Hi Robert, 20.01.2015 11:30, Robert Jonsson wrote: > 2015-01-20 7:06 GMT+01:00 Andrew Deryabin <[email protected]>: >> Hi guys! >> >> For about a week I was testing qt5 port to be sure that it can be used >> in everyday life. And now I can say - yes, it can! Port is working as >> stable as qt4 version - no random crashes, all gui elements are >> functional and it seems, that gui became more responsive. >> > Awesome man! > So, my prophecy is that we will meet a tall dark stranger, err sorry, > I mean: release a new MusE release in the recent future :) Yeah, may be it will be a spring present? :) >> As usual, there is a new demo song made from scratch only in qt5 MusE >> version: >> http://www.youtube.com/watch?v=Wvx_-ND0iSM >> >> I also added it to the demos page: >> http://muse-sequencer.org/index.php/Demos >> >> Here are the list of recent changes (all available in qt5 branch): >> - Ported code to Qt5. Port works rather stable - no crashes >> and artefacts >> during 4 days testing (danvd) >> - Added scrollable submenus option to global settings (danvd) >> - Added CPU load toolbar (danvd) >> - Optimized RT Fifo class used for lv2 plugin <-> UI >> communication. >> This reduced CPU load by 1-2% :) (danvd) >> - Integrated Presets submenu into midi track's programs menu >> for LV2 synths (danvd) >> - Implemented presets saving and updating for LV2 synths/rack >> plugins (danvd) >> - LV2 presets are stored in common format and can be >> shared between lv2 hosts (danvd) > Many cool improvements, great! > >> - Integrated lv2,lilv,sord,serd,sratom,zix libraries into >> MusE's source tree. >> Now MusE can be compiled without external LV2 dependences at >> all (danvd) >> - Reworked Appearance::defaultStyle static QString variable >> (static var init fiasco fix) (danvd) > There was atleast one report of someone failing to start MusE until > -style was set during startup, this wouldn't happen to have to do with > that? It's very likely - random crashes at startup and/or shutdown were due to not proper initialization of this variable by the compiler. And it's interesting, that recompilation can fix this error, and another - to bring it back. > >> The more noticeable are: >> 1. Now all lv2 dependences (liblv2, lilv,serd,sord,sratom) are bundled >> with MusE source. I adopted code to cmake. The main reason of this was >> the lack of lilv_world_unload_resource function in earlier lilv versions >> that made it impossible to correctly support dynamic presets update and >> save. > Could you explain a bit more about what did not work? > Maybe we need to discuss if we really should do this, aren't there > potential drawbacks, like compatibility with already present > libraries? I spent several hours to make this decision. Here is a reason: There is a function called 'lilv_world_unload_resource' which makes it possible to dynamically update presets after saving new or replacing existing. It simply can unload all entries associated with given resource from internal library cache. That bad thing is that this function is only available in later lilv versions (may be in the last). Only my system (arch linux) has it, but different ubuntas, fedoras etc. uses version without this function. So there were 3 variants: 1) to make hacks for existing lib version (which I really don't like at all) 2) to drop presets save/update function (no, no, no! :) ) 3) to bundle library source into muse. I've chosen the last. What about conflicts with installed versions - there should be none. These libraries are compiled into static lib, which is then linked to libmusecore.so and all include paths are sent to compiler before system ones. It's like with suil library - I had to rewrite gui support from scratch, but in the case of turtle/strings handling - I don't want to dive into that nightmare :). > >> 2. New cpu load toolbar. Now it uses jack cpu info, but I want to make >> internal performance testers for every synth and plugin. >> >> >> Now in plans is to get rid of warnings generated by clang compiler (over >> 2000) - this problem is common to master qt4 branch also. >> >> May be it's an early question but.. What do you think about switching to >> qt5 in master for the main development? > Indeed, this is the future :) > Tim, how is it going, walking the tightrope? Still standing? Yes, connections restoring will be a great improvement, I already want to test it! :)
Regards, Andrew >> P.S. I've already applied all recent changes in master to qt5 branch. >> >> -- >> Regards, >> Andrew >> > Regards, > Robert > > ------------------------------------------------------------------------------ > New Year. New Location. New Benefits. New Data Center in Ashburn, VA. > GigeNET is offering a free month of service with a new server in Ashburn. > Choose from 2 high performing configs, both with 100TB of bandwidth. > Higher redundancy.Lower latency.Increased capacity.Completely compliant. > http://p.sf.net/sfu/gigenet > _______________________________________________ > Lmuse-developer mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/lmuse-developer -- Regards, Andrew ------------------------------------------------------------------------------ New Year. New Location. New Benefits. New Data Center in Ashburn, VA. GigeNET is offering a free month of service with a new server in Ashburn. Choose from 2 high performing configs, both with 100TB of bandwidth. Higher redundancy.Lower latency.Increased capacity.Completely compliant. http://p.sf.net/sfu/gigenet _______________________________________________ Lmuse-developer mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/lmuse-developer
