SIGSEGV (Address boundary error)

[xsh005]

tldr:

local fzy = require("fzy")
fzy.filter("foo", { string.rep(" ", 1039) })
-- 1039 and below: no problem
-- 1040/1041: LuaJIT ASSERT lj_api.c:162: lua_settop: bad stack slot -2
-- fish: Job 1, 'nvim -l test.lua' terminated by signal SIGABRT (Abort)
-- 1042 and above: fish: Job 1, 'nvim -l test.lua' terminated by signal SIGSEGV (Address boundary error)

long story:

Hello! I am experimenting with this library and implementing a fuzzy finder frontend. When I was testing on a big repo, I noticed that my little script segfault-ed. I am able to narrow the problem down to one file:

My setup:

-rw-r--r--. 1 xiaoshihou xiaoshihou    205 Jan 20 16:10 fzy.lua
-rwxr-xr-x. 1 xiaoshihou xiaoshihou  29688 Jan 20 16:10 fzy_native.so
-rw-r--r--. 1 xiaoshihou xiaoshihou    337 Jan 20 16:23 test.lua
-rw-r--r--. 1 xiaoshihou xiaoshihou 158538 Jan 20 16:20 test.txt

test.txt is this file here

test.lua:

local fzy = require("fzy")
local uv = vim.uv
local path = "/tmp/test.txt"

local fd = assert(uv.fs_open(path, "r", 438))
local stat = assert(uv.fs_fstat(fd))
local data = assert(uv.fs_read(fd, stat.size, 0))
assert(uv.fs_close(fd))
data = vim.split(data, "\n") -- this outputs fine if I print it
fzy.filter("ex", data)

Then:

❯ nvim -l test.lua
fish: Job 1, 'nvim -l test.lua' terminated by signal SIGSEGV (Address boundary error)

I know there's a bunch of horrible \xc0 stuff in the file, but that shouldn't make fzy segfault 🤔
Thank you for your time!

[xsh005]

update: this particular line would segfault

     decoderArray = ConvArray "\x0\x0\x1\x0\x2\x0\x3\x0\x9c\x0\x9\x0\x86\x0\x7f\x0\x97\x0\x8d\x0\x8e\x0\xb\x0\xc\x0\xd\x0\xe\x0\xf\x0\x10\x0\x11\x0\x12\x0\x13\x0\x9d\x0\x85\x0\x8\x0\x87\x0\x18\x0\x19\x0\x92\x0\x8f\x0\x1c\x0\x1d\x0\x1e\x0\x1f\x0\x80\x0\x81\x0\x82\x0\x83\x0\x84\x0\xa\x0\x17\x0\x1b\x0\x88\x0\x89\x0\x8a\x0\x8b\x0\x8c\x0\x5\x0\x6\x0\x7\x0\x90\x0\x91\x0\x16\x0\x93\x0\x94\x0\x95\x0\x96\x0\x4\x0\x98\x0\x99\x0\x9a\x0\x9b\x0\x14\x0\x15\x0\x9e\x0\x1a\x0\x20\x0\xa0\x0\xe2\x0\xe4\x0\xe0\x0\xe1\x0\xe3\x0\xe5\x0\xe7\x0\xf1\x0\xa2\x0\x2e\x0\x3c\x0\x28\x0\x2b\x0\x7c\x0\x26\x0\xe9\x0\xea\x0\xeb\x0\xe8\x0\xed\x0\xee\x0\xef\x0\xec\x0\xdf\x0\x21\x0\x24\x0\x2a\x0\x29\x0\x3b\x0\xac\x0\x2d\x0\x2f\x0\xc2\x0\xc4\x0\xc0\x0\xc1\x0\xc3\x0\xc5\x0\xc7\x0\xd1\x0\xa6\x0\x2c\x0\x25\x0\x5f\x0\x3e\x0\x3f\x0\xf8\x0\xc9\x0\xca\x0\xcb\x0\xc8\x0\xcd\x0\xce\x0\xcf\x0\xcc\x0\x60\x0\x3a\x0\x23\x0\x40\x0\x27\x0\x3d\x0\x22\x0\xd8\x0\x61\x0\x62\x0\x63\x0\x64\x0\x65\x0\x66\x0\x67\x0\x68\x0\x69\x0\xab\x0\xbb\x0\xf0\x0\xfd\x0\xfe\x0\xb1\x0\xb0\x0\x6a\x0\x6b\x0\x6c\x0\x6d\x0\x6e\x0\x6f\x0\x70\x0\x71\x0\x72\x0\xaa\x0\xba\x0\xe6\x0\xb8\x0\xc6\x0\xa4\x0\xb5\x0\x7e\x0\x73\x0\x74\x0\x75\x0\x76\x0\x77\x0\x78\x0\x79\x0\x7a\x0\xa1\x0\xbf\x0\xd0\x0\xdd\x0\xde\x0\xae\x0\x5e\x0\xa3\x0\xa5\x0\xb7\x0\xa9\x0\xa7\x0\xb6\x0\xbc\x0\xbd\x0\xbe\x0\x5b\x0\x5d\x0\xaf\x0\xa8\x0\xb4\x0\xd7\x0\x7b\x0\x41\x0\x42\x0\x43\x0\x44\x0\x45\x0\x46\x0\x47\x0\x48\x0\x49\x0\xad\x0\xf4\x0\xf6\x0\xf2\x0\xf3\x0\xf5\x0\x7d\x0\x4a\x0\x4b\x0\x4c\x0\x4d\x0\x4e\x0\x4f\x0\x50\x0\x51\x0\x52\x0\xb9\x0\xfb\x0\xfc\x0\xf9\x0\xfa\x0\xff\x0\x5c\x0\xf7\x0\x53\x0\x54\x0\x55\x0\x56\x0\x57\x0\x58\x0\x59\x0\x5a\x0\xb2\x0\xd4\x0\xd6\x0\xd2\x0\xd3\x0\xd5\x0\x30\x0\x31\x0\x32\x0\x33\x0\x34\x0\x35\x0\x36\x0\x37\x0\x38\x0\x39\x0\xb3\x0\xdb\x0\xdc\x0\xd9\x0\xda\x0\x9f\x0"#

local fzy = require("fzy")
local str =
[[     decoderArray = ConvArray "\x0\x0\x1\x0\x2\x0\x3\x0\x9c\x0\x9\x0\x86\x0\x7f\x0\x97\x0\x8d\x0\x8e\x0\xb\x0\xc\x0\xd\x0\xe\x0\xf\x0\x10\x0\x11\x0\x12\x0\x13\x0\x9d\x0\x85\x0\x8\x0\x87\x0\x18\x0\x19\x0\x92\x0\x8f\x0\x1c\x0\x1d\x0\x1e\x0\x1f\x0\x80\x0\x81\x0\x82\x0\x83\x0\x84\x0\xa\x0\x17\x0\x1b\x0\x88\x0\x89\x0\x8a\x0\x8b\x0\x8c\x0\x5\x0\x6\x0\x7\x0\x90\x0\x91\x0\x16\x0\x93\x0\x94\x0\x95\x0\x96\x0\x4\x0\x98\x0\x99\x0\x9a\x0\x9b\x0\x14\x0\x15\x0\x9e\x0\x1a\x0\x20\x0\xa0\x0\xe2\x0\xe4\x0\xe0\x0\xe1\x0\xe3\x0\xe5\x0\xe7\x0\xf1\x0\xa2\x0\x2e\x0\x3c\x0\x28\x0\x2b\x0\x7c\x0\x26\x0\xe9\x0\xea\x0\xeb\x0\xe8\x0\xed\x0\xee\x0\xef\x0\xec\x0\xdf\x0\x21\x0\x24\x0\x2a\x0\x29\x0\x3b\x0\xac\x0\x2d\x0\x2f\x0\xc2\x0\xc4\x0\xc0\x0\xc1\x0\xc3\x0\xc5\x0\xc7\x0\xd1\x0\xa6\x0\x2c\x0\x25\x0\x5f\x0\x3e\x0\x3f\x0\xf8\x0\xc9\x0\xca\x0\xcb\x0\xc8\x0\xcd\x0\xce\x0\xcf\x0\xcc\x0\x60\x0\x3a\x0\x23\x0\x40\x0\x27\x0\x3d\x0\x22\x0\xd8\x0\x61\x0\x62\x0\x63\x0\x64\x0\x65\x0\x66\x0\x67\x0\x68\x0\x69\x0\xab\x0\xbb\x0\xf0\x0\xfd\x0\xfe\x0\xb1\x0\xb0\x0\x6a\x0\x6b\x0\x6c\x0\x6d\x0\x6e\x0\x6f\x0\x70\x0\x71\x0\x72\x0\xaa\x0\xba\x0\xe6\x0\xb8\x0\xc6\x0\xa4\x0\xb5\x0\x7e\x0\x73\x0\x74\x0\x75\x0\x76\x0\x77\x0\x78\x0\x79\x0\x7a\x0\xa1\x0\xbf\x0\xd0\x0\xdd\x0\xde\x0\xae\x0\x5e\x0\xa3\x0\xa5\x0\xb7\x0\xa9\x0\xa7\x0\xb6\x0\xbc\x0\xbd\x0\xbe\x0\x5b\x0\x5d\x0\xaf\x0\xa8\x0\xb4\x0\xd7\x0\x7b\x0\x41\x0\x42\x0\x43\x0\x44\x0\x45\x0\x46\x0\x47\x0\x48\x0\x49\x0\xad\x0\xf4\x0\xf6\x0\xf2\x0\xf3\x0\xf5\x0\x7d\x0\x4a\x0\x4b\x0\x4c\x0\x4d\x0\x4e\x0\x4f\x0\x50\x0\x51\x0\x52\x0\xb9\x0\xfb\x0\xfc\x0\xf9\x0\xfa\x0\xff\x0\x5c\x0\xf7\x0\x53\x0\x54\x0\x55\x0\x56\x0\x57\x0\x58\x0\x59\x0\x5a\x0\xb2\x0\xd4\x0\xd6\x0\xd2\x0\xd3\x0\xd5\x0\x30\x0\x31\x0\x32\x0\x33\x0\x34\x0\x35\x0\x36\x0\x37\x0\x38\x0\x39\x0\xb3\x0\xdb\x0\xdc\x0\xd9\x0\xda\x0\x9f\x0"#]]
fzy.filter("foo", { str })