Why is "no code allowed to be all ones" in libjpeg's Huffman decoding? -

i'm trying satisfy myself meteosat images i'm getting ftp server valid images. doubt arises because tools i've used far complain "bogus huffman table definition" - yet when comment out error message, image appears quite plausible (a greyscale segment of earth's disc).

from https://github.com/libjpeg-turbo/libjpeg-turbo/blob/jpeg-8d/jdhuff.c#l379:

while (huffsize[p]) {   while (((int) huffsize[p]) == si) {     huffcode[p++] = code;     code++;   }   /* code 1 more last code used codelength si;    * must still fit in si bits, since no code allowed ones.    */   if (((int32) code) >= (((int32) 1) << si))     errexit(cinfo, jerr_bad_huff_table);   code <<= 1;   si++; } 

if comment out check, or add check huffsize[p] nonzero (as in containing loop's controlling expression), djpeg manages convert image bmp can view few problems.

why comment claim all-ones codes not allowed?

it claims because not allowed. doesn't mean there can't images out there don't comply standard.

the reason not allowed (from standard):

making entropy-coded segments integer number of bytes performed follows: huffman coding, 1-bits used, if necessary, pad end of compressed data complete final byte of segment.

if 1's code allowed, end ambiguity in last byte of compressed data padded 1's coded symbol.