8bcc4b41d1
it seems to me that there is still something wrong:
1) the new is_qdtext_char() is incorrect.
There is a trap if char is implemented as signed char.
Suppose that c is '\xFF', it will be -1 when converted to an int.
By definition, c should be qdtext:
qdtext = <any TEXT except <">>
TEXT = <any OCTET except CTLs, but including LWS>
OCTET = <any 8-bit sequence of data>
2) the character after '\\' could be either part of a quoted-pair
(together with '\\'), or a normal qdtext, since '\\' itself can
be treated as a qdtext. This is equivalent to saying that the
character after '\\' in a quoted string could be ANY octet.
A patch based on the above two observations is attached.
Peng
(cherry picked from commit
|
||
---|---|---|
build | ||
docs/dist | ||
ixml | ||
m4 | ||
threadutil | ||
upnp | ||
.gitignore | ||
AUTHORS | ||
bootstrap | ||
ChangeLog | ||
configure.ac | ||
COPYING | ||
Doxyfile | ||
INSTALL | ||
libupnp.pc.in | ||
libupnp.spec | ||
LICENSE | ||
Makefile.am | ||
NEWS | ||
README | ||
svnignore | ||
THANKS | ||
TODO |
Portable SDK for UPnP* Devices (libupnp) Copyright (c) 2000-2003 Intel Corporation - All Rights Reserved. Copyright (c) 2005-2006 Rémi Turboult <r3mi@users.sourceforge.net> Copyright (c) 2006 Michel Pfeiffer and others <virtual_worlds@gmx.de> See LICENSE for details. This file contains information about the above product in the following sections: 1. Release Contents 2. Package Contents 3. System Requirements 4. Build Instructions 5. Install/Uninstall Instructions 6. Product Release Notes 7. New Features 8. Support and Contact Information 9. IXML support for scriptinglanguages 1) Release Contents ------------------------------------------- The Portable SDK for UPnP Devices is an SDK for development of UPnP device and control point applications. It consists of the core UPnP protocols along with a UPnP-specific eXtensible Markup Language (XML) parser supporting the Document Object Model (DOM) Level 2 API and an optional, integrated mini web server for serving UPnP related documents. 2) Package Contents ------------------------------------------- The SDK for UPnP Devices contains the following: README This file. Contains the installation and build instructions. LICENSE The licensing terms the SDK is distributed under. NEWS Changes and new features. ixml\doc The files for generating the XML parser documentation from the source code. ixml\inc The public include files required to use the XML parser. ixml\src The source code to the XML parser library. threadutil\inc The public include files required to the threading utility library. threadutil\src The source code to the threading utility library. upnp\doc The files for generating the SDK documentation from the source code. upnp\inc The public include files required to use the SDK. upnp\src The source files comprising the SDK, libupnp.so. upnp\sample A sample device and control point application, illustrating the usage of the SDK. 3) System Requirements ------------------------------------------- The SDK for UPnP Devices is designed to compile and run under several operating systems. It does, however, have dependencies on some packages that may not be installed by default. All packages that it requires are listed below. libpthread The header and library are installed as part of the glibc-devel package (or equivalent). Additionally, the documentation for the SDK can be auto-generated from the UPNP.H header file using DOC++, a documentation system for C, C++, IDL, and Java*. DOC++ generates the documentation in HTML or TeX format. Using some additional tools, the TeX output can be converted into a PDF file. To generate the documentation these tools are required: DOC++ The homepage for DOC++ is http://docpp.sourceforge.net/. The current version as of this release of the SDK is version 3.4.9. DOC++ is the only requirement for generating the HTML documentation. LaTeX/TeX To generate the PDF documentation, LaTeX and TeX tools are necessary. The tetex and tetex-latex packages provide these tools. dvips dvips converts the DVI file produced by LaTeX into a PostScript* file. The tetex-dvips package provides this tool. ps2pdf The final step to making the PDF is converting the PostStript into Portable Document Format. The ghostscript package provides this tool. For the UPnP library to function correctly, networking must be configured properly for multicasting. To do this: route add -net 239.0.0.0 netmask 255.0.0.0 eth0 where 'eth0' is the network adapter that the UPnP library will use. Without this addition, device advertisements and control point searches will not function. 4) Build Instructions ------------------------------------------- CORE LIBRARIES The in the examples below, replace $(LIBUPNP) with "libupnp-x.y.z", with x, y, and z corresponding to the version of the library that you have. All pieces of the SDK are configured and built from the $(LIBUPNP) directory. % cd $(LIBUPNP) % ./configure % make will build a version of the binaries without debug support, and with default options enabled (see below for options available at configure time). % cd $(LIBUPNP) % ./configure CFLAGS="-DSPARC_SOLARIS -mtune=<cputype> -mcpu=<cputype>" % make will build a Sparc Solaris version of the binaries without debug support and with default options enabled (see below for options available at configure time). Please note: <cputype> has to be replaced by a token that fits to your platform and CPU (e.g. "supersparc"). To build the documentation, assuming all the necessary tools are installed (see section 3) : To generate the HTML documentation: % cd $(LIBUPNP) % make html To generate the PDF file: % cd $(LIBUPNP) % make pdf A few options are available at configure time. Use "./configure --help" to display a complete list of options. Note that these options may be combined in any order. After installation, the file <upnp/upnpconfig.h> will provide a summary of the optional features that have been included in the library. % cd $(LIBUPNP) % ./configure --enable-debug % make will build a debug version with symbols support. To build the library with the optional, integrated mini web server (note that this is the default): % cd $(LIBUPNP) % ./configure --enable-webserver % make To build without: % cd $(LIBUPNP) % ./configure --disable-webserver % make The SDK also contains some additional helper APIs, declared in inc/tools/upnptools.h. If these additional tools are not required, they can be compiled out: % cd $(LIBUPNP) % ./configure --disable-tools % make By default, the tools are included in the library. To further remove code that is not required, the library can be build with or with out the control point (client) or device specific code. To remove this code: % cd $(LIBUPNP) % ./configure --disable-client % make to remove client only code or: % cd $(LIBUPNP) % ./configure --disable-device % make to remove device only code. By default, both client and device code is included in the library. The integrated web server is automatically removed when configuring with --disable-device. To build the library without large-file support (enabled by default) : % cd $(LIBUPNP) % ./configure --disable-largefile % make To remove all the targets, object files, and built documentation: % cd $(LIBUPNP) % make clean CROSS COMPILATION To cross compile the SDK, a special "configure" directive is all that is required: % cd $(LIBUPNP) % ./configure --host=arm-linux % make This will invoke the "arm-linux-gcc" cross compiler to build the library. SAMPLES The SDK contains two samples: a TV device application and a control point that talks with the TV device. They are found in the $(LIBUPNP)/upnp/sample directory. To build the samples (note: this is the default behaviour): % cd $(LIBUPNP) % ./configure --enable-samples % make will build the sample device "$(LIBUPNP)/upnp/tv_device" and sample control point "$(LIBUPNP)/upnp/tv_ctrlpt". Note : the sample device won't be built if --disable-device has been configured, and the sample control point won't be build if --disable-client has been configured. To run the sample device, you need to create a tvdevice directory and move the web directory there, giving: "$(LIBUPNP)/upnp/sample/tvdevice/web". To run the sample invoke from the command line as follows: % cd $(LIBUPNP)/upnp/sample/tvdevice % ../tv_device SOLARIS BUILD The building process for the Solaris operating system is similar to the one described above. Only the call to ./configure has to be done using an additional parameter: ./configure CFLAGS="-mcpu=<cputype> -mtune=<cputype> -DSPARC_SOLARIS" where <cputype> has to be replaced by the appropriate CPU tuning flag (e.g. "supersparc"). Afterwards make make install can be called as described above. WINDOWS BUILD In order to build libupnp under Windows the pthreads-w32 package is required. You can download a self-extracting ZIP file from the following location: ftp://sources.redhat.com/pub/pthreads-win32/pthreads-w32-2-7-0-release.exe or possibly newer versions if available. Execute the self-extracting archive and copy the Pre-build.2 folder to the top level source folder. Rename Pre-build.2 to pthreads. Open the provided workspace build\libupnp.dsw with Visual C++ 6.0 and select Build->Build libupnp.dll (F7) In the build directory there are also VC8, VC9 and VC10 folders containing solution files for Visual Studio 2005/2008/2010 respectively. If you use newer versions to build libupnp, eg Visual Studio 2003 or later, then you need to rebuild the pthreads package so it uses the same VC runtime as libupnp to prevent cross boundary runtime problems (see http://msdn.microsoft.com/en-us/library/ms235460%28v=VS.100%29.aspx). Just replace the files in the Pre-build.2 folder (renamed to pthreads as mentioned above) with the newly build versions. If you also use a newer version of pthreads-win32 then you should also replace the header files in that directory structure (obviously). For building a static library instead of a DLL and for using the static pthreads-w32 library following switches need to be defined additionally: UPNP_STATIC_LIB - for creating a statically linkable UPnP-library PTW32_STATIC_LIB - for using the static pthreads32 library 5) Install/Uninstall Instructions ------------------------------------------- Install The top-level makefile for the UPnP SDK contains rules to install the necessary components. To install the SDK, as root: make install Uninstall Likewise, the top-level makefile contains an uninstall rule, reversing the steps in the install: make uninstall 6) Product Release Notes ------------------------------------------- The SDK for UPnP Devices v1.2.1a has these known issues: - The UPnP library may not work with older versions of gcc and libstdc++, causing a segmentation fault when the library loads. It is recommended that gcc version 2.9 or later be used in building library. - The UPnP library does not work the glibc 2.1.92-14 that ships with Red Hat 7.0. For the library to function, you must updated the glibc and glibc-devel packages to 2.1.94-3 or later. There is some issue with libpthreads that has been resolved in the 2.1.94 version. 7) New Features ------------------------------------------- See NEWS file. 8) Support and Contact Information ------------------------------------------- Intel is not providing support for the SDK for UPnP Devices. Mailing lists and discussion boards can be found at http://www.libupnp.org/. If you find this SDK useful, please send an email to upnp@intel.com and let us know. * Other brands, names, and trademarks are the property of their respective owners. 9. IXML support for scriptinglanguages ------------------------------------------- The treestructure of XML documents created by IXML is hard to maintain when creating a binding for a scripting language. Even when many elements may never be used on the script side, it requires copying the entire tree structure once you start accessing elements several levels deep. Hence scriptsupport was added. To enable it compile while IXML_HAVE_SCRIPTSUPPORT has been defined (enabled by default). This allows control using only a list instead of a tree-like structure, and only nodes actually accessed need to be created instead of all the nodes in the tree. Here's how its supposed to work; * The scriptsupport allows you to add a callback when a node is freed on the C side, so appropriate action can be taken on the script side, see function ixmlSetBeforeFree(). * Instead of recreating the tree structure, an intermediate object should be created only for the nodes actually accessed. The object should be containing a pointer to the node and a 'valid flag' which is initially set to TRUE (the valid flag, can simply be the pointer to the node being NULL or not). Before creating the intermediate object, the custom tag 'ctag' can be used to check whether one was already created. * the node object gets an extra 'void* ctag' field, a custom tag to make a cross reference to the script side intermediate object. It can be set using ixmlNode_setCTag(), and read using ixmlNode_getCTag(). Whenever a new intermediate object is created, the ctag of the coirresponding node should be set to point to this intermediate object. * The tree structure traversal is done on the C side (looking up parents, children and siblings) * Every intermediate object created should be kept in a list (preferably a key-value list, where the key is the pointer to the node and the value is the pointer to the intermediate object) * when the callback is called, the node should be looked up in the list, the flag set to false, the pointer to the C-side node be cleared and on the C-side the ctag should be cleared. * whenever the intermediate object is accessed and its flag is set to False, an error should be thrown that the XML document has been closed. Freeing resources can be done in 2 ways, C side by simply calling the free node methods, or script side by the garbage collector of the scriptengine. Scriptside steps; * if the valid flag is set to False (XML document is closed), then the intermediate object can be destroyed, no further action * if the node has a parent, then the intermediate object can be destroyed after the ctag on the corresponding node has been cleared. Nothing needs to be freed on the C-side. * if the node has no parent, then the node must be freed on the C side by calling the corresponding free node methods. This will result in a chain of callbacks closing the node and all underlying nodes.