. What they were asking me to do was to rewrite a small portion of Rhythmbox as a QT app to expose the bug. I still think it was more reasonable for them to just get Rhythmbox from the repository and run qdbusviewer under a debugger. Did I mention that I am not a QT programmer?
- dkersten 16y agoI get that and understand that (if they're on a debian based system) they can probably just apt-get install rythembox and test it for themselves (I don't use a debian-based system, but I imagine I can get it just as easily with yaourt -S rythembox). That would be the easiest. But, for whatever reason they cannot or will not install it and I don't think we can criticize them for that decision. Anyway, since you were saying that it is definitely not rythemboxes fault, I would suggest you 1) show the dbus-monitor trace of the d-feet so that they can see that rythembox and d-feet do work perfectly and 2) show them, again, the dbus-monitor output for qdbusviewer, in as much detail as you can and clearly highlighting where it diverges from the expected. If that isn't enough for them, then they definitely are being stubborn... The sad thing is, even if you do as they ask and write a Qt program, it really won't give them anything over what they'd have if you give them logs, since, as you say, the problem is not with rythembox. Sigh. In any case, good luck.
- jda2000 16y agoSince they closed the last bug report I had to open a new one to provide the dbus-monitor logs: http://bugreports.qt.nokia.com/browse/QTBUG-16529 http://bugreports.qt.nokia.com/browse/QTBUG-16529 These logs prove to a level of mathematical certainty that the bug is somewhere in qdbusviewer rather than the original bug report where it was merely obvious.... If they still insist on "QT only" bug reports (a bit contrary to the spirit of DBUS), I may have to ask for help from an experienced QT programmer.