WBXML dissectors by define a new protocol handle in the WBXML dissector.
This is necessary as the Profile document (XML or WBXML) has no associated
media type (or not a useful registration).
New WBXML media: "UAPROF for Profile-Diff header (WSP)"
svn path=/trunk/; revision=11251
leftovers from before the EXT_T fix where tvb_strsize() was being called.
This patch fixes this (both in TAG and ATTR processors).
svn path=/trunk/; revision=10895
WV-CSP 1.0 but became a String from WV-CSP 1.1 onwards).
The token code page 3 of WV-CSP 1.2 is the same as in WV-CSP 1.1 (not 1.0).
svn path=/trunk/; revision=10643
Provide rendering of OPAQUE data for media types, based on a framework of 4 functions that take care of well-known and literal tags and attribute names for deciding on the meaning of the OPAQUE data. Applied to SI, EMN and WV-CSP.
Update reference URLs for WAP, OMA and related specs.
Note: as the WBXML common code only takes ~2000 lines while the token mappings take ~4000 lines, maybe it is time to envisage writing the per-media mappings to dedicated source files (packet-wbxml-wml.c, packet-wbxml-wv_csp.c etc)?
svn path=/trunk/; revision=10615
- Control the rendering of WBXML tokens to a media type token mapping.
- Control the dissection of the tokens that make up the WBXML body.
Both preferences are OFF by default.
Fix some comments as they did not reflect the current state anymore.
svn path=/trunk/; revision=10608
WV-CSP 1.2 specifications will be corrected and in approved state, we can
get rid of those comments (#ifdef ... #endif).
svn path=/trunk/; revision=10396
New simpler lists for both known registered WBXML public IDs and unknown
lists based on content type value. The latter have the possibility to add
a discriminator which facilitates a more correct/appropriate selection of
the token map in cases where more (backwards incompatible) versions exist.
Now the EXT_T tokens are correctly decoded, and a mapping function can be
assigned. This has been done in WML (existing code reused there) and in WV-CSP
(new functions looking up a value in a value_string based on the offset after
the EXT_T_0 token).
Add the textual WV-CSP media type to packet-text-media.c.
TODO - write the discriminators for WML, SyncML and WV-CSP. Add the code for
running the discriminators and possibly iterating over a list of mappings.
svn path=/trunk/; revision=10352
Track differences between WV-CSP versions in the token definitions.
Get rid of 8-bit characters in some comments.
Get rid of 'unused symbol' warnings.
svn path=/trunk/; revision=10344
Flag a known VERY NASTY bug in WV-CSP 1.2 where a token is being defined
twice with the same binary identifier. It's already a mess that they don't
provide backwards compatible encodings; you can now write a WV-CSP document
without being able to parse it because of version issues (hint: 1.2 has no
public WBXML identifier yet and the spec explicitly asks to set it to
"Unknown/Missing" so you could write an 1.1 message in 1.2 format, which
looks *really* weird as the tokens are mixed in both codings).
Consequence: I have to foresee a discriminator for the WBXML dissectors
so I *may* predict the correct version. I like reviewing in engineering :)
svn path=/trunk/; revision=10308
as the 2 most significant bits of every token byte represent whether the
tag has content and/or has an attribute list.
svn path=/trunk/; revision=10307
for custom EXT_T_x handlers as in WV-CSP you're allowed to *not* have the
index point to a string in the string table (verified in WBXML specs).
Unfortunately, WV-CSP is *never* backwards compatible. Additionally, things
get messed up when the WBXML document does *not* contain the WV-CSP DocType.
As a result, I eventually require to let the end-user choose which WV-CSP
version they want to see in such cases.
Fix code page errors with global tokens (don't depend on code pages).
Add some comment at the top level of the file.
svn path=/trunk/; revision=10280
Rename the WAP protocols so that they reflect the protocol name and not the
protocol stack name.
Fix the WSP content-type header.
Rename the WSP header fields from 'hdr' to 'header'.
Rename the WSP Openwave header fields from header code page 'x-up-1':
replace 'openwave' with 'x_up_1' (is more intuitive).
Remove unused WSP header fields.
Rename the WTLS header fields so they do not contain 'wsp' anymore.
Correct the WBXML code page switches in the token parsers.
Fix the code page support of multi-codepage content.
Update the PROV WBXML content (and find bugs in the new OMA Prov spec :).
svn path=/trunk/; revision=9291
* Add a "match_string" field to the "packet_info" structure,
saving the string value that matched in a string dissector
lookup, by analogy to "match_port" - this was required for
dissection with token rendering of WBXML content when no public
ID was given (e.g. Nokia/Ericsson OTA provisioning data).
* Add support for textual content type based WBXML token
mapping.
* Add extra WBXML public identifiers.
* Add the Nokia/Ericsson OTA provisioning (version 7) token
definitions.
* Inform the user when a content-type based token match is found.
svn path=/trunk/; revision=9061
* Fix an incorrectly #ifdef'ed return introduced by the previous
patch
* Remove the parsed_length parameter from the function signature
of the tag and attribute dissectors - have it return the
parsed length, instead
* Throw ReportedBoundsError exceptions if we run past the end of
the tvbuff, rather than putting a special entry in
* Produce consistent and extensive debug logging (enabled by
defining DEBUG_wbxml)
* Get rid of the dozens of "#ifdef DEBUG ... #endif" constructs
for debug logging
svn path=/trunk/; revision=9047
Rename WSP defines to avoid collisions with Windows defines.
Rename WSP dissector table for integer-value content types.
Add WSP dissector table for literal content types.
Add WSP dissection registration for literal content types.
Register literal and extra WBXML content types to WSP dissector.
Register textual MMS content type to WSP dissector.
svn path=/trunk/; revision=8648
characters don't show up. (We really need to handle strings better in
Ethereal, so that we can handle various character encodings in the
packets; GTK+ 2.x uses UTF-8 internally, so we could at least map
everything to that when displaying, and I think GTK+ 1.x for Win32 does
so as well, and but we need to clean up the GTK+ 1.x-for-X11 stuff to
handle the various font encodings.)
Get rid of "strtbl_lookup()" in favor of the appropriate direct calls.
svn path=/trunk/; revision=7800
* Support correct processing of code page switches
* Provide a new datatype containing scarce array of
const value_string arrays (value_valuestring)
allowing an efficient support for WBXML code pages.
* Minor fix (XML PI was always decoded without
mapping known tokens)
* Add support for decoding of more WBXML content
types (SyncML 1.0, SyncML 1.1, CHANNEL 1.0, EMN 1.0).
svn path=/trunk/; revision=7626
- Move all static definitions from packet-wbxml.h to packet-wbxml.c
- Comment out inclusion of packet-wbxml.h in packet-wbxml.c
- Append WBXML + version + public ID to the Info column
Then, while we're at it, get rid of packet-wbxml.h (we can reinstate it
if there's any functionality to export other than the dissector) and get
rid of the include of packet-wbxml.h (and update comments not to refer
to it).
svn path=/trunk/; revision=7153
add string table parsing and displaying;
add registration to more content types;
LITERAL tag indentation fix;
add and a change in unrendered WBXML: from confusing "<Tag
0x37>" type display to "<Tag_0x37>" (and same in attribute
state).
svn path=/trunk/; revision=7129
Olivier Biot