What causes "JPEG quantization tables are numbered 0..%d " in DrKLO/Telegram?

Use only table indices 0 through NUM_QUANT_TBLS-1 (typically 0-3).

JPEG quantization tables are numbered 0..%d

Error message

JPEG quantization tables are numbered 0..%d

What it means

The set_quant_slots function processes the -qslots argument, which maps color components to quantization table indices. JPEG supports NUM_QUANT_TBLS tables (typically 4, numbered 0-3). If any table index in the argument is negative or >= NUM_QUANT_TBLS, the function prints this error and returns FALSE. This validates the component-to-table mapping before compression begins.

Source

Thrown at TMessagesProj/jni/mozjpeg/rdswitch.c:590

set_quant_slots (j_compress_ptr cinfo, char *arg)
/* Process a quantization-table-selectors parameter string, of the form
 *     N[,N,...]
 * If there are more components than parameters, the last value is replicated.
 */
{
  int val = 0;                  /* default table # */
  int ci;
  char ch;

  for (ci = 0; ci < MAX_COMPONENTS; ci++) {
    if (*arg) {
      ch = ',';                 /* if not set by sscanf, will be ',' */
      if (sscanf(arg, "%d%c", &val, &ch) < 1)
        return FALSE;
      if (ch != ',')            /* syntax check */
        return FALSE;
      if (val < 0 || val >= NUM_QUANT_TBLS) {
        fprintf(stderr, "JPEG quantization tables are numbered 0..%d\n",
                NUM_QUANT_TBLS-1);
        return FALSE;
      }
      cinfo->comp_info[ci].quant_tbl_no = val;
      while (*arg && *arg++ != ','); /* advance to next segment of arg
                                        string */
    } else {
      /* reached end of parameter, set remaining components to last table */
      cinfo->comp_info[ci].quant_tbl_no = val;
    }
  }
  return TRUE;
}


GLOBAL(boolean)
set_sample_factors (j_compress_ptr cinfo, char *arg)
/* Process a sample-factors parameter string, of the form

View on GitHub (pinned to 45ab8f4308)

Solutions

Use only table indices 0 through NUM_QUANT_TBLS-1 (typically 0-3).
Remember indices are 0-based: 0=luminance, 1=chrominance is the standard mapping.
For a standard YCbCr image, -qslots 0,1,1 maps Y to table 0 and Cb/Cr to table 1.

Example fix

# before
cjpeg -qslots 0,5 image.bmp > out.jpg
# after
cjpeg -qslots 0,1,1 image.bmp > out.jpg

Defensive patterns

Strategy: validation

Validate before calling

// Pre-validate qslots argument before calling set_quant_slots
static boolean validate_qslots_arg(const char *arg) {
  const char *p = arg;
  while (*p) {
    int val;
    char ch = ',';
    if (sscanf(p, "%d%c", &val, &ch) < 1) return FALSE;
    if (val < 0 || val >= NUM_QUANT_TBLS) return FALSE;
    while (*p && *p++ != ',') ;
  }
  return TRUE;
}

Prevention

Use only table indices 0 to NUM_QUANT_TBLS-1 (typically 0-3).
Remember indices are 0-based.
For standard YCbCr: -qslots 0,1,1 is the default luminance/chrominance mapping.
Validate each comma-separated value is in range before passing.

When it happens

Trigger: Passing -qslots with a value outside 0..NUM_QUANT_TBLS-1 (typically 0..3). For example 'cjpeg -qslots 4 image.bmp' or '-qslots 0,5' where 5 exceeds the table limit. The sscanf parses the integer and the bounds check fires.

Common situations: Misunderstanding the number of available quantization tables; passing a 1-based index instead of 0-based; using a table number that was defined in a different build or library version; typos in the argument string.

Related errors

AI-assisted analysis of DrKLO/Telegram@45ab8f4308 (2026-08-14). Data as JSON: /api/errors/abf53159398fa0fc. Report an issue: GitHub.