<oriansj>Melg8[m]: looks like I solve the Global buffer overflow in collect_weird_string and only have 1 left to go (assuming the other 2 are good) <stikonas>oriansj: you probably want to close solved issues... <stikonas>or are you waiting for Melg8[m] to confirm... <oriansj>stikonas: I try to confirm if possible before closing issues as I believe issues need to be fully addressed. The only time I close an issue without discussing it first is if they don't respond within a reasonable amount of time (6months usually) <oriansj>Although without the -fsanitize=address it does show such memory leaks but honestly some of them don't seem valid like: by 0x10A466: new_primitive (cc_types.c:60) (which is allocating a type node found in the global types list) unless I am supposed to allocate more than just the size of the struct??? <Melg8[m]>oriansj: valgrind and sanitizers doesn't work together, so it meant to be one or another - without address sanitizer but with valgrind, or compile with sanitizer, but run without valgrind. Usually they report mem leaks if something were allocated with malloc/calloc - call, but never got deallocated with free until the end of program. <oriansj>in which case, we absolutely do not free anything in M2-Planet because everything allocated into memory *might* be needed for introspection or absolutely *is needed* for basic function of the compiler. <oriansj>and going through the trouble of freeing everything right before an exit doesn't seem that useful. ***ChanServ sets mode: +o oriansj