Create two new files (ws_strsplit.[ch]) that use GTK2 code to override
the buggy g_strsplit() function when compiling for GTK1. Include this
work-around function (ws_strsplit) in libwireshark.def. Add notes on usage
to README.developer. Include epan/ws_strsplit.h in all files that use
g_strsplit().
svn path=/trunk/; revision=20804
fix this, by providing required functions in the new file file_util.c - it's mostly copied from GLib (g_open alike - that take UTF8 as filename format but don't use msvcrt.dll V6 for this as the glib files do)
"link" to these functions in file_util.h: #define eth_open eth_stdio_open
revert changes (from SVN 20282) throughout the code related to these file functions which were introduced with the first tries of MSVC 2005 ...
Hopefully I've done everything right with the new file_util.c ...
svn path=/trunk/; revision=20402
- Microsoft .Net Framework SDK Version 1.1
- Microsoft .Net Framework 2.0 SDK
... which seems to work fine now (except for the open problems like zlib) :-)
Some of the plugins need different linker flags depending on the MSVC. I've set a "define" in config.nmake (PLUGIN_LDFLAGS) and used it in the affected plugins/.../Makefile.nmake. Maybe we should generally change the plugins that way.
btw: The "Microsoft Visual C++ Toolkit 2003" (mentioned some time ago on the list) doesn't work as some important files are missing - and we have much easier alternatives now :-)
svn path=/trunk/; revision=20332
Lex code is often hard for humans to understand... neither of the error conditions found would be happening...
One was marked as error because coverity does not understand that yyterminate() breaks execution.
The other marked as an error because coverity is not able to figure out that the error appears in an actions against two different rules that will not ever match toghether.
svn path=/trunk/; revision=17584
Dissectors registered with register_postdissector() will be called after all other dissectors have been called.
Use it to register mate.
svn path=/trunk/; revision=17089
include fails as we don't do -Iwiretap.
If we have it, include <sys/stat.h> in epan/filesystem.c - we need it
for stat() and the macros and structures it uses.
svn path=/trunk/; revision=16410
to do this, I've added file_util.h to wiretap (would file_compat.h be a better name?), and provide compat_macros like eth_open() instead of open(). While at it, move other file related things there, like #include <io.h>, definition of O_BINARY and alike, so it's all in one place.
deleted related things from config.h.win32
As of these massive changes, I'm almost certain that this will break the Unix build. I'll keep an eye on the buildbot so hopefully everything is working again soon.
svn path=/trunk/; revision=16403
print register numbers as unsigned (they're guint32);
when printing a PUT_FVALUE instruction, show the value as well
as the type of the value.
That requires that a bunch of types get to_repr methods; add them for
PCRE (FTREPR_DFILTER-only - show the regular expression as text),
tvbuffs (FTREPR_DFILTER_only - show the data as a hex string), integral
types, string types other than FT_STRING, and FT_IPv6.
That means we can use fvalue_to_string_repr() for FT_IPXNET and FT_IPv6
in proto_construct_dfilter_string(), and that we don't need to handle
integer and floating types specially in MATE.
Fix some problems with the PCRE execution code for tvbuff types.
svn path=/trunk/; revision=16369
that we don't have all the lexical analyzers in libethereal share them
(note that they're already static in radius_dict.l, so they weren't
sharing with any other lexical analyzer), and so that OS X 10.3.9's
run-time linker doesn't get upset at finding them defined in libethereal
and the MATE plugin.
Fix up indentation.
svn path=/trunk/; revision=15961
I've done more than a day to change the timestamp resolution from microseconds to nanoseconds. As I really don't want to loose those changes, I'm going to check in the changes I've done so far. Hopefully someone else will give me a helping hand with the things left ...
What's done: I've changed the timestamp resolution from usec to nsec in almost any place in the sources. I've changed parts of the implementation in nstime.s/.h and a lot of places elsewhere.
As I don't understand the editcap source (well, I'm maybe just too tired right now), hopefully someone else might be able to fix this soon.
Doing all those changes, we get native nanosecond timestamp resolution in Ethereal. After fixing all the remaining issues, I'll take a look how to display this in a convenient way...
As I've also changed the wiretap timestamp resolution from usec to nsec we might want to change the wiretap version number...
svn path=/trunk/; revision=15520
"make distclean", even though they come with the distribution, and
make the one in plugins/xml/Makefile.nmake match the one in some other
files.
svn path=/trunk/; revision=15214
returned quite a list of files. Add them to MAINTAINERCLEANFILES.
Whitespace changes (replace multiple spaces by TABs, in a few cases this
needed to be done at the beginning of Makefile lines.
svn path=/trunk/; revision=14891
- pdu transports were upside down (ip/tcp vs tcp/ip)
- the gop to gog index was not being populated
- tell svn to ignore generated c files
svn path=/trunk/; revision=14416
it serves the same purpose as the register routine in a built-in
dissector, and don't require all dissectors to have one, as they might
just be taps.
Get rid of the stats tree's init routine, as it's just a tap, and as it
doesn't do anything.
Update the idl2eth Python script to generate plugins with register routines.
svn path=/trunk/; revision=13644
Introduce a new init routine for plugins,
which does not take the plugin api table as an
argument and allows etheral to distinguish
between plugins using the old and the new api.
Update README.plugins accordingly
Change all g_warnings() in epan/plugins.c to report_failue().
On windows we do not have a log console open while
loading the plugins, because a log console cannot be opened before the prefs have been read. Thus g_warnings()
does not work for reporting problems with plugins.
svn path=/trunk/; revision=13596
GopTree={BasicTree|NoTree|FullTree}
NoTree: just an item for each gop
BasicTree: (default) some essential information regarding the gop
FullTree: the Gop's full tree
added Id and eol-style to examples
svn path=/trunk/; revision=13435
files. Do this with GENERATED_HEADER_FILES, GENERATED_C_FILES, and
GENERATED_FILES macros in Makefile.common files, along the lines of what
wiretap/Makefile.common has.
Clean up "*~" files with "make clean" rather than only "make distclean"
in some additional places.
Add "maintainer-clean" rules to the Makefile.nmake files, paralelling
the ones in the automake-generated Makefile.in files, using the
GENERATED_FILES macros from Makefile.common files. In some cases, move
the cleanup of files from "make distclean" to "make maintainer-clean",
and in other cases, put in a comment indicating why we're not doing that
(because some files that are distributed in the source tarballs, namely
Flex output, were built with a UN*X Flex and won't compile on Windows,
so we get rid of them with "make distclean" so you can clean up stuff
that *has* to be re-generated for Windows).
Clean up some *CLEANFILES definitions - get rid of ones that no longer
apply as files were moved or that add to the definition a name that's
already there.
svn path=/trunk/; revision=13402
- fixes a leak in analize_pdu()
- fixes a typo in analize_pdu()
- fixes a crash in scs_unsubscribe()
- save a malloc and a free in get_pdu_fields()
svn path=/trunk/; revision=13335
- fix a bug at gog reinit that disabled gogs after reload of a file.
- fix some crashes when Debug_GoG>0
- cleanup the debug output (no CR at the end is needed)
- Gops start when no GopStart is given and match an existing GogKey
svn path=/trunk/; revision=13318
The HP-UX compilers don't like non-constant initializers (C99'ism).
Patch attached for plugins/mate/mate_setup.c.
Fix a few warnings (change some () to (void))
svn path=/trunk/; revision=13274
- renamed analyze_frame() into mate_analyze_frame().
- prepare mate to be statically linked
- use report_open_failure() , report_read_failure() and
report_failure() in case there's an error while loading the config
- some cleanup in mate_util.c
- some fixes and cleanup in examples and matelib
svn path=/trunk/; revision=13001
A new patch for mate
- changed the default config file name to ""
- avoid trying to confgure mate when the config filename is ""
- add filename and line to config errors, that way one can actually
find errors in config files
- teady up the config error strings ( made them uniform )
- fixed a crash when missing index AVPs (Name,For,On etc) in config AVPLS
- make sure we do not delete the config AVPs and AVPLs to avoid
references to unsubscribed strings been left arround
- fixed the len on LoALs
- changed the way loal_from_file reports an error.
- matelib/sip.mate was missing the Name in a PduDef (from Julien Leproust)
- matelib/h225_ras had a wrong field name (from Julien Leproust)
At this point mate is silent when it's not expressely configured by the user.
From me:
reduce compiler noise on MSVC
svn path=/trunk/; revision=12985
Another patch for mate:
- adds a "GopTree={TRUE|FALSE}" attribute to Gogs
- changes "ShowPduTree={NoTree|PduTree|FrameTree}"
- adds "implicit stop" to Gops with no Stop declared
svn path=/trunk/; revision=12968
- moved gop and gog indexes into gopcfgs, which is a propedeutic
change for upcoming changes in the way gops are to be grouped
- changed the way gog-keys are kept in memory
- every gopkey attribute is copied into the gop->extras to avoid
redundancy in the configuration
- added timers to gogs mate.gog_type.StartTime and mate.gog_type.Time
- fixed a bug in scs_subscribe that mangled some strings
- minor interface improvement to scs propedeutic to having types avp
values in a future
- changed medium and large into mate_medium and mate_large in the
scs_collection
- fixed Mode=Replace in Transforms, now it works
- fixed a crash at reinit due to impropper initialization of mate_items
svn path=/trunk/; revision=12902
- fixes crash at reinit when DiscardPduData=TRUE: check if there's an
avpl before deleting it.
- mate.pdu_type.RelativeTime had a dual use, add
mate.pdu_type.TimeInGop and leave the other alone.
- other fixes related to times (they now seem to be right)
- scs_subscribe doesn't use g_malloc anymore. it uses GMemChunks
(I had to rename two GMemChunks to make MSVC happy)
- fixes crash at reinit when DiscardPduData=TRUE: check if there's an
avpl before deleting it.
- mate.pdu_type.RelativeTime had a dual use, add
mate.pdu_type.TimeInGop and leave the other alone.
- other fixes related to times (they now seem to be right)
- fixed few things at reinit:
- reset last_id
- remove the nodes in the items hash tables
svn path=/trunk/; revision=12820
From Luis Ontanon:
- moves mate configuration from proto_register to proto_register_handoff
- add the config file protocol preference
- every item (gop,gog,pdu) has it's own ett
- the tap doesn't do nothing, it just primes the tree
- analyze_frame() what once was the tap now is called by the dissector
- should work with tethereal now (to be tested)
svn path=/trunk/; revision=12763
- make rd static (no purpose for it to be global)
- remove outdated comments
- add "PduCriteria" feature (as already stated in the wiki)
- add "DiscardUnassignedPdu" feature (as already stated in the wiki)
- removed ".Id" from abbrev for "mate.item_name"
- in radius.mate: Don't use the port to determine the Gop key.
svn path=/trunk/; revision=12735
As suggested by Martin Regner:
- Use strtod() instead of strtof()
From Luis Ontanon:
- changes the id of mate items to be integer (the old string ID
imposed a lenght limit and as mate fileds had become dynamic it makes
no more sense anymore)
- fixes a huge avp leak on reinit (every avp object was leaked when a
new file was loaded)
- adds the "Lib" AVP to the Action=Include AVPL to include definitions
>from matelib
- rename mate.[dll/so] to zzmate.[dll/so] so it gets initialized as
the very last protocol (so that fields from every dissector can be
used).
svn path=/trunk/; revision=12728
MATE -- Meta Analysis and Tracing Engine
Won't be compiled by default.
It is still not possible to link the plugin on Win32.
svn path=/trunk/; revision=12716
Ulf Lamping