From: alin w. <ali...@ca...> - 2005-05-08 13:18:11
|
=A0 > De: Mathias Lundgren > A: alin weiller > Objet: Re: [Lmuse-develope= r] Sysex transfer in MESS softsynth > Date: 08 May 2005 09:57:05 +0200 > Hi there! > > s=F6n 2005-05-08 klockan 01.07 skrev alin weiller: > > Hi the= re , > > I am trying to add the "save in project" feature in DeicsOnze, > >=20 I meet the following problem : > > The sysex data array transmited in the i > > nitialization of the project in cut after the first '\0'. I unders tand th > > at it is useful because '\0' designs the end of a string, how ever it is n > > ot necessary since the length of the data array is defin ed in the sysex e > > vent. In my case, having data cut after '\0' is not convenient. Would tha > > t be possible to change that? If you think so, I can make the work, I hav > > en't found yet where is the function that cut the data, help would be wel > > come. > > > > That sounds strange, t he length of the init event should be specified in > the length-parameter that is passed to the Mess in the getInitData(int* > length, const unsig ned char** data), and it sounds like a bug if it > doesn't follow the len gth that you specify. I did some updates on > simpledrums yesterday, in t he 0.7-branch, and everything seems to be > working there. When the data=20 is gotten from the softsynth to the med file when saving everything is wo rking well, but when the sysex message is sent back to the softsynth when loading, then it is seems to be cut after '\0'. Now I am going to try wi th the 0.7 branch. > > > > Another thing, I think it would better to orga nize the data in t > > he med file gotten from getInitData, in a sequence like for instance 0344 > > 3516f55a instead of > > 03 > > 44 > > 35 > >=20 > > 16 > > f5 > > 5a > > the current organization multiply > > by 10 the=20 size of the softsynth data, with the other organization it shou > > ld be only twice. If you don't have any argument against that I can make > > t he work also. > > > > I'm not sure about this, I've found it very helpful to be able to look > at the data in the .med-file when ironing out bugs. Earlier, the data > looked something like this: > > 03 44 35 16 f5 5a >=20 . .. .. This one is probably the best and multiplies only 3 times. /A-LI N > > ...Not sure if Qt:s XML-stuff has anything to do with this. I think that > looked better. Having everything written together like 03443516f5 5a > would make it difficult to read (since everything is written in text , a > hex-editor wouldn't help in that case either). > > > Note that thes e remarks are based on the CVS of the 0.8 > > branch downloaded for 1 mon th ago (I should compile the last CVS soon bu > > t not yet). > > > > Las t time (yesterday) I tried latest CVS I didn't get the softsynths to > wo rk (actually didn't have any midi output at all) - CVS of the > 0.8-branc h from 1 month ago is more "stable". I didn't make any deep > investigati on though, but something tells me there are quite a few loose > ends ther e, since Werner has a new midi routing concept coming up (it > looks grea t btw). I think the best thing is to keep working on Deicsonze > in the 0 7-branch and then make all the necessary modifications for 0.8 > when 0. 8 has started to spin again. > > /Mathias > > > Regards, > > > > A-LIN >=20 > > > Protek-on: CaraMail met en oeuvre un no > > uveau Concept de S=E9cu= ri t=E9 Globale - www.caramail.com Protek-on: CaraMail met en oeuvre un nouveau Concept de S=E9curit=E9 Glob= ale - www.caramail.com |