Frank, thanks for your answer,

fluidsynth did not get compiled with LADSPA i think as the summary after ./configure indicates:
LADSPA support : no

that's what i got with gdb (as a matter of fact, i am not used to this tool)
btw, i struggled a lit bit just to find these simplistic two first lines in the mailing list, may be we should just add them on puredata.info?

> gdp pd
(gdb) r
Starting program: /usr/local/bin/pd
[Thread debugging using libthread_db enabled]
[New Thread 1076218336 (LWP 7428)]

Program received signal SIGILL, Illegal instruction.
[Switching to Thread 1076218336 (LWP 7428)]
0x40266267 in fluid::fluid_init () from
/usr/local/lib/pd/extra/fluid~.pd_linux
(gdb) bt
#0  0x40266267 in fluid::fluid_init () from /usr/local/lib/pd/extra/fluid~.pd_linux
#1  0x40265da5 in fluid::__init__ () from /usr/local/lib/pd/extra/fluid~.pd_linux
#2  0x40268339 in flext_obj_single::obj_new () from /usr/local/lib/pd/extra/fluid~.pd_linux
#3  0x080a2b6a in pd_typedmess (x=0x8209bd4, s=0x0, argc=1, argv=0x81037e8) at m_class.c:664
#4  0x080a25c4 in new_anything (dummy=0x8220a88, s=0x8209bd4, argc=1, argv=0x81037e8) at m_class.c:530
#5  0x080a27a6 in pd_typedmess (x=0x8209bd4, s=0x0, argc=1, argv=0x81037e8) at m_class.c:749
#6  0x080a5c84 in binbuf_eval (x=0x8220a88, target=0x8209bd4, argc=0, argv=0x0) at m_binbuf.c:574
#7  0x080622c5 in canvas_objtext (gl=0x8227718, xpix=136448648, ypix=136448648, selected=0, b=0x8226088) at g_text.c:94
#8  0x08060639 in canvas_obj (gl=0x8227718, s=0x82136a0, argc=4, argv=0x81037c0) at g_text.c:145
#9  0x080a2b57 in pd_typedmess (x=0x8227718, s=0x0, argc=4, argv=0x81037c0) at m_class.c:665
#10 0x080a5c84 in binbuf_eval (x=0x8220a88, target=0x8227718, argc=0, argv=0x0) at m_binbuf.c:574
#11 0x080a93a3 in binbuf_evalfile (name=0x821fea8, dir=0x821fed0) at m_binbuf.c:1193
#12 0x080a9ba1 in glob_evalfile (ignore=0x8209be0, name=0x821fea8, dir=0x821fed0) at m_binbuf.c:1209
#13 0x080a2960 in pd_typedmess (x=0x8209be0, s=0x0, argc=136445648, argv=0x0) at m_class.c:732
#14 0x080a5c84 in binbuf_eval (x=0x8220a88, target=0x8209be0, argc=0, argv=0x0) at m_binbuf.c:574
#15 0x080ae080 in socketreceiver_read (x=0x821f530, fd=0) at s_inter.c:540
#16 0x080afbbf in sys_domicrosleep (microsec=136448648, pollem=1) at s_inter.c:180
#17 0x080af2e8 in sys_microsleep (microsec=136448648) at s_inter.c:202
#18 0x080aaabf in m_scheduler () at m_sched.c:487
#19 0x080b545b in main (argc=136448648, argv=0x8220a88) at s_entry.c:27




greetings,
vincent

ps: 
strangely enough, i finally managed to run fluid~ and pd39, using an older version of fluid~ i compiled some time ago for pd38.

Frank Barknecht wrote: 
Hallo,
vincent rioux hat gesagt: // vincent rioux wrote:

  
i just tried to compile fluid~ (from externals cvs) and Pd39.2
i also compiled fluidsynth1.0.6.
    

Hopefully you compiled fluidsynth1.0.6 with LADSPA disabled? 

  
the fluid~ lib loads fine at startup but the creation of the fluid~ 
object crashes pd (signal 4).
would it be related to the flext layer may be?
    

If you could run it under a debugger like gdb and send a backtrace,
that would a a lot of help. (How to run gdb was described in the
list archives several times.)

Ciao