To make output more useful fields from the protocol should have
additional information like description, name, type and so on.
List of attributes added:
- ws_desc, just a simple description;
- ws allow to specify a description and a name. Useful as easier to
type instead of using two attributes;
- ws_type allows to override type detected, for instance to
specify is a boolean instead of an int;
- ws_base like ws_type but for base (hexadecimal instead of decimal)
These bases are the based defined in epan/proto.h, currently:
NONE, DEC, HEX, OCT, DEC_HEX and HEX_DEC;
- ws_txt and ws_txt_n allows to specify formatting. The difference
between formatting and description is that description is static
while these new attributes contain a formatting string as first
argument followed by arguments representing fields or the special
value "INDEX" for the index if we are dumping a structure contained
in an array.
The distinction between ws_txt and ws_txt_n is whether the item
is contained in an array or not;
- ws_inline, this is tricky attribute. It allows to embed structure
dissecting in the same function. This allows format string and other
fields (for instance switch or array sizes) to be seen by the
function currently being generated;
- ws_as, dissect this structure as another type. Useful to avoid
changing the protocol but show a slightly modified version.
This is used for instance to show two fields like x and y as a
single point. Could also be used to dump a binary data with
more detail but avoid to change marshalling/demarshalling.
This is also used to dump binary data in wireshark with more details
without changing the current marshalling/demarshalling code (eg agent
messages or QUIC image data).
These attributes required some changes in the parser as previously
arguments could be only integers and identifiers while they require
string and multiple identifiers (like "this.that").
In wireshark names are important as they can be used to do
queries about packet with specific features.
Signed-off-by: Frediano Ziglio <fziglio@xxxxxxxxxx>
---
python_modules/ptypes.py | 27 +++++++++++++++++++++++++--
python_modules/spice_parser.py | 13 +++++++++++--
2 files changed, 36 insertions(+), 4 deletions(-)
diff --git a/python_modules/ptypes.py b/python_modules/ptypes.py
index 7ab2771..311285a 100644
--- a/python_modules/ptypes.py
+++ b/python_modules/ptypes.py
@@ -108,6 +108,24 @@ valid_attributes=set([
# for a switch this indicates that on network
# it will occupy always the same size (maximum size required for all members)
'fixedsize',
+
+ # description
+ 'ws_desc',
+ # combine description + name
+ # name will be used for filtering
+ 'ws',
+ # type (BOOLEAN, STRINGZ)
+ 'ws_type',
+ # force wireshark base
+ 'ws_base',
+ # text to use for format, you can specify parameters
+ 'ws_txt',
+ 'ws_txt_n',
+ # put structure not in a separate function
+ # used to be able to retrieve fields from parent
+ 'ws_inline',
+ # handle this container as another type
+ 'ws_as',
])
attributes_with_arguments=set([
@@ -119,6 +137,13 @@ attributes_with_arguments=set([
'minor',
'bytes_count',
'virtual',
+ 'ws',
+ 'ws_desc',
+ 'ws_type',
+ 'ws_base',
+ 'ws_txt',
+ 'ws_txt_n',
+ 'ws_as',
])
def fix_attributes(attribute_list):
@@ -131,8 +156,6 @@ def fix_attributes(attribute_list):
if not name in attributes_with_arguments:
if len(lst) > 0:
raise Exception("Attribute %s specified with options" % name)
- elif len(lst) > 1:
- raise Exception("Attribute %s has more than 1 argument" % name)
attrs[name] = lst
return attrs
diff --git a/python_modules/spice_parser.py b/python_modules/spice_parser.py
index 97af8b2..a0f969a 100644
--- a/python_modules/spice_parser.py
+++ b/python_modules/spice_parser.py
@@ -3,7 +3,7 @@ import six
try:
from pyparsing import Literal, CaselessLiteral, Word, OneOrMore, ZeroOrMore, \
Forward, delimitedList, Group, Optional, Combine, alphas, nums, restOfLine, cStyleComment, \
- alphanums, ParseException, ParseResults, Keyword, StringEnd, replaceWith
+ alphanums, ParseException, ParseResults, Keyword, StringEnd, replaceWith, QuotedString
except ImportError:
six.print_("Module pyparsing not found.")
exit(1)
@@ -77,20 +77,29 @@ def SPICE_BNF():
switch_ = Keyword("switch")
default_ = Keyword("default")
case_ = Keyword("case")
+ ws_ = Keyword("@ws")
+ ws_txt_ = Keyword("@ws_txt") | Keyword("@ws_txt_n")
+ ws_desc_ = Keyword("@ws_desc")
identifier = Word( alphas, alphanums + "_" )
+ multi_identifier = Word( alphas, alphanums + "_." )
enumname = Word( alphanums + "_" )
integer = ( Combine( CaselessLiteral("0x") + Word( nums+"abcdefABCDEF" ) ) |
Word( nums+"+-", nums ) ).setName("int").setParseAction(cvtInt)
+ string = QuotedString(quoteChar='"', escChar='\\')
+
typename = identifier.copy().setParseAction(lambda toks : ptypes.TypeRef(str(toks[0])))
# This is just normal "types", i.e. not channels or messages
typeSpec = Forward()
attributeValue = integer ^ identifier
- attribute = Group(Combine ("@" + identifier) + Optional(lparen + delimitedList(attributeValue) + rparen))
+ ws = ws_ + lparen + string + comma + multi_identifier + rparen
+ ws_desc = ws_desc_ + lparen + string + rparen
+ ws_txt = ws_txt_ + lparen + string + ZeroOrMore(comma + multi_identifier) + rparen
+ attribute = Group(ws | ws_desc | ws_txt | (Combine ("@" + identifier) + Optional(lparen + delimitedList(attributeValue) + rparen)))
attributes = Group(ZeroOrMore(attribute))
arraySizeSpecImage = Group(image_size_ + lparen + integer + comma + identifier + comma + identifier + rparen)
arraySizeSpecBytes = Group(bytes_ + lparen + identifier + comma + identifier + rparen)
--
2.1.0
_______________________________________________
Spice-devel mailing list
Spice-devel@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/spice-devel