Re: [Wireshark-dev] Eclipse project for Wireshark
On 20/01/2011 21:26, Dietfrid Mali wrote: I could as well analyze the automake hell Wireshark comes with. ^_^ Btw, I would have happily used the Wireshark VS solution to see how to add spandsp codecs, but to my horror I found out that it used some nmake hell just as well ... and required bash (Cygwin/mingw?) which I don't have installed on my Windows machine (why should I). If you take the time to setup a Windows build environment (basically Cygwin + Python + VS), then you can use the Visual Studio debugger. Note that Wireshark is still built using nmake rather than a VS project but VS can attach to the process to debug it. The reason that the windows build still uses other tools and nmake is that no-one has found the time and motivation to create, and most importantly maintain, a VS solution for Wireshark. -- Regards, Graham Bloice ___ Sent via:Wireshark-dev mailing list wireshark-dev@wireshark.org Archives:http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe
Re: [Wireshark-dev] Eclipse project for Wireshark
On Jan 21, 2011, at 12:46 AM, Graham Bloice wrote: The reason that the windows build still uses other tools and nmake is that no-one has found the time and motivation to create, and most importantly maintain, a VS solution for Wireshark. Note that maintain includes allow developers using UN*X not to have to somehow manually update a VS project file when they add a new source file to Wireshark. At least with VS nmake we can have both the auto* Makefile.am files and the VS nmake files both include Makefile.common, so adding a new source file can, in 99% of the cases, be done by a UN*X developer using the auto* files or a Windows developer using the VS nmake files and it'll work for the other group. CMake is multi-platform and even *that* isn't always updated when people add files to the project. If we can arrange that CMake handle *all* the weird platform checks that the auto* stuff does, or are willing to abandon the platform versions/variants it can't handle, and are willing to require that all developers install CMake (and we either have binary packages available for all platforms that matter, or can have CMake generate release source tarballs that don't require CMake, or are willing to require that *end users* install CMake if they can't install from a binary package), we might be able to switch to CMake, in which case it might be able to generate VC project files (at least as I understand what the CMake site says). CMake *might* work better with Eclipse than auto*: http://www.cmake.org/Wiki/CMake:Eclipse_UNIX_Tutorial (It might say UNIX in the URL, but the page says These instructions are written with OS X /Linux in mind but should be applicable to Windows Operating Systems.) ___ Sent via:Wireshark-dev mailing list wireshark-dev@wireshark.org Archives:http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe
Re: [Wireshark-dev] snmp decoding ...ubuntu smi issue ?... different then on windows XP ...?
Apologies in advance for the tone, but some actions are inexcusable. Jaap Keuter wrote: Hi, Some history: libsmi doesn't provide MIBs any more, due to Debian bug 498476 [7] The bug belongs to Debian, not libsmi. limsmi provides MIBs. Debian chooses to not include them. MIB installation was spun out into an independent package snmp-mibs-downloader [8] This package downloads the MIBs directly from their source [9], so these are not cleaned up for libsmi. Debian is choosing to use the MIB writers version - not the one in libsmi's source. They would rather have an unusable package than a working one. The better solution would have been for Debian to package the whole of libsmi as non-free, or for Wireshark to provide a Debian safe version without SNMP support. IANA and IETF provided MIBs have numerous syntax errors. The libsmi authors have gone to a lot of trouble to correct the MIBs they provide in their codebase. I can not think of a single manufacturer whose MIBs will all go into libsmi without some syntactic and semantic corrections. NetSNMP does not have the same level of strictness in its parser, so it was not a problem until the switch was made to libsmi. FYI OpenNMS runs into the same problems with faulty MIBs. Andrew -- There's no point in being grown up if you can't be childish sometimes. -- Dr. Who ___ Sent via:Wireshark-dev mailing list wireshark-dev@wireshark.org Archives:http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe
Re: [Wireshark-dev] Eclipse project for Wireshark
Hi Jaap, thanks for the info. I have given up on trying to build Wireshark with Eclipse. It's too much of a mess. How do I configure the regular Linux make for debugging? With --enable_debug=yes, or how else please? Date: Thu, 20 Jan 2011 23:01:00 +0100 From: jaap.keu...@xs4all.nl To: wireshark-dev@wireshark.org Subject: Re: [Wireshark-dev] Eclipse project for Wireshark Hi, Your a bit on your own here, but lets get an bottom up approach going You'll need files from: /-codecs/all subfolders -epan/all subfolders -gtk -plugins/all subfolders -win32 -wiretap -wsutil Be aware of the fact that stuff is compiled into libraries, which can be put together in the various working parts of the suite: Wireshark and the command line tools. I wonder if you could just build using the autotools (gasp) and then launched your debugger on the result. Should figure out where to look just fine. Adding the ASN.1 directory you are interested in to the GDB source path makes that work too. Thanks, Jaap On 01/20/2011 10:26 PM, Dietfrid Mali wrote: Date: Thu, 20 Jan 2011 12:32:44 -0700 From: st...@stephen-fisher.com To: wireshark-dev@wireshark.org Subject: Re: [Wireshark-dev] Eclipse project for Wireshark On Thu, Jan 20, 2011 at 07:45:23PM +0100, Dietfrid Mali wrote: I have been trying to build Wireshark using Eclipse, but I am running into missing header and C files all the way (particularly with many of the decoders in the ASN1 source tree). The reason is that I want to add a few new codecs using libspandsp, have failed in properly adding them so far The asn1 source tree contains template files and specification files that are merged together and use #line and #file pre-processor directives to show where everything came from. There shouldn't be a problem finding any headers. When throwing a bunch of C source files at Eclipse, it simply assumes it should build all of them and create an executable from them. While I could easily tell it not to build the tools the source of which is contained in the source code I have checked out from the Wireshark SVN repo, I cannot tell which other files are auxiliary, so Eclipse is indeed throwing a lot of file not found errors both for C source and header files, particularly when building stuff from the ASN1 subfolder. So? would like to be able to conveniently debug Wireshark using Eclipse's built-in graphical gdb frontend. You could always try emacs :) I could also sow a button to my cheek ... or use DDD (Eclipse's gdb GUI is much better though). Avoiding good tools when developing software is a kind of masochism I don't share. ;) Is there an Eclipse project available for building Wireshark on Linux? Not that I'm aware of. If no, is there general information available how to build Wireshark w/o using the automake stuff it comes with by default? We have cmake build files. I could as well analyze the automake hell Wireshark comes with. ^_^ Btw, I would have happily used the Wireshark VS solution to see how to add spandsp codecs, but to my horror I found out that it used some nmake hell just as well ... and required bash (Cygwin/mingw?) which I don't have installed on my Windows machine (why should I). ___ Sent via:Wireshark-dev mailing list wireshark-dev@wireshark.org Archives:http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe ___ Sent via:Wireshark-dev mailing list wireshark-dev@wireshark.org Archives:http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe
[Wireshark-dev] Header field (hf_) in packet-ieee80211.c
Hi, I work on the file packet-ieee80211.c and I found that some header field ( tag_interpretation, cf_version ...) will not begin by hf_ There is a historic reason ? Because if I'm not mistaken, The checkhf.pl file only verified header fields starting with hf_ ? Regards, ___ Sent via:Wireshark-dev mailing list wireshark-dev@wireshark.org Archives:http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe
Re: [Wireshark-dev] Header field (hf_) in packet-ieee80211.c
Alexis La Goutte skrev 2011-01-21 17:32: Hi, I work on the file packet-ieee80211.c and I found that some header field (tag_interpretation, cf_version ...) will not begin by hf_ There is a historic reason ? Yes, the historic reason being that it slipped trough ;-) There is a lot of stuff that should be fixed like misuse of tvb_get_ptr hf naming etc. Any effort to clean it up would be appreciated. I think the hf variables should be prefixed with hf_ieee80211 to follow the style of other dissectors, unfortunately that's a bit long, hf_80211, hf_wifi? if hf_ieee80211 is used throughout it's ease to shorten it later. If you decide to clean it up a set of smaller patches is probably better than a huge one. Best regards Anders Because if I'm not mistaken, The checkhf.pl http://checkhf.pl file only verifiedheader fields starting with hf_? Regards, ___ Sent via:Wireshark-dev mailing listwireshark-dev@wireshark.org Archives:http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe ___ Sent via:Wireshark-dev mailing list wireshark-dev@wireshark.org Archives:http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe
Re: [Wireshark-dev] Eclipse project for Wireshark
On Jan 21, 2011, at 3:57 AM, Dietfrid Mali wrote: How do I configure the regular Linux make for debugging? With --enable_debug=yes, or how else please? For debugging in what sense? If you're building from SVN, I think it compiles with -g by default (it certainly does so on OS X, and I think it did so when I was debugging a problem that only showed up in a Turkish locale, which I could get for my Ubuntu virtual machine but couldn't as conveniently set up on my Mac). ___ Sent via:Wireshark-dev mailing list wireshark-dev@wireshark.org Archives:http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe
Re: [Wireshark-dev] ANSI C12.22
Dubrawsky, Ido Ido.Dubrawsky@... writes: I’ve noticed that an ANSI C12.22 dissector is now a part of the 1.5.0 development branch in Wireshark (it appears to have been added in around the time of SVC Rev. 35292). I’ve downloaded the 1.5.0 SVN Rev 35597 build for Win32 executable but cannot seem to find the ANSI C12.22 protocol in the protocol list. Can someone point me in the right direction? A dissector for ANSI C12.22 has been submitted in bug 5531 (See https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=5531), but it looks like nobody has had the time to review it and commit it yet. ___ Sent via:Wireshark-dev mailing list wireshark-dev@wireshark.org Archives:http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe
Re: [Wireshark-dev] Header field (hf_) in packet-ieee80211.c
On Fri, Jan 21, 2011 at 06:55:51PM +0100, Anders Broman wrote: I think the hf variables should be prefixed with hf_ieee80211 to follow the style of other dissectors, unfortunately that's a bit long, hf_80211, hf_wifi? if hf_ieee80211 is used throughout it's ease to shorten it later. How about just 80211? I noticed yesterday that the protocol column is quite wide when looking at raw 802.11 traffic (such as that from AirPcap w/o decryption), it shows IEEE 802.11 when 802.11 would probably suffice. ___ Sent via:Wireshark-dev mailing list wireshark-dev@wireshark.org Archives:http://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-requ...@wireshark.org?subject=unsubscribe