Revision history [back]

The test on line 8709 is part of a loop that checks whether the "abbreviation" for a named field contains only:

• ASCII alphabetic characters (A-Z, a-z);
• digits (0-9);
• ASCII dash (hyphen);
• period (.);
• underscore (_).

It also requires that the period not be at the beginning or the end of the abbreviation, that there are not two periods in a row.

So you must update your dissector so that if, in the abbreviations of the fields that it registers, it's using any characters other than the ones listed there, or if it puts a period at the beginning or end of the abbreviation, or if it has two or more periods in a row, it no longer does so.

The descriptive name of the field, which is what's used in the packet details display, is not restricted in that fashion. The abbreviation, which is what's used in packet-matching expressions ("display filters"), and when telling Wireshark or TShark to use certain fields when drawing graphs or when printing particular field values, is restricted in that fashion. So a dissector can, for example, have a field registered as

{ &hf_ip_hdr_len,
  { "Header Length", "ip.hdr_len", FT_UINT8, BASE_DEC,
    NULL, 0x0, "Header length in 32-bit words", HFILL }},

with a space in the description a space in the "blurb" ("Header length in 32-bit words"), but you can't have a field registered as

{ &hf_ip_hdr_len,
  { "Header Length", "ip hdr len", FT_UINT8, BASE_DEC,
    NULL, 0x0, "Header length in 32-bit words", HFILL }},

with spaces in the abbreviation.