Rename proto_construct_dfilter_string() to
authorguy <guy@f5534014-38df-0310-8fa8-9805f1628bb7>
Thu, 23 Nov 2006 21:16:46 +0000 (21:16 +0000)
committerguy <guy@f5534014-38df-0310-8fa8-9805f1628bb7>
Thu, 23 Nov 2006 21:16:46 +0000 (21:16 +0000)
commit6184103cd8aa0aa6eb79b4ceb62e0d0f7d3831a2
treeb8a4fc55ca8524b944595a7ef65d16849af5ef9a
parent4c4e03685231bd66a2fb3c29d165625b65b7e233
Rename proto_construct_dfilter_string() to
proto_construct_match_selected_string() to indicate what it does - and
have it return a Boolean indication of whether the string could be
built, returning the string through a pointer, and, if that pointer is
null, have it just return the Boolean and not construct the string.

Get rid of proto_can_match_selected() -
proto_construct_match_selected_string() can be used for that, which
means we have only one piece of code that knows whether a "match
selected" string can be constructed or not.

Have proto_construct_match_selected_string() support matching
zero-length FT_NONE (and FT_PCRE, but that shouldn't happen) fields even
if there's no epan_dissect_t, as such a match just checks whether the
field is present.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@19967 f5534014-38df-0310-8fa8-9805f1628bb7
epan/proto.c
epan/proto.h
gtk/expert_comp_table.c
gtk/main.c
gtk/menu.c
print.c
tap-protocolinfo.c