Hallo, B. Bogart hat gesagt: // B. Bogart wrote:

I posted this a while back, I fixed it by turning typetags off.

I could swear that I already tried this, but it seems not.

Problem is I can't send file paths with typetags off so I switch it on and off depending on the message I send (yuck!)

Martin Peach posted a patch/fix I think, check the archives.

Does the maintainer for OSCx read this list??

Yes, I guess, but if I find the patch, I'll try to put it into the CVS, if noone minds (and if I manage to do that...)

ciao

Frank Barknecht _ ______footils.org__

Notice that this fix was for the problem with negative ints. The floats would be repaired like this: around line 750: ### case f: { int i = ntohl(*((int *)p)); to case f: { int i = (signed)ntohl(*((int *)p)); ### Now negative ints and floats seem to work, for me at least... Martin

Frank Barknecht wrote:

Hallo, B. Bogart hat gesagt: // B. Bogart wrote:

...

I posted this a while back, I fixed it by turning typetags off.

Problem is I can't send file paths with typetags off so I switch it on and off depending on the message I send (yuck!)

Martin Peach posted a patch/fix I think, check the archives.

Thanks, I've found it, Martin wrote: ### In OSCx/OSC/dumpOSC.c line 743, in the function dumpOSC_PrintTypeTaggedArgs():

change SETFLOAT(mya+myargc,ntohl(*((int *) p))); to SETFLOAT(mya+myargc,(signed)ntohl(*((int *) p)));

...since ntohl returns an unsigned integer, you get 0xFFFFFFFF for -1, which equals 4294967295 decimal... ###

I'll try this tomorrow and after that will commit it to CVS.

ciao

In your original post you said "it occured to me, that sending negative floats over OSC doesn't work here " ... I admit I never tried it myself before 'fixing' it, but it still works...

Perhaps in the longer term it would be better to send XML format such as <float>-1.2345</float> so that the vagaries of stuffing machine-dependent representations of floats into machine-dependent representations of integers can be completely avoided, and other formats (like imaginary numbers and expressions) could be included as well. That would require a complete rewrite of the OSC standard though....

Martin

Frank Barknecht wrote:

Hallo, Martin Peach hat gesagt: // Martin Peach wrote:

...

Notice that this fix was for the problem with negative ints. The floats would be repaired like this: around line 750: ### case f: { int i = ntohl(*((int *)p)); to case f: { int i = (signed)ntohl(*((int *)p)); ### Now negative ints and floats seem to work, for me at least...

Floats seemed to work okay already, I thought. Didn't they?

I've put your first patch into the CVS now.

ciao