From: D. M. M. <mic...@ro...> - 2009-12-26 15:57:49
|
On Saturday 26 December 2009, Heinz Wiesinger wrote: > Very nice! > I had a go at building the beta on Slackware and it was no problem at all. > Thumbs up! :) Good stuff! > It seems that the editing area of the notation view is limited to a fixed > size. Everything below a certain point is black. (screenshot here: > http://www.liwjatan.at/files/screenshots/rosegarden-notation-view.png) I forgot about trying to do something with that one. It doesn't react to resizing in the vertical axis well, or at all. Thanks for the reminder. > Also there's a rendering issue when I maximize the main window after a > fresh start. It works fine if I drag the lower edge of the window > manually, but if I click on the maximize icon in the window decoration it > leaves garbage. (screenshot here: > http://www.liwjatan.at/files/screenshots/rosegarden-main- window.png) That one I can't repeat, and your setup is quite similar to mine apart from the gcc version. That doesn't give me anything I could compile and test against, so I don't have much hope of sorting that out. I'm left having to hope that you can get around that problem using the native graphics system. Try: ./rosegarden -graphicssystem native If that works, I'm going to have to revisit the idea of making the graphics system configurable through the preferences interface. That particular detail can't be stored in QSettings, so I'll have to come up with some workaround. > Last is a package issue. "make install" runs update-mime-database at the > end. For packaging rosegarden I would need to patch the Makefile to remove > that command, because update-mime-database should be called on > installation of the package, and not at the end of "make install". I can't > really think of a better solution than patching right now, but maybe > someone else has a good idea on how we can avoid that. Is running update-mime-database out of an install target abnormal and wrong for every distro? I don't think I'd ever written an install target from scratch, and it's entirely possible that was just a stupid thing to do. I don't remember if our old build system used to do that at some point or not. I'll be happy to honor whatever the best standard practice is. I'm not trying to make life hard for maintainers. > I can also open reports in the bugtracker if it's better located there. It isn't necessary to open bugs on the tracker, but I think we're at a point where it's OK to go ahead and start using it again, and I removed the "don't use the tracker yet" instructions from the latest round of texts I published. At this point the main thing I care about is reporting the bugs *somewhere* I can see them, and this was just fine. Thank you for the reports! > Lastly, out of curiosity, I recognized the cmake build system is gone. Will > it come back at some point or will you stay with autotools? We were all building with some static hand-written makefile for months in preference to anybody trying to get CMake working again. It fell on my shoulders to get it to go, and I was just not able to do so. Chris had an alternate plan for doing up a simple Autotools based build system, and we went with that and got it to work. I'm not strongly opposed to going back to CMake, but I am strongly opposed to having a build system that's dependent on one single individual to maintain. People lose interest and move away sometimes. If Chris or I lose interest and move away, the project is most likely dead, so if we have a build system both Chris and I can maintain ourselves, that's as close as we can come to guaranteeing somebody will always be around who knows how to maintain it. So I think we're probably sticking with Autotools, with some sad regret about having to dump such a nice thing as we had, and some relief about maintainability over time mixed in as well. After all, we're here to write software first, not build systems. If the thing compiles, it gets the job done. -- D. Michael McIntyre |