suds-community

• Fix undeclared variables, found via linting.
• Allow subclassing Builder to always initialize optional arrays with empty lists
• [ci] Drop py3.6 testing, add 3.10
• Allow subclasses of Builder to determine if children are initialized.
• Don't ignore underscore prefixed attribute

version 1.0.0, 1.0.0-beta.1, 1.0.0-beta.2 (2021-09-13)

• Drop support for python 2.x

• R4201 Any PASSWORD MUST specify a Type attribute
• Compress or decompress message when Content-Encoding is present in headers
• Do not force byte-compilation optimization at level 1

• Add per invocation timeout
• Ensure request headers set in options are used to fetch definitions

• Fix pypi description format

• Fix bug introduced in 0.8.0, initialize optional arrays with empty lists (@guifran001)

• [Breaking] Objects should not be instantiated with empty optional attributs, see https://github.com/suds-community/suds/issues/14 for more information. (@guifran001)
• Add WSSE password digest and nonce encoding type (@ovnicraft)

• Pass header while requesting a WSDL (@guifran001)

• Fix Travis deploy (@btb)

• Documentation updates (@phillbaker)
• Add travis deploy (@phillbaker)

• Based on revision 712 (1e48fd79a1fc323006826439e469ba7b3d2b5a68) from the original suds Python library development project's Subversion repository.
  • Last officially packaged & released suds Python library version - 0.4.1.
• Supported Python versions.
  • Intended to work with Python 2.4+, except for the following versions:
    • Python 3.0.x - not supported by setuptools, pip or pytest.
  • Basic sources prepared for Python 2.x.
  • For using Python 3 the sources first processed by the Python py2to3 tool during the setup procedure.
  • Tested in the environments listed in the [.travis.yml]{.title-ref}
• Improved support for decimal XSD types.
  • Now modeled internally using Python's decimal.Decimal type instead of float - see the new suds.xsd.sxbuiltin.XDecimal class.
  • Based on a patch included with #454 for the original suds library implementation.
  • In order to get a decimal value formatted correctly in constructed SOAP request XML documents, pass it to suds as decimal.Decimal or an int/long.
    • In general, passing a value of a Python type other than decimal.Decimal causes that type's native string representation to be used which might not strictly match the lexical representation rules defined in the XSD specification for the decimal XSD type. For instance, a float value may be represented using scientific notation, or a fractions.Fraction may be represented using its numerator & denominator values.
    • Specific user applications can easily register their own customized XDecimal implementation using suds.xsd.sxbuiltin.Factory.maptag() if they want to use more specialized decimal value handling.
• Updated how suds constructs its cached WSDL & XML identifiers to allow cached data reuse between different processes with Python's hash randomization feature enabled.
  • Previously constructed using the built-in Python hash() function, while now it gets constructed using md5 hash.
  • Python's hash randomization (implemented since Python 2.6.8, enabled by default since Python 3.3) was causing different processes to mangle their cached data names differently.
  • Many thanks to Eugene Yakubovich for reporting the issue as well as providing the initial fix.
• Fixed loading recursive WSDL imports.
• Fixed loading recursive XSD imports/includes.
• Fixed an infinite recursion bug encountered when looking for an XSD type in an XSD schema that does not contain it but itself defines a type referencing a recursively defined type from an imported XSD schema.
  • Kudos to Kevin Pors ([krpors]{.title-ref} on BitBucket) for detecting, analysing & reporting the issue, plus preparing a working quick-fix.
• Removed never actually used suds.mx.appender.DictAppender class.
  • All code paths that could potentially lead to this class getting used convert any encountered dictionaries to suds.sudsobject.Object instances and report an error in case a corresponding XSD type can not be found.
• Now marshalling passed empty object optional params as empty SOAP request XML elements.
  • Before, passing an empty suds object as an optional parameter value was treated the same as not passing that parameter's value or passing it None - the value would not get marshalled into the constructed SOAP request at all.
  • Now, user can still not have the value marshalled by passing nothing or None, but passing an empty object will get marshalled as an actual SOAP request XML element.
  • Kudos to Nicholas Chen (nicholaschen at BitBucket) & Mark Saniscalchi (msaniscalchi at BitBucket) for reporting the issue and preparing the initial fix.
• Made suds no longer eat up, log & ignore exceptions raised from registered user plugins (detected & reported by Ezequiel Ruiz & Bouke Haarsma, patch & test case contributed by Bouke Haarsma).
• Fixed places in code where suds could eat up & silently ignore internal Python exceptions like KeyboardInterrupt or SystemExit.
• Fixed the exception message used when attempting to construct a suds.sax.element.Element with a non-Element parent.
• suds.xsd.sxbase.SchemaObject.content() now runs in linear instead of quadratic time.
• DepList class replaced with a simple dependency_sort() function taking a single dependency dictionary as input.
  • The original implementation's interface was too heavy-weight with no added value.
  • Anything tried with the original interface outside the basic use-case covered by dependency_sort() was actually or could be easily broken.
  • suds.xsd.deplist module renamed to suds.xsd.depsort.
• Global XSD elements (i.e. top-level + reference elements) are now correctly always considered qualified and their form attribute values are ignored.
  • Many thanks to Andrew Yager from BitBucket for reporting the issue.
• suds.cache module cleanup.
  • Fixed FileCache default cache location related security issue.
    • Each process now uses a separate temporary folder as its default cache location.
    • Different FileCache instances within the same process still use the same default cache location and user may still explicitly specify a non-default location for each FileCache instance.
    • Default cache location now gets removed automatically on process exit. User code may disable this removal by setting the FileCache.remove_default_location_on_exit class attribute to False.
    • Additional external information on this issue:
      • Red Hat bug 978696
      • CVE-2013-2217
      • Ubuntu USN-2008-1: Suds vulnerability
      • http://lists.opensuse.org/opensuse-updates/2013-07/msg00062.html
      • https://bugzilla.novell.com/show_bug.cgi?id=827568
      • http://www.openwall.com/lists/oss-security/2013/06/27/8
    • Many thanks to Rolf Krahl for the initial report, providing links to related external resources as well as helping brainstorm the whole issue.
  • Fixed a bug causing DocumentCache to never actually cache any documents since one of the last commits made to the original suds project.
    • That commit refactored suds.sax.document.Document so it is no longer derived from suds.sax.element.Element while the suds.cache.DocumentCache.put() implementation simply did nothing when passed something other than a suds.sax.element.Element instance. suds.reader.DocumentReader on the other hand always passes suds.sax.document.Document instances to its cache's put() method.
    • Many thanks to bgr_ at BitBucket for reporting the issue.
  • Fixed a bug causing DocumentCache & ObjectCache to not remove their cached files when failing to read data from them or process the data read from them.
  • FileCache version file operations now take care to close the file in case of a failed read/write operation.
  • Removed FileCache.setlocation() method as it was never used inside suds and if used from user code would have caused the cache to use a specific folder but without making sure that the data already stored in it has been prepared for the correct suds version, as done when passing a location parameter to the FileCache constructor.
  • Private and protected FileCache interface methods renamed to use leading underscores.
    • FileCache.getf() --> FileCache._getf().
    • FileCache.__fn() --> FileCache.__filename().
    • FileCache.checkversion() --> FileCache.__check_version().
    • FileCache.mktmp() --> FileCache.__mktmp().
    • FileCache.open() --> FileCache.__open().
    • FileCache.setduration() --> FileCache.__set_duration().
    • FileCache.validate() --> FileCache.__remove_if_expired().
  • Updated FileCache duration implementation.
    • FileCache construction now takes standard datetime.timedelta duration related keyword arguments instead of just weeks, days, hours, minutes & seconds. More to the point, it now also supports milliseconds & microseconds keyword arguments.
    • Corrected FileCache docstring stating that it accepted a months keyword argument. Using that argument would actually have caused a failure when passing it to a datetime.timedelta initializer internally.
    • You may now specify multiple duration keyword arguments in FileCache construction and they will all get summed up when constructing the internal datetime.timedelta duration representation. Before, you could specify such multiple arguments, but that would only make the FileCache silently use duration 0, i.e. its cache entries would never expire.
• Fixed suds.sax.document.Document str conversion broken around the end of 2011 by some accidental interaction between our Python 3 compatibility fixes and one of the final official suds project commits making suds.sax.document.Document no longer be derived from suds.sax.element.Element.
  • Many thanks to Ezequiel Ruiz (emruiz81 at BitBucket) for detecting & reporting the issue, as well as providing the initial patch.
• Cleaned up suds.transport ASCII/unicode URL/data handling.
  • suds.transport.Request now allows specifying its URL input as either a byte or a unicode string with any Python version. Internally that URL information is always converted to the used Python interpreter's native str data type (byte string for Python versions prior to 3.0, or unicode string for later ones).
  • Given URLs must not contain any non-ASCII characters, and any attempt to create a suds.transport.Request with such an invalid URL is reported as a UnicodeError (either UnicodeDecodeError or UnicodeEncodeError depending on the exact Python version and the given URL data type used).
  • suds.transport.Reply & suds.transport.Request string representation cleaned up and no longer raises an error when their message data contains non-ASCII characters.
• suds.client module cleanup.
  • Removed unused suds.client.Client.messages attribute.
  • Renamed private SoapClient & SimClient classes:
    • SoapClient --> _SoapClient.
    • SimClient --> _SimClient.
  • Several private methods renamed:
    • _SoapClient.location() --> _SoapClient.__location().
    • _SoapClient.get_fault() --> _SoapClient.__get_fault().
    • _SoapClient.headers() --> _SoapClient.__headers().
  • RequestContext no longer has client & original_envelope attributes.
    • client attribute seems unnecessary.
    • original_envelope was an incorrectly documented bug trap - it represented the XML request envelope as a SAX XML document from after being processed by registered marshalled plugins, but before being processed by registered sending plugins. Users should use the envelope attribute instead which can easily be converted into a SAX XML document if needed by parsing it using suds.sax.parser.Parser.parse(). That envelope has been consistently processed by all relevant registered plugins and matches the data to be sent over the registered transport exactly.
  • Cleaned up _SoapClient debug log messages a bit.
• suds.reader module cleanup.
  • Several private methods renamed:
    • DocumentReader.cache() --> DocumentReader.__cache()
    • DocumentReader.download() --> DocumentReader.__fetch()
    • DefinitionsReader.cache() --> DefinitionsReader.__cache()
• Updated the BuildError exception message.
  • Reformatted.
  • Converted to a unicode string.
• Marked suds.mx.core.Core.node() as abstract since this base class variant is never actually used (both Encoded & Literal derived classes use a different implmentation).
• suds.binding.Binding converted to a new-style class.
• suds.tostr() utility function may no longer silently eat internal Python exceptions like KeyboardInterrupt or SystemExit.
• Removed the unused SoapHeadersNotPermitted exception class.
• Extra input arguments now reported when invoking web service operations taking no input parameters.
• Using injected requests/replies/error-information with a web service operation taking at least one input parameter no longer causes suds to report an invalid extra argument error.
• Internal project development improvements.
  • The project will from now on be distributed as a wheel as well as a source distribution.
  • Added a script for automatically setting up required development Python environments for this project, hopefully supporting the full range of supported Python versions out of the box.
  • Improved internal project HACKING.rst documentation.
• setup.py improvements.
  • Python 3.0.x releases explicitly marked as not supported.
  • Attempting to run setup.py in an unsupported Python environment now reports a clean error message.
  • Now uses setuptools 1.4.2 with Python 2.4 & 2.5, and setuptools 5.1 with all more recent Python releases.
  • Project may now be installed without even in environments when you can not install setuptools.
    • In such cases setup.py will attempt to use any preinstalled setuptools version, and if none is available, it will disable some of its features and fall back to using a plain distutils based setup. See the setup.py script comments for a more detailed listing of all setup.py features affected by this.
  • Several installation issues fixes when installing into Python 3.x environments prior to Python 3.2.3.
  • When installing the project into a Python 3.x environment prior to Python 3.2, setuptools is not installed automatically since one of its test modules contains UTF-8 BOM characters, which would cause such automated installation to fail.
    • If needed, setuptools can still be installed into such environments by manually running its ez_setup.py installation script. Such an installation will encounter the same errors but will ignore them, effectively just leaving the installed setuptools package with one defective test module, but fully operational at run-time.
  • When installing the project into a Window Python 2.5 environment, you no longer need to manually install a compatible colorama package versions in order to be able to run the project tests.
  • Package meta-data may now contain non-ASCII characters on platforms where that is allowed, namely with all Python versions except Python 3.x prior to 3.2.2.
  • setup.py test command improvements.
    • Now works in Python 2.4.x environments.
    • Now reports cleanly if it can not be used for some reason, both when run and in the command's --help-commands listing.
    • Better commented the related implementation.
• Test suite improvements.
  • Test suite no longer installed together with the project, thus no longer causing confusion by existing in the target Python environment as a global tests package.
    • The tests may now be run from the source archive, and will always run on the suds version found installed in the used Python environment.
  • Refactored the quick & dirty batch script used to run all the project tests in multiple Python environments to remove much code duplication.
  • Automated project testing in several additional Python environment versions.
  • Added more detailed XSD modeling tests.
  • Added tests demonstrating how additional or replacement built-in XSD types can be registered with suds.
  • All project tests now using Python 2 & 3 compatible source code and so no longer need to be built separately for Python 3.
  • Added new and updated existing suds.cache module related tests.
  • Documented that all pytest test parametrizations should be prepared so they get ordered the same on all test runs. See Project implementation note #1 in HACKING.rst for more detailed information.
    • Many thanks to Bruno Oliveira (nicoddemus at BitBucket) for researching related pytest xdist usage problems, discovering & explaining the underlying issue as well as providing an initial project patch for it.