src/flisp/equalhash.c reading uninitialized data

[alexharris1976]

This happens with SANITIZE_ADDRESS build of flisp on Aarch64 Linux:

==56350==WARNING: MemorySanitizer: use-of-uninitialized-value
    #0 0xaaaaaab9ca6c in equalhash_lookup_bp_r /home/alg/julia/src/flisp/equalhash.c:22:1
    #1 0xaaaaaab5dcfc in get_type /home/alg/julia/src/flisp/types.c:11:17
    #2 0xaaaaaab73a0c in cvalues_init /home/alg/julia/src/flisp/cvalues.c:883:5
    #3 0xaaaaaab73a0c in lisp_init /home/alg/julia/src/flisp/flisp.c:2404:5
    #4 0xaaaaaab73a0c in fl_init /home/alg/julia/src/flisp/flisp.c:2433:5
    #5 0xaaaaaaba7198 in main /home/alg/julia/src/flisp/flmain.c:34:5
    #6 0xfffff7cd2290 in __libc_start_call_main csu/../sysdeps/nptl/libc_start_call_main.h:58:16
    #7 0xfffff7cd2374 in __libc_start_main csu/../csu/libc-start.c:360:3
    #8 0xaaaaaaab73ac in _start (/home/alg/julia/src/flisp/flisp+0x173ac) (BuildId: b810d08aaa933c974ef7e91bcef250ea18b13445)

equalhash.c is failing with:
   if (tab[index] == ((void *)1)) {

Note that src/support/htable.c htable_new function does not clear memory.

You will probably argue that this likely doesn't matter.

[topolarity]

Note that src/support/htable.c htable_new function does not clear memory.

It looks to me like it does:

julia/src/support/htable.c

Lines 37 to 38 in 796a849

	for (i = 0; i < size; i++)
	h->table[i] = HT_NOTFOUND;

[vtjnash]

I do seem to see many memory bugs like that in relocate_htable, but I don't see that being used in this code path. Maybe there is a aarch64-specific bug in MSan?