A syntax to match a specific layer in the protocol stack has been added. For example “ip.addr#2 == 1.1.1.1” matches only the inner layer in an IP-over-IP packet.

Set elements must be separated using a comma, e.g: {1, 2, "foo"}. Using only whitespace as a separator was deprecated in 3.6 and is now a syntax error.

Support for some additional character escape sequences in double quoted strings has been added. Along with octal (\<number>) and hex (\x<number>) encoding, the following C escape sequences are now supported with the same meaning: \a, \b, \f, \n, \r, \t, \v. Previously they were only supported with character constants.

Unrecognized escape sequences are now treated as a syntax error. Previously they were treated as a literal character. In addition to the sequences indicated above, backslash, single quotation and double quotation mark are also valid sequences: \, \', \".

The display filter engine now uses PCRE2 instead of GRegex (GLib’s bindings to the older and end-of-life PCRE library). PCRE2 is compatible with PCRE so any user-visible changes should be minimal. Some exotic patterns may now be invalid and require rewriting.

A new strict equality operator "===" or "all_eq" has been added. The expression "a === b" is true if and only if all a’s are equal to b. The negation of "===" can now be written as "!==" (any_ne).

The aliases "any_eq" for "==" and "all_ne" for "!=" have been added.

The operator "~=" is deprecated and will be removed in a future version. Use "!==", which has the same meaning instead.

Dates and times can be given in UTC using ISO 8601 (with 'Z' timezone) or by appending the suffix "UTC" to the legacy formats. Otherwise local time is used.

Integer literal constants may be written in binary (in addition to decimal/octal/hexadecimal) using the prefix "0b" or "0B".

A new syntax to disambiguate literals from identifiers has been added. Every value with a leading dot is a protocol or protocol field. Every value with a leading colon or in between angle brackets is a literal value. See the User’s Guide for details.

Floats must be written with a leading and ending digit. For example the values ".7" and "7." are now invalid as floats. They must be written "0.7" and "7.0" respectively.

The "bitwise and" operator is now a first-class bit operator, not a boolean operator. In particular this means it is now possible to mask bits, e.g.: frame[0] & 0x0F == 3.

Arithmetic is supported for numeric fields with the usual operators “+”, “-”, “*”, “/”, and “%”. Arithmetic expressions must be grouped using curly brackets (not parenthesis).

Logical AND now has higher precedence than logical OR, in line with most programming languages.

New display filter functions max(), min() and abs() have been added.

Functions can accept expressions as arguments, including other functions. Previously only protocol fields and slices were syntactically valid function arguments.

text2pcap supports writing the output file in all the capture file formats that wiretap library supports, using the same -F option as editcap, mergecap, and tshark.

text2pcap supports selecting the encapsulation type of the output file format using the wiretap library short names with an -E option, similiar to the -T option of editcap.

text2pcap has been updated to use the new logging output options and the -d flag has been removed. The "debug" log level corresponds to the old -d flag, and the "noisy" log level corresponds to using -d multiple times.

text2pcap and “Import from Hex Dump” support writing fake IP, TCP, UDP, and SCTP headers to files with Raw IP, Raw IPv4, and Raw IPv6 encapsulations, in addition to Ethernet encapsulation available in previous versions.

text2pcap supports scanning the input file using a custom regular expression, as supported in “Import from Hex Dump” in Wireshark 3.6.x.

In general, text2pcap and wireshark’s “Import from Hex Dump” have feature parity.