On Sat, May 19, 2012 at 08:51:42PM +0200, gabor papp wrote: > maybe valgrind would help:
Valgrind has some sort of problem and pretends he's done before I get a scratchpad window. Before that I get some interesting things, like this; ==4253== Conditional jump or move depends on uninitialised value(s) ==4253== at 0x46DDBC2: GC_mark2 (newgc.c:2988) ==4253== by 0x46DE664: GC_mark2_variable_stack (var_stack.c:73) ==4253== by 0x46DE7A4: GC_mark_variable_stack (newgc.c:2031) ==4253== by 0x46DB889: garbage_collect (newgc.c:4492) ==4253== by 0x46E07CE: GC_switch_out_master_gc (newgc.c:2808) ==4253== by 0x44DB25E: scheme_engine_instance_init (env.c:283) ==4253== by 0x44DE91C: scheme_basic_env (env.c:221) ==4253== by 0x80565C0: fluxus::Interpreter::Initialise() (Interpreter.cpp:83) ==4253== by 0x8063D2B: run(void*) (main.cpp:270) ==4253== by 0x44ABED6: scheme_main_stack_setup (salloc.c:194) ==4253== by 0x8061FC0: main (main.cpp:437) ==4253== ...but nothing that looks relevant to the midi position. I'll build a version from plain master to rule out that some of the redacted MIDI stuff somehow interferes with this. Kas.
