entrouvert
/
debian-suds-jurko
17
0
Fork 0
Code Issues Pull Requests Releases Activity
debian-suds-jurko/TODO.txt

2375 lines
115 KiB
Plaintext
Raw Permalink Blame History

PRIORITIZED:
=================================================
(21.12.2011.)
(+) * (Jurko) Prepare for the initial forked project release.
(+) * Add todo list.
(+) * Document how to access this forked project's development sources &
(+) released files.
(+) * Jurko's Mercurial repository hosted at BitBucket and accessible
(+) from 'http://bitbucket.org/jurko/suds'.
(+) * Already documented in README.
(+) * Add more details to HACKING.
(+) * Suds library Python 3 patches hosted in a Mercurial patch queue
(+) repository at BitBucket and accessible from
(+) 'http://bitbucket.org/bernh/suds-python-3-patches'.
(+) * Already documented in HACKING.
(+) * (Jurko) Minor stylistic changes & typo corrections.
(+) * Code.
(+) * 'tranparent' --> 'transparent'.
(+) * 'if tns' --> 'if tns'.
(+) * 'docuemnt' --> 'document'.
(+) * '('restriction', 'any', 'list',)' --> '('restriction', 'any',
(+) 'list')'.
(+) * And other unnecessary trailing tuple commas.
(+) * 'Qualfied' --> 'Qualified'.
(+) * 'Resolveds' --> 'Resolves'.
(+) * 'describe a port and it's list of methods' --> 'describe a port
(+) and its list of methods'.
(+) * 'dependancies' --> 'dependencies'.
(+) * 'imcoming' --> 'incoming'.
(+) * 'relavent' --> 'relevant'.
(+) * 'inidcat' --> 'indicat'.
(22.12.2011.)
(+) * (Jurko) Prepare for the initial forked project release.
(+) * Rename top level project documentation files to use the .txt extension
(+) to make them friendlier to Windows users.
(+) * Research release procedure.
(+) * Open PyPI account.
(+) * How to prepare a source distribution package.
(+) * Change author information.
(+) * Include tests.
(+) * Include all the top-level documentation files.
(+) * 'README'.
(+) * 'LICENSE'.
(+) * 'HACKING'.
(+) * 'TODO'.
(+) * Note the original project author in the package description.
(+) * Include correct license information.
(+) * See what the difference between author and maintainer
(+) information is and where it can be seen.
(+) * Try using 'setuptools_hg' to simplify specifying the project
(+) sources.
(+) * Failed when used under Python 3.
(+) * How to upload the prepared distribution packages.
(+) * Should upload a source distribution only.
(23.12.2011.)
(+) * (Jurko) Prepare for the initial forked project release.
(+) * Research release procedure.
(+) * How to upload the prepared distribution packages.
(+) * PyPI.
(24.12.2011.)
(+) * (Jurko) Prepare for the initial forked project release.
(+) * Research release procedure.
(+) * How to upload the prepared distribution packages.
(+) * BitBucket.
(+) * Document the project's official download URL.
(+) * Document how to access this forked project's development sources &
(+) released files.
(+) * Released project packages accessible from PyPI & BitBucket.
(+) * Installing the project using distribute or pip.
(+) * Document release procedure.
(+) * Version identification.
(+) * Remove the '(development)' suffix for official release builds.
(+) * Format '<base-suds-version> jurko #', e.g. '0.4.1 jurko 1'.
(+) * Tag in Hg.
(+) * Name the tag like 'release-<version-info>', e.g.
(+) 'release-0.4.1 jurko 1'.
(+) * Prepare official releases based only on tagged commits.
(+) * Prepare source distribution package, register the new release
(+) at PyPI and upload the prepared source package.
(+) * Run 'setup.py sdist register upload'.
(+) * Upload the prepared source package to the project site.
(+) * Archive the prepared source release locally if needed.
(+) * Next development version identification.
(+) * Bump up the forked project version counter.
(+) * Add back the '(development)' suffix.
(+) * Commit all local changes.
(+) * (Jurko) Constructing a SOAP request containing data stored in a sequence
(+) inside a choice.
(+) * Test scenario (syntax not precise).
(+) <choice>
(+) <element "a" - string/>
(+) <element "s">
(+) <sequence>
(+) <element "s1" - string/>
(+) <element "s2" - string/>
(+) </sequence>
(+) </element>
(+) </choice>
(+) * When 's' is None and 'a' is not - 'a' should be used.
(+) * When 'a' is None and 's' is not - 's' should be used.
(+) * When 's' is used, all of its child elements should be used independent
(+) of whether they are None or not.
(+) * Add related test.
(+) * (Jurko) Prepare the '0.4.1 jurko 1' release.
(+) * Follow the documented release procedure.
(+) * Update version information.
(+) * Tag in Hg.
(+) * Upload the source package.
(+) * Project site.
(+) * PyPI.
(+) * (Jurko) Fix getting a suds.client object's string representation when the
(+) client is initialized with the following WSDL. Calling 'str(client)'
(+) reports 'IndexError: list index out of range'.
(+) * WSDL.
(+) <?xml version='1.0' encoding='UTF-8'?>
(+) <wsdl:definitions targetNamespace="my-namespace"
(+) xmlns:wsdl="http://schemas.xmlsoap.org/wsdl/"
(+) xmlns:ns="my-namespace"
(+) xmlns:soap="http://schemas.xmlsoap.org/wsdl/soap/">
(+) <wsdl:types>
(+) <xsd:schema targetNamespace="my-namespace"
(+) elementFormDefault="qualified"
(+) attributeFormDefault="unqualified"
(+) xmlns:xsd="http://www.w3.org/2001/XMLSchema">
(+) <xsd:element name="Elemento">
(+) <xsd:complexType>
(+) <xsd:sequence>
(+) <xsd:element name="x1" type="xsd:string" />
(+) <xsd:element name="x2">
(+) <xsd:complexType>
(+) <xsd:choice>
(+) <xsd:element name="u1" type="xsd:string" />
(+) <xsd:element name="u2" type="xsd:string" />
(+) <xsd:element name="u3" type="xsd:string" />
(+) </xsd:choice>
(+) </xsd:complexType>
(+) </xsd:element>
(+) <xsd:element name="x3" type="xsd:string" />
(+) </xsd:sequence>
(+) </xsd:complexType>
(+) </xsd:element>
(+) </xsd:schema>
(+) </wsdl:types>
(+) <wsdl:message name="fRequestMessage">
(+) <wsdl:part name="parameters" element="ns:Elemento" />
(+) </wsdl:message>
(+) <wsdl:portType name="dummyPortType">
(+) <wsdl:operation name="f">
(+) <wsdl:input message="ns:fRequestMessage" />
(+) </wsdl:operation>
(+) </wsdl:portType>
(+) <wsdl:binding name="dummy" type="ns:dummyPortType">
(+) <soap:binding style="document"
(+) transport="http://schemas.xmlsoap.org/soap/http" />
(+) <wsdl:operation name="f">
(+) <soap:operation soapAction="f" style="document" />
(+) <wsdl:input><soap:body use="literal" /></wsdl:input>
(+) <wsdl:output><soap:body use="literal" /></wsdl:output>
(+) </wsdl:operation>
(+) </wsdl:binding>
(+) <wsdl:service name="dummy">
(+) <wsdl:port name="dummy" binding="ns:dummy">
(+) <soap:address location="https://localhost/dummy" />
(+) </wsdl:port>
(+) </wsdl:service>
(+) </wsdl:definitions>
(+) * Research.
(+) * Caused by undocumented suds.client.Client behaviour - it was
(+) expecting a specifically formatted suds.__build__ string which was
(+) not the case with the forked project release.
(+) * Add a test.
(+) * Fix.
(+) * Plan preparing a patched release.
(+) * (Jurko) Prepare the '0.4.1 jurko 2' release.
(+) * Update release notes.
(+) * Follow the documented release procedure.
(+) * Update version information.
(+) * Tag in Hg.
(+) * Upload the source package.
(+) * Project site.
(+) * PyPI.
(25.12.2011.)
(+) * (Jurko) Printing out a list of function parameters should not print an
(+) additional trailing comma after the last parameter.
(+) * Research.
(+) * suds.servicedefinition.ServiceDescription.description() code needs
(+) to be changed.
(+) * Prepare test.
(+) * Update code.
(+) * (Jurko) 'suds.xsd.xsbasic.Enumeration' objects should list their value in
(+) their string representation.
(+) * Research.
(+) * Prepare test.
(+) * Update code.
(+) * (Jurko) 'suds.sudsobject.Metadata' __unicode__()/__str__()/__repr__()
(+) functions should not raise an AttributeError.
(+) * Research.
(+) * There should be no need to access a 'suds.sudsobject.Metadata'
(+) object's __metadata__ member as done for 'suds.sudsobjects.Facade'
(+) class instances.
(+) * Prepare test.
(+) * Update code.
(26.12.2011.)
(+) * (Jurko) Clean up suds.xsd.sxbasic.TypedContent.resolve().
(+) * Research.
(+) * Base class resolve() behaviour.
(+) * Other resolve() functions in suds.
(+) * 'resolve()' related caching.
(+) * Clean up the SchemaObject resolve() implementations.
(+) * Caching not needed in TypedContent base classes.
(+) * Document.
(+) * Returns the same XSD node when the node does not have an
(+) explicitly specified external type.
(+) * When called for an XSD node with an explicitly specified external
(+) type returns that type's XSD node.
(+) * (Jurko) Clean up suds.xsd.sxbasic.TypedContent.resolve().
(+) * Research WSDL structure related to the resolving type references.
(+) * 'type'.
(+) * 'ref'.
(+) * Prepare additional resolve() tests.
(+) * 'ref'.
(+) * Valid.
(+) * Recursive.
(+) * Invalid.
(+) * References to nodes referencing other nodes.
(+) * There seems to be no way to do this in WSDL so seems no
(+) reason to keep the complicated and potentially buggy
(+) recursive resolve() implementation.
(+) * Refactor the resolve() implementation to remove recursion.
(+) * Todo items obsoleted by this refactoring.
(+) * Prevent possible endless resolve() loops due to resolve()
(+) directly or indirectly returning the same TypedContent
(+) instance.
(+) * Refactor to cache the final resolved type instead of a possibly only
(+) partially resolved one when resolving without allowing resolving to
(+) builtin types.
(+) * Research.
(+) * Prepare test.
(+) * Update code.
(+) * (Jurko) Check and remove detected potential unused imports if they are no
(+) longer needed.
(+) * splitPrefix.
(+) * DefinitionsReader.
(+) * (Jurko) Prepare the '0.4.1 jurko 3' release.
(+) * Update release notes.
(+) * Follow the documented release procedure.
(+) * Update version information.
(+) * Tag in Hg.
(+) * Upload the source package.
(+) * Project site.
(+) * PyPI.
(+) * (Jurko) Look into suds.xsd.sxbase.SchemaObject.unbounded(). It seems to
(+) return True even when the object is bounded with a max value greater than
(+) 1.
(+) * Research.
(+) * Add tests.
(+) * 'min'.
(+) * 'max'.
(+) * 'optional'.
(+) * 'required'.
(+) * 'unbounded'.
(+) * Update code - rename unbounded to multi_occurrence.
(27.12.2011.)
(+) * (Jurko) Get calling a web service operation taking no parameters to work
(+) correctly.
(+) * Research.
(+) * Seems to work fine. The original problem triggering this task
(+) seems to have been caused by an invalid WSDL.
(+) * Add a task to add more detailed test cases for this.
(17.04.2012.)
(+) * (Jurko) Merge upstream changes from the original suds development
(+) repository.
(+) * (Jurko) Update embedded author values so they do not include non-ASCII
(+) characters causing problems with the 'distribute' based setup procedure
(+) which erroneously assumes they have been prepared using the user's local
(+) code-page.
(+) * (Jurko) Process received pull requests on BitBucket.
(+) * (Jurko) Prepare the '0.4.1 jurko 4' release.
(+) * Update release notes.
(+) * Follow the documented release procedure.
(+) * Update version information.
(+) * Tag in Hg.
(+) * Upload the source package.
(+) * Project site.
(+) * PyPI.
(28.02.2013.)
(+) * (Jurko) Merge changes prepared by Juraj Ivan<61>i<EFBFBD>.
(+) * Update the original Hg repository containing Python 3 related fixes.
(+) * (Jurko) Process received pull requests.
(01.03.2013.)
(+) * (Jurko) Sync with external related repositories.
(+) * 'http://bitbucket.org/palday/suds'.
(27.03.2013.)
(+) * (Jurko) Fix buggy Python 3 support patch related to reading the cache
(+) version.
(+) * The cache version file should be read as a text and not as a binary
(+) file.
(+) * (Jurko) Fix test_enumeration_type_string_should_contain_its_value test
(+) under Python 2.
(+) * (Jurko) Test & fix Python 2.4 compatibility.
(+) * (Jurko) Fix input/output binding usage.
(+) * Incorrect binding was being used in several places.
(+) * Some of the uses related to processing the SOAP Fault error reporting
(+) element seem to be 'fake', needed only because of a a bit messy
(+) design. Planned to be fixed soon.
(28.03.2013.)
(+) * (Jurko) Add web service reply processing related unit tests.
(+) * (Jurko) Remove undocumented, untested & unused binding.replyfilter
(+) functionality.
(+) * (Jurko) Remove seeming unused SoapClient last_sent() and last_received()
(+) functionality.
(+) * (Jurko) Add a test for unicode Fault data processing.
(+) * (Jurko) Merge SoapClient failed() & succeeded() functions into the same
(+) process_reply() function.
(+) * (Jurko) Make binding classes no longer have anything to do with method
(+) independent Fault element processing.
(+) * (Jurko) Make reply XML processing check the namespace used for Envelope &
(+) Body elements.
(+) * (Jurko) Make SOAP Fault processing check the namespaces used for all
(+) relevant tags.
(+) * (Jurko) Make HTTP status code 200 XML replies containing a Fault element
(+) consistently reported as SOAP faults (plus issue a warning about the
(+) non-standard HTTP status code) both when reporting such faults using
(+) exceptions or by returning a (status, reason) tuple.
(+) * Currently this is done only when reporting them using exceptions.
(+) * (Jurko) Make plugins received() & parsed() calls now process both success
(+) & error replies.
(+) * (Jurko) SOAP fault reports with invalid Fault structure should not cause
(+) suds code to break with an 'invalid attribute' exception.
(+) * (Jurko) SOAP fault reports with no <detail> tag (optional) should not
(+) cause suds code to break with an 'invalid attribute' exception when run
(+) with the suds 'faults' option set to false.
(+) * (Jurko) Clean up message reply processing return codes with suds 'faults'
(+) option set to both true & false.
(+) * (Jurko) Reorganize SimClient injection keywords.
(+) * 'msg' - request message.
(+) * 'reply' - reply message ('msg' must not be set).
(+) * 'status' - HTTP status code accompanying the 'reply' message.
(+) * 'description' - description string accompanying the 'reply' message.
(+) * (Jurko) Check failing tests.
(+) * All tests now pass except for ones related to SOAP Fault unicode
(+) faultstring processing.
(29.03.2013.)
(+) * (Jurko) Sync with external related repositories.
(+) * 'http://bitbucket.org/blarghmatey/suds-blarghmatey'.
(+) * (Jurko) Additional SOAP web service reply tests.
(+) * (Jurko) Fix detected unicode problems.
(+) * Remove invalid WebFault fix merged from an external source.
(+) * All suds exception classes now contain unicode description messages.
(+) * Undo a hasty unicode related WebFault fix merged from an external
(+) source in revision 16b084e8eea6511981d171e63cada98b58720c38.
(+) * Rename smart_str class to byte_str and make it accept only string
(+) parameters.
(+) * Clean Python2/3 compatibility DocumentStore fix.
(+) * Now contains raw data instead of unicode strings.
(+) * This also fixes a problem where unicode data read from the
(+) DocumentStore would fail to be encoded using the default encoding,
(+) which would then get reported as 'document not found'.
(+) * SAX parser now accepts only byte string content instead of also
(+) accepting unicode strings containing latin1 characters only.
(+) * Make tests now specify their fixed WSDL & reply content as byte
(+) strings only.
(+) * Make all tests pass.
(+) * Python 2.4.
(+) * Python 2.7.3.
(+) * Python 3.2.3.
(+) * (Jurko) Remove Python 2/3 unicode encoding compatibility support assuming
(+) that its encoded unicode representations contain only latin1 characters.
(+) * SoapClient 'location' cleanup.
(+) * Should be stored as a unicode object instead of being converted
(+) from its byte representation assuming it was encoded using a
(+) specific encoding, e.g. if read from a WSDL schema, it should be
(+) decoded the same as the rest of the WSDL schema.
(+) * Remove str2bytes() & bytes2str().
(+) * (Jurko) Remove the str_to_utf8_in_py2() Python 2/3 unicode encoding
(+) compatibility support function as it no longer seems to be needed.
(+) * (Jurko) Add tests for web service operation input & output element types.
(30.03.2013.)
(+) * (Jurko) Improve suds tests.
(31.03.2013.)
(+) * (Jurko) Add tests for wrapped suds operation input & output data.
(01.04.2013.)
(+) * (Jurko) Add tests for wrapped suds operation output data.
(+) * (Jurko) Merge patches sent in by Juraj Ivan<61>i<EFBFBD> from PKE sistemi.
(+) * Add tests for disabled wrapped suds operation input & output data
(+) support.
(+) * Add code for disabling suds library wrapped parameter support.
(02.04.2013.)
(+) * (Jurko) Restriction support cleanup based on patches sent in by Juraj
(+) Ivan<61>i<EFBFBD> from PKE sistemi.
(+) * Research.
(+) * Not enough time to research this thoroughly and come up with a
(+) complete and well tested solution.
(+) * Prepare and commit related tests.
(+) * Mark the tests as 'expected to fail' & comment the reasons.
(+) * Commit as a separate unfinished private branch.
(+) * (Jurko) Prepare the '0.4.1 jurko 5' release.
(+) * Update release notes.
(08.05.2013.)
(+) * (Jurko) Make suds construct SOAP requests with correct element namespaces
(+) when their XSD schema definition nodes reference other nodes with a
(+) different namespace.
(+) * Research.
(+) * Add test.
(+) * Implement.
(+) * Report back to Jens Arm from KabelDeutschland who reported the issue.
(+) * (Jurko) Support specifying a custom DocumentStore instance for a specific
(+) Client.
(+) * Support.
(+) * Update test code adding documents to the global DocumentStore instance
(+) to use a local one instead.
(+) * Cleanup.
(+) * DocumentStore.open() can return the bytes object directly instead
(+) of having to wrap it inside a BytesIO instance.
(+) * Remove unnecessary Cache functions.
(+) * getf() should be left in the FileCache class only.
(+) * putf() should be removed completely.
(+) * Add tests.
(+) * Separate DocumentStore instances must not share content.
(+) * Not specifying a DocumentStore instance uses the default global
(+) one.
(+) * Default content.
(+) * Updating content.
(+) * Accessing existing content.
(+) * Accessing missing content.
(17.06.2013.)
(+) * (Jurko) Upgrade the setup procedure to use the latest setuptools 0.7.2
(+) release instead of the now deprecated 'distribute' Python package.
(+) * Research.
(+) * Implement.
(18.06.2013.)
(+) * (Jurko) Upgrade the setup procedure to use the latest setuptools 0.7.2
(+) release instead of the now deprecated 'distribute' Python package.
(+) * Add automated setuptools installation (downloaded on-demand from
(+) PyPI).
(+) * Fix issues with installing on Python 2.4.
(+) * Add project installation troubleshooting notes to the main readme.
(+) * (Jurko) See how to allow using the setup script's 'test' option to run the
(+) project's pytest based test suite.
(19.06.2013.)
(+) * (Jurko) Resolve test failures caused by suds generating slightly different
(+) SOAP requests when using Python 3.3.
(+) * Tests should not depend on the order in which XML attributes are
(+) specified for a single XML element where this is not necessary.
(11.11.2013.)
(+) * (Jurko) Prepare the '0.4.1 jurko 5' release.
(+) * Follow the documented release procedure.
(+) * Update release notes.
(+) * Update version information.
(+) * Tag in Hg.
(+) * Upload the source package.
(+) * Project site.
(+) * PyPI.
(18.11.2013.)
(+) * (Jurko) Fix suds time-zone handling according to a pull request received
(+) on BitBucket from MDuggan1.
(+) * Research.
(+) * Suds assumes that all timezones have a full-hour offset from the
(+) UTC timezone and does not work correctly with those that do not.
(+) * This seems to be a suds specific problem and not a more
(+) general Python issue as some information on the net implies.
(+) * FixedOffsetTimezone.
(+) * datetime.tzinfo subclass.
(+) * Used only in test code.
(+) * Represents fixed offset timezones with no daylight saving
(+) time.
(+) * Start morphing the current suds & test code base towards the suggested
(+) patched code.
(19.11.2013.)
(+) * (Jurko) Fix suds time-zone handling according to a pull request received
(+) on BitBucket from MDuggan1.
(+) * Research.
(+) * Prepare date/time string parsing tests.
(20.11.2013.)
(+) * (Jurko) Start documenting the upcoming suds 0.4.1 jurko 6 release.
(+) * (Jurko) Fix suds time-zone handling according to a pull request received
(+) on BitBucket from MDuggan1.
(+) * Research.
(+) * Implement parsing.
(+) * DateTime no longer derived from Date & Time.
(+) * Date constructed from datetime.datetime should hold a datetime.date.
(+) * Research.
(+) * See if sax.date.Date("1900-01-01+02:00") should hold a timezone
(+) aware date object.
(+) * Related test: TestDate.testStringToValue().
(+) * Timezone data.
(+) * See when we need to specify timezone information.
(+) * In tests when we create DateTime/Time objects.
(+) * Default timezone when parsing web service responses.
(+) * YAGNI - for now left to user code reading specific
(+) DateTime/Time objects.
(+) * Default timezone when constructing web service requests.
(+) * YAGNI - for now left to user code creating specific
(+) DateTime/Time objects.
(+) * Implement.
(+) * Contained datetime.DateTime/Time objects should hold their
(+) timezone information.
(+) * Research.
(+) * Test TestDate.testStringToValue_failure() fails on Python 2 but
(+) passed on Python 3.
(+) * Fixed by the latest implementation changes.
(21.11.2013.)
(+) * (Jurko) Fix suds time-zone handling according to a pull request received
(+) on BitBucket from MDuggan1.
(+) * Check for feedback from users requesting this patch.
(+) * Add tests.
(+) * FixedOffsetTimezone class.
(+) * Fix FixedOffsetTimezone.tzname() output string formatting bug with
(+) negative timezone offsets.
(+) * Add tests.
(+) * UtcTimezone class.
(+) * (Jurko) Remove support for timezone specifiers including seconds as such
(+) are not supported by either Python or the XSD data types specification.
(+) * Add/update tests.
(+) * Input like "+10:10:10" should be rejected.
(+) * Timezone offset timedelta objects containing more detailed than
(+) minute information.
(+) * Remove support.
(+) * (Jurko) Add tests making sure timezone indicator strings without a colon
(+) between hours and minutes are not accepted.
(+) * This leads to border cases like "+121" where you do not know whether
(+) this represents "+01:21" or "+12:01".
(22.11.2013.)
(+) * (Jurko) Remove date/time related test code duplication.
(+) * Date, DateTime & Time classes.
(+) * Clean up test function names.
(+) * Test construction from unexpected objects.
(+) * Test construction from datetime.date/datetime/time objects.
(+) * str() tests.
(+) * XDate & Date class.
(+) * XDateTime & DateTime class.
(+) * XTime & Time class.
(+) * Timezone handling checks in XDateTime & XTime classes.
(23.11.2013.)
(+) * (Jurko) Make converting datetime/time to string output subsecond
(+) information without trailing zeroes.
(+) * Test.
(+) * Implement.
(+) * Discard as the implementation complexity seems way too great and the
(+) gain does not seem to reciprocate the cost.
(+) * (Jurko) Update the project's versioning scheme to no longer have pip
(+) detect suds-jurko releases as 'prerelease' only due to our version tag
(+) formatting.
(+) * Accept that the original suds project has died and continue with the
(+) natural version number progression.
(+) * (Jurko) Plan a new release.
(+) * (Jurko) Update used setuptools version.
(+) * (Jurko) Remove unused project files inherited from the original suds
(+) project.
(25.11.2013.)
(+) * (Jurko) Test the project with different Python installations.
(+) * Python 2.4.3/x86, on Windows 7/SP1/x64.
(+) * Install.
(+) * 'setuptools'.
(+) * 'pip'.
(+) * Describe encountered problems in 'HACKING.txt'.
(+) * 'pytest'.
(+) * Describe encountered problems in 'HACKING.txt'.
(+) * Run tests.
(+) * Python 2.4.4/x86, on Windows 7/SP1/x64.
(+) * Install.
(+) * 'setuptools'.
(+) * 'pip'.
(+) * Describe encountered problems in 'HACKING.txt'.
(+) * 'pytest'.
(+) * Describe encountered problems in 'HACKING.txt'.
(+) * Run tests.
(+) * Python 2.7.6/x64, on Windows 7/SP1/x64.
(+) * Install.
(+) * 'setuptools'.
(+) * 'pip'.
(+) * 'pytest'.
(+) * Run tests.
(+) * Python 3.2.5/x64, on Windows 7/SP1/x64.
(+) * Install.
(+) * 'setuptools'.
(+) * 'pip'.
(+) * 'pytest'.
(+) * Run tests.
(+) * Python 3.3.3/x86, on Windows 7/SP1/x64.
(+) * Install.
(+) * 'setuptools'.
(+) * 'pip'.
(+) * 'pytest'.
(+) * Run tests.
(+) * Python 3.3.3/x64, on Windows 7/SP1/x64.
(+) * Install.
(+) * 'setuptools'.
(+) * 'pip'.
(+) * 'pytest'.
(+) * Run tests.
(+) * (Jurko) Document the test environment setup in HACKING.txt.
(+) * (Jurko) Prepare a new suds-jurko 0.5 release.
(28.11.2013.)
(+) * (Jurko) Look into a reported problem with how unicode data gets encoded
(+) inside a suds SOAP request with Python 2. Reported by Alexey Sveshnikov
(+) and mduggan1.
(+) * Get a reproducible use case from Alexey Sveshnikov.
(+) * Research.
(+) * Data in HTTP requests needs to be encoded as defined by the
(+) Content-Type header given in that request (ISO-8859-1 being the
(+) default).
(+) * Suds set the "Content-Type=text/xml; charset=utf-8" HTTP request
(+) header for all of its SOAP requests.
(+) * Python's http module (used internally by the urllib module) add
(+) the given message data to its existing data. Suds gives its
(+) message data to urllib as an utf-8 encoded bytes object. If
(+) existing message data is unicode, it will attempt to forcefully
(+) convert the given message data to unicode assuming all it contains
(+) is ASCII characters.
(+) * Suds message data is already a utf-8 encoded bytes object.
(+) * With Python-3 httplib's previous message data is a bytes and not a
(+) string object.
(+) * The reason why httplib's message content is converted to unicode
(+) is that with Python 2 the initial header is a unicode object
(+) u'POST /service HTTP/1.1' while with Python 3 it is a bytes object
(+) b'POST /service HTTP/1.1'.
(+) * Python 2.
(+) * Affects Python 2.7.
(+) * Does not affect Python 2.4.
(+) * Python 3.
(+) * Its httplib Request object automatically converts the
(+) passed URL to a bytes object (assumes it contains only
(+) ASCII characters) which then prevents all the other
(+) request data from being forcibly converted to unicode.
(29.11.2013.)
(+) * (Jurko) Look into a reported problem with how unicode data gets encoded
(+) inside a suds SOAP request with Python 2. Reported by Alexey Sveshnikov
(+) and mduggan1.
(+) * Reduce the reproducible use case.
(+) * Should not require an external web service.
(+) * Integrate into regular suds-jurko tests.
(+) * Make the reproducible test case not attempt to connect to the network
(+) if the test passes.
(+) * Fix the issue.
(+) * Confirm with Alexey Sveshnikov that it is ok with him to make the
(+) reproducible use case public.
(+) * Close related project pull requests on 'bitbucket.org'.
(+) * (Jurko) Add a test for handling actual non-ASCII unicode service location
(+) data.
(30.11.2013.)
(+) * (Jurko) See if the suds HttpTransport.open() method ever gets called.
(+) * Yup, transport open() methods get called from DocumentReader, e.g.
(+) when downloading a WSDL schema from the net.
(+) * It seems like Transport's open() & send() methods might be mergeable,
(+) but that would first require further research. For now - YAGNI.
(+) * (Jurko) Process pull requests received on 'bitbucket.org'.
(+) * Fix setup.py current working folder path comparison so it works with
(+) links in the path. Contributed by ryanpetrello.
(23.12.2013.)
(+) * (Jurko) Prepare a basic development script for running the full suds test
(+) suite using multiple Python interpreter versions.
(26.12.2013.)
(+) * (Jurko) Process patches sent in by Bouke Haarsma on BitBucket.
(+) * Unicode logging issue.
(+) * Research.
(+) * Prepare tests.
(+) * Merge.
(+) * Thorough code review.
(+) * Implement a cleaner fix for both Reply & Request classes.
(+) * Update release notes.
(+) * (Jurko) Fix possible typo in the suds.transport.Request string/unicode
(+) representation where there seems to be a missing space after a colon just
(+) before the URL information.
(+) * (Jurko) Remove unnecessary logger objects.
(+) * (Jurko) Process the project issue #2 reported on BitBucket by Arthur
(+) Clune, related to not being able to set the option cache location if the
(+) default cache location is not a writable folder.
(+) * Research.
(+) * Prepare tests.
(+) * Default cache.
(+) * Default ObjectCache instance should be created only if no other
(+) cache has been explicitly specified during suds.client.Client()
(+) construction.
(+) * Fix.
(+) * Update release notes.
(+) * Report back & close the project issue on BitBucket.
(21.01.2014.)
(+) * (Jurko) Process patch sent in by Bouke Haarsma on BitBucket to 'make sure
(+) all web service operation parameters are consumed'.
(+) * This patch has been worked on on a separate feature branch for close
(+) to a month now.
(+) * Update todo list.
(+) * Update release notes.
(+) * Commit.
(+) * Push changes to BitBucket.
(23.01.2014.)
(+) * (Jurko) Prepare internal documentation notes folder.
(+) * Add folder & add a descriptive readme.txt file to it.
(+) * Include in the source distribution folder.
(+) * Do not include it in the installation.
(+) * Note the new documentation notes folder in the project's HACKING.rst
(+) documentation.
(24.01.2014.)
(+) * (Jurko) Process Jurko's research & todo notes collected while working on
(+) reporting extra parameter errors.
(+) * (Jurko) Prepare a new suds-jurko 0.6 release.
(+) * Check release notes.
(+) * Test.
(+) * Update version tag.
(+) * Tag in Hg.
(+) * Update version information.
(+) * Retag in Hg
(+) * Package the release.
(+) * Distribute the new release.
(+) * PyPI.
(+) * BitBucket.
(+) * Notify Gauthier Bastien - see comments for commit
(+) 9da81de891958292850a242781b30a3493f617 on BitBucket.
(+) * Prepare project information for the next development cycle.
(25.01.2014.)
(+) * (Jurko) Look into issue #12 & the related pull request #22 reported on
(+) BitBucket - suds-jurko incorrectly returning a raw Fault() instead of a
(+) WebFault() object when faults=False. Reported as a regression against the
(+) base suds project.
(+) * Research.
(+) * Most likely introduced by commit
(+) '506c8362c81343f1f629906606db23daf8b426ec'.
(+) * Try to reproduce using an injected reply.
(+) * Test against the current HEAD commit.
(+) * Find a commit not that works correctly.
(+) * Status.
(+) * Could not reproduce the issue. Truly suds-jurko returns a
(+) suds.sudsobject.Fault object, but the original suds library
(+) implementation does not return a suds.WebFault object either.
(+) Further work on this can be done only once the exact desired
(+) behaviour is defined.
(+) * Ask for feedback containing a reproducible example.
(26.01.2014.)
(+) * (Jurko) Code cleanup.
(+) * Stylistic cleanup.
(+) * (Jurko) Process decimal type support patch sent in by pendletongp on
(+) BitBucket.
(+) * Research.
(+) * Where a Python type to SOAP representation transformation is
(+) performed.
(+) * Expected locations.
(+) * When constructing a web service operation invocation
(+) request, i.e. when marshaling Python data to a SOAP
(+) request XML (mx package).
(+) * Where a SOAP representation to Python type transformation is
(+) performed.
(+) * Expected locations.
(+) * When processing a web service operation reply, i.e. when
(+) unmarshaling Python data from a SOAP response XML (umx
(+) package).
(28.01.2014.)
(+) * (Jurko) Process decimal type support patch sent in by pendletongp on
(+) BitBucket.
(+) * Prepare tests.
(+) * Representation tests for different data types.
(+) * Transformation tests for different data types.
(+) * Thorough code review.
(+) * Add XFloat tests.
(29.01.2014.)
(+) * (Jurko) Update XFloat to correctly translate decimal.Decimal &
(+) numbers.Real (on Python 2.6+) values to their XSD value representation.
(+) Currently, inconsistently with how other types are handled, they get
(+) translated to the same decimal/fraction value which then gets changed to a
(+) string later on when actually writing it to an XML string. This works out
(+) fine for decimals in the end but not for rationals/fractions.
(+) * Research.
(+) * decimal.Decimal input values already get entered correctly into
(+) constructed SOAP XML requests.
(+) * It seems not all numbers.Real derived classes can be converted to
(+) their XSD value representation the same way.
(+) * Some examples.
(+) * float --> str(x)
(+) * fractions.Fraction --> str(float(x))
(+) * decimals.Decimal --> str(x)
(+) * int --> str(x)
(+) * bool --> str(int(x))
(+) * Adding support for them would pair down to simply listing all
(+) supported types explicitly and adding code for translating
(+) each of them separately. We can not possibly do this for 'all
(+) data types in the world' so the current implementation seems
(+) as good as any.
(+) * A user application requiring support for a specific type can
(+) always add it itself by implementing a replacement XFloat class
(+) and registering it using suds.xsd.xbuiltin.Factory.maptag().
(+) * Discard the task.
(+) * Add a new development note document describing XType class usage for
(+) built-in XSD type value translation between Python objects and their
(+) XSD value representations.
(+) * (Jurko) Clean up different XType.translate() methods. See if there is any
(+) difference in them returning a string or any other Python object
(+) convertible to a string using str().
(+) * If there is no difference, clean up the code to simply return the
(+) original value where possible.
(+) * (Jurko) Process decimal type support patch sent in by pendletongp on
(+) BitBucket.
(+) * Research.
(30.01.2014.)
(+) * (Jurko) Process decimal type support patch sent in by pendletongp on
(+) BitBucket.
(+) * Research.
(+) * XSD type value representation is incorrect.
(+) * Apply equivalent changes together with adding related unit tests.
(+) * Comment on the original issue #3 & patch request #19 on BitBucket.
(+) * Update release notes.
(+) * Later ideas to consider.
(+) * 'float' --> 'decimal.Decimal' conversion which has not been
(+) implemented in Python before Python 2.7.
(+) * YAGNI for now.
(+) * Update release notes.
(31.01.2014.)
(+) * (Jurko) Restore last_sent()/last_received() getters for last used SOAP
(+) request/response messages.
(+) * Research background.
(+) * Requested by.
(+) * andreebrazeau as BitBucket.
(+) * 'http://bitbucket.org/andreebrazeau'.
(+) * FerCesar at BitBucket.
(+) * 'http://bitbucket.org/FerCesar'.
(+) * Lucas Sampaio at BitBucket.
(+) * 'http://bitbucket.org/lucassmagal'.
(+) * Related links.
(+) * 'http://stackoverflow.com/questions/4426204/
(+) how-can-i-output-what-suds-is-generating-receiving'.
(+) * 'http://jortel.fedorapeople.org/suds/doc/
(+) suds.client.Client-class.html'
(+) * Code removed in commit 'f0034d6826c179625478bc19ae2a39a0b803fc3a'.
(03.02.2014.)
(+) * (Jurko) Restore last_sent()/last_received() getters for last used SOAP
(+) request/response messages.
(+) * Clean up related code.
(05.02.2014.)
(+) * (Jurko) Clean up SoapClient/SimClient related code a bit.
(+) * Generic code cleanup.
(+) * Mark private methods using leading underscores.
(+) * Remove RequestContext.original_envelope attribute.
(+) * Stop logging incorrect 'original_envelope' value.
(+) * Rename to _SoapClient/_SimClient.
(06.02.2014.)
(+) * (Jurko) Add tests for the functionality potentially affected by a planned
(+) _SoapClient/_SimClient refactoring.
(+) * Cache - update existing tests.
(+) * File operation failures.
(07.02.2014.)
(+) * (Jurko) Update FileCache duration handling.
(+) * Research.
(+) * Currently FileCache.__set_duration() implementation allows using
(+) only a single unit.
(+) * Check whether Python 2.4 datetime.timedelta class implementation
(+) allows it to be constructed from multiple units.
(+) * Store FileCache.duration as a datetime.timedelta instance.
(+) * Support for specifying FileCache durations with multiple time units.
(+) * Add tests.
(+) * Default suds.Client cache.
(+) * Already tested in test_default_cache_construction().
(+) * Class.
(+) * Duration.
(+) * Default duration.
(+) * Set duration (<0, 0, >0, single/multiple duration arguments).
(+) * Exception on duration < 0.
(+) * YAGNI. Current implementation, using negative values as
(+) regular durations and simply adding them to a specific file
(+) timestamp to see whether it expired seems good enough as well
(+) as intuitive enough for the user.
(+) * Expiration after expiration time with duration > 0.
(+) * Not expiring after expiration time with duration = 0
(+) * Not expiring before expiration time with duration > 0.
(+) * Not expiring before expiration time with duration = 0.
(+) * Duration testing for all FileCache derived classes - DocumentCache
(+) & ObjectCache.
(+) * FileCache and derived classes - one item expiring should not
(+) affect other unexpired items.
(+) * Update release notes.
(+) * (Jurko) Per process default FileCache folder randomization + removal on
(+) exit.
(+) * Research.
(+) * Related BitBucket issue #15.
(+) * Links from the BitBucket issue #15.
(+) * 'https://bugzilla.redhat.com/show_bug.cgi?id=978696'.
(+) * 'http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2013-2217'.
(09.02.2014.)
(+) * (Jurko) Per process default FileCache folder randomization + removal on
(+) exit - fix CVE-2013-2217.
(+) * Add tests.
(+) * Implement.
(+) * Update release notes.
(+) * Update related issues on BitBucket.
(+) * (Jurko) Extract all suds.client.Client related tests related to how it
(+) uses its worker cache/store/transport components into separate
(+) test_client.py test module. Modules like test_transport.py or
(+) test_cache.py should concentrate on testing that specific component
(+) functionality and not how the component gets used by suds.client.Client.
(+) * Test modules to check.
(+) * 'test_cache.py'.
(+) * 'test_client_cache.py'.
(+) * 'test_transport.py'.
(+) * Several transport usage tests already prepared in test_transport_1.py
(+) on Jurko's notebook.
(10.02.2014.)
(+) * (Jurko) Add tests for the functionality potentially affected by a planned
(+) _SoapClient/_SimClient refactoring.
(+) * Client's cache/store/transport usage.
(+) * WSDL access.
(+) * If an object is found in cache it should not be looked up in
(+) the given document store or transported.
(+) * If an object is found in a given document store it should not
(+) be transported.
(14.02.2014.)
(+) * (Jurko) Clean up test_client.py test module - extract all remaining
(+) suds.client.Client related tests related to how it uses its worker cache/
(+) store/transport components.
(+) * Test modules to check.
(+) * 'test_transport_http.py'.
(+) * (Jurko) Update test_sending_non_ascii_data_to_unicode_URL() test to test
(+) both send() & open() transport operations.
(17.02.2014.)
(+) * (Jurko) Clean up Transport ASCII/Unicode URL/data handling.
(+) * Design desired behaviour.
(+) * URL may be specified as either a byte string or a unicode string.
(+) * URL input data may contain ASCII characters only.
(+) * URL stored internally as a native str type (byte string with
(+) Python versions prior to 3, unicode string with Python 3+).
(+) * Using it in a different format breaks the underlying
(+) HttpTransport httplib implementation under some Python
(+) versions.
(+) * Bytes data used directly.
(+) * Unicode data used directly with the underlying implementation
(+) converting the data to bytes or raising an error.
(+) * HttpTransport.
(+) * Python2 - converts to bytes if ASCII chars only.
(+) * Python2 - error for non-ASCII chars.
(+) * Python3 - error.
(+) * Fix suds.transport Reply & Request string representation.
(+) * Update suds.transport.Request.
(+) * Check for non-ASCII URL characters.
(+) * Hold it internally as a native str type.
(+) * Update suds.transport.http.HttpTransport.
(+) * Expects the suds.transport.Request to have already checked its URL
(+) content.
(+) * Document the desired design (docstrings).
(+) * Request.
(+) * Transport.
(+) * Update tests.
(+) * Request tests.
(+) * Bytes URL.
(+) * Must be used directly on Python 2.
(+) * Must be converted to str URL internally on Python 3.
(+) * Error if it includes non-ASCII character codes.
(+) * Unicode URL.
(+) * Must be converted to str URL internally on Python 2.
(+) * Must be used directly on Python 3.
(+) * Error if it includes non-ASCII character codes.
(+) * HttpTransport tests.
(+) * Sending data through the network.
(+) * Remove tests made redundant by the new Request testing.
(+) * Update release notes.
(18.02.2014.)
(+) * (Jurko) Add tests for the functionality potentially affected by a planned
(+) _SoapClient/_SimClient refactoring.
(+) * Client's cache/store/transport usage.
(+) * If an object is found in cache it should not be looked up in the
(+) given document store or transported.
(+) * WSDL document.
(+) * Imported XSD schema.
(+) * Included XSD schema.
(19.02.2014.)
(+) * (Jurko) Add a failing test illustrating a WSDL import detected when
(+) attempting to test importing a cached WSDL schema.
(+) * test_WSDL_import().
(+) * (Jurko) Add tests for the functionality potentially affected by a planned
(+) _SoapClient/_SimClient refactoring.
(+) * Client's cache/store/transport usage.
(+) * If an object is found in cache it should not be looked up in the
(+) given document store or transported.
(+) * Imported WSDL document.
(+) * If an object is found in a given document store it should not be
(+) transported.
(+) * WSDL document.
(+) * Imported WSDL document.
(+) * Imported XSD schema.
(+) * Included XSD schema.
(+) * If an object is not cached or found in a given document store it
(+) should not transported.
(+) * WSDL document.
(+) * Imported WSDL document.
(+) * Imported XSD schema.
(+) * Included XSD schema.
(+) * Client's cache/store/transport usage.
(+) * Different cachingpolicy option values.
(+) * cachingpolicy == 0 - XML documents are cached.
(+) * Already tested.
(+) * cachingpolicy == 1 - final WSDL objects are cached.
(+) * Loading a WSDL object from cache should avoid attempting
(+) to fetch any additional external data either from the
(+) cache, the document store or the registered transport.
(20.02.2014.)
(+) * (Jurko) Add tests for the functionality potentially affected by a planned
(+) _SoapClient/_SimClient refactoring.
(+) * Transport.
(+) * Base Transport, Reply & Request classes already tested.
(+) * HttpTransport.
(+) * urllib2.HTTPError raised from urllib's open() operation.
(+) * ACCEPT/NO_CONTENT status handling - buggy but YAGNI for
(+) now.
(+) * Add a separate todo item to look into this at some
(+) later time.
(+) * Other HTTPError exceptions.
(+) * Other functionality already tested well enough.
(+) * _SoapClient - currently can only be tested via suds.client.Client.
(+) * Using Transport.
(+) * 'nosend' option avoids transport usage.
(+) * Sending data via & using data returned by transport.
(+) * Opening a WSDL already tested.
(+) * Sending a web service operation invocation request.
(+) * Handling transport errors.
(+) * This seems like a complex topic. Prepare a rough test and
(+) add a todo item to deal with this in more detail later on.
(+) * Embed a TODO comment next to the new test.
(23.02.2014.)
(+) * (Jurko) Refactor suds.transport.http module unit tests to make them
(+) clearer.
(27.02.2014.)
(+) * (Jurko) Fix the raw try:/except: exception handling suds which may eat up
(+) internal exceptions like SystemExit or KeyboardInterrupt but should not.
(+) * Just replace with Exception subclass catching. A more detailed
(+) exception class can be used later on if needed.
(28.02.2014.)
(+) * (Jurko) Fix Exception message used when attempting to construct a
(+) suds.sax.element.Element with a non-Element parent. It seems like someone
(+) forgot to apply % formatting there.
(+) * (Jurko) Extract existing XML comparison test utilities into a separate
(+) tests/test_utility.py module (CompareSAX class).
(03.03.2014.)
(+) * (Jurko) CompareSAX test utility improvements.
(+) * Use pytest assertions to report errors.
(+) * Add tests.
(+) * Extract tests into a separate module.
(+) * (Jurko) Fix the problem with different processes using different cache ids
(+) for matching documents, causing them not to be able to reuse each other's
(+) cache entries. Original suds implementation used the built-in hash
(+) function which should be replaced with a md5 hash.
(+) * On Python 3.3 and above hash function results are seeded with a random
(+) value for each process.
(+) * 32-bit & 64-bit Python implementations use different hash
(+) implementations giving different results.
(+) * Update readme.
(+) * Problem reported by Eugene Yakubovich at bitbucket.
(+) * See project related comments at: 'http://bitbucket.org/eyakubovich/
(+) suds/commits/fec6efd9c10b48114bbd9c06847a264453a51620'.
(04.03.2014.)
(+) * (Jurko) suds.sax.enc module cleanup.
(+) * Code & comment cleanup.
(+) * Add unit tests.
(+) * (Jurko) Look into CDATA related encoding issues reported at BitBucket
(+) (issue #14 + pull requests #25 & #26).
(+) * Prepare tests.
(+) * Prepare profiling scripts for comparing different solutions.
(+) * Design input types.
(+) * Long input.
(+) * Replacements.
(+) * None.
(+) * Rare.
(+) * Lots.
(+) * CDATA.
(+) * None.
(+) * Short.
(+) * Long.
(+) * Short input.
(+) * Has replacements.
(+) * False.
(+) * True.
(+) * Has CDATA
(+) * False.
(+) * True.
(05.03.2014.)
(+) * (Jurko) Consider reimplementing suds.sax.enc.Encoder.decode() using
(+) xml.sax.saxutils.unescape() as suggested by Stephen Fuhry (fuhrysteve at
(+) BitBucket).
(+) * Profile using the prepared tests.profiling.profile_sax_encoder module.
(+) * Python 2.4.3.
(+) * About the same for data containing many replacements.
(+) * New solution much slower for data containing no replacements.
(+) * Python 3.3.3.
(+) * New solution a bit slower for data containing many
(+) replacements.
(+) * New solution much slower for data containing no replacements.
(+) * Discard the new solution as it works the same or worse than the
(+) current one.
(20.03.2014.)
(+) * (Jurko) Internal project development cleanup.
(+) * Setting up Python installations used for testing suds.
(+) * Research exact steps for installing required Python packages for
(+) each specific Python version used.
(+) * Prepare Windows batch script performing this operation en-masse.
(23.03.2014.)
(+) * (Jurko) Report 'except:' issues in setuptools project's ez_setup.py
(+) script.
(+) * (Jurko) Update the project setup procedure to use any preexisting
(+) setuptools installation of version >= 1.4 but install the latest
(+) compatible setuptools version if a suitable one is not already installed.
(+) * YAGNI - using an older and potentially untested setuptools version
(+) risks our project installation failing without the user 'doing
(+) anything wrong', and gains us only a bit lighter installation in some
(+) scenarios.
(+) * (Jurko) Update the project setup procedure to use setuptools 1.4.2 for
(+) Python versions prior to 2.6, and the latest setuptools version 3.3 for
(+) all newer Python versions.
(+) * Update the project's setup.py script.
(+) * Use setuptools 1.4.2 with Python < 2.6.
(+) * Use setuptools 3.3 with Python >= 2.6.
(+) * Update project hacking documentation.
(+) * Check for newer setuptools versions with every new suds release
(+) and update the project's setup.py script as needed.
(+) * Update release notes.
(15.05.2014.)
(+) * (Jurko) Process work done 'as time allowed' during the last two months.
(+) * Commit prepared basic environment setup scripts.
(+) * Mark related todo items now completed or rendered obsolete.
(+) * (Jurko) See if we can make the project setup.py script work when run from
(+) a folder other than the project's root folder.
(+) * Managed to improve the support a bit, but distutils still has issues
(+) regarding this. Should be checked again after a while to see if the
(+) situation has changed.
(+) * (Jurko) Internal project development cleanup.
(+) * Setting up Python installations used for testing suds.
(+) * Prepare a script for installing required Python packages.
(+) * Convert prepared Windows batch script to Python.
(+) * Required Python packages.
(+) * setuptools.
(+) * pip.
(+) * pytest.
(+) * Update project hacking documentation.
(+) * Allow setting up only specific Python versions.
(+) * Can be configured using setup.cfg for now.
(+) * Update to only install packages that have not already been
(+) installed.
(+) * Transformed to embedded TODO comments.
(+) * Make sure the prepared Python script runs using any of the
(+) suds project supported Python versions.
(+) * Collect all downloaded data under a single folder.
(+) * Targeted content.
(+) * ez_setup.py downloaded data.
(+) * easy_install downloaded data.
(+) * pip downloaded data.
(+) * See if using the pip download cache folder is useful at all
(+) since we already pre-download all the required installation
(+) packages and use them from there.
(+) * Prepare a test script installing suds into multiple Python versions.
(+) * Without setuptools preinstalled.
(+) * With setuptools preinstalled.
(+) * setuptools 1.4.
(+) * Latest compatible setuptools release.
(+) * setuptools 1.4.2 for Python releases prior to 2.6.
(+) * (Jurko) Plan preparing a wheel based distribution.
(+) * Test it.
(+) * Document.
(+) * Setting up the necessary development environment.
(+) * Release notes.
(+) * (Jurko) Find a cleaner way to install suds tests. Currently they get
(+) installed into a top-level 'tests' folder and so may cause conflicts with
(+) some other 'tests' package that might exist in the target Python
(+) environment.
(+) * Ideas to consider.
(+) * We want all suds users to have easy access to the suds test suite.
(+) * Suds users are programmers themselves, and so we want to make
(+) it easy for them to reproduce their issues as easily,
(+) concisely and directly as possible, and using existing suds
(+) tests seems like a perfect starting point for doing just that.
(+) * Placing the 'tests' folder under the 'suds' folder.
(+) * Placing the top level suds folder outside the egg folder (similar
(+) to how this is done for the pytest package), not having the egg
(+) folder added to the Python path and leaving the tests folder
(+) inside the egg folder.
(+) * Have setup.py install suds tests as a setuptools 'extra' - only if
(+) explicitly requested.
(+) * At first glance, this does not seem like something we would
(+) like done, as we want all suds users to have easy access to
(+) the suds test suite and not just those installing suds from
(+) its source distribution and providing some magic
(+) '--install-tests'-like command-line option.
(+) * Keep the tests in the source distribution and do not install them.
(+) This seems like the simplest way, so lets go with this for now.
(16.05.2014.)
(+) * (Jurko) Find a cleaner way to install suds tests. Currently they get
(+) installed into a top-level 'tests' folder and so may cause conflicts with
(+) some other 'tests' package that might exist in the target Python
(+) environment.
(+) * Update the tests so we no longer need to transform them using py2to3
(+) in order to run them using Python 3.
(+) * Try using the six compatibility package.
(+) * Manual install.
(+) * Update setup.py.
(+) * Update tools/setup_base_environments.py.
(+) * Update tests.
(+) * Make the tests run using the 'currently installed' suds version. In
(+) order to test the current development version, user should install the
(+) project in development/editable mode using 'pip install -e' or
(+) 'setup.py develop'.
(+) * Running 'setup.py test' should still work using both Python 2 & 3.
(+) * Make project tests run from the root project folder use the correct
(+) installed Python 3 sources if run under Python 3 instead of using the
(+) ones from the current folder.
(+) * No easy way to do this for now. Should look into it again at some
(+) later time.
(+) * Update project release docs.
(+) * Update project HACKING.rst docs.
(+) * Using the 'six' Python 2/3 compatibility package.
(+) * Running project tests using Python 3 - current working folder.
(22.05.2014.)
(+) * (Jurko) Test using Python 3.4.1.
(26.05.2014.)
(+) * (Jurko) Remove code duplication between setup.py &
(+) tools/setup_base_environments.py, e.g. with regards to setting up the
(+) project's testing requirement packages (argparse, colorama, py, pytest).
(29.05.2014.)
(+) * (Jurko) Internal project development cleanup.
(+) * Re-implement the existing run_all_tests.cmd Windows batch script for
(+) running a full suds test suite on multiple Python versions in Python.
(+) * Read the list of target Python environments from the project's
(+) setup.cfg configuration file.
(+) * Make sure the prepared Python script runs using any of the suds
(+) project supported Python versions.
(+) * Add support for disabling testing a specific Python platform.
(+) * YAGNI for now. Can be done by commenting out appropriate
(+) environment definition sections in the main project 'setup.py'
(+) configuration file and will be affected later on by porting
(+) our testing system over to use the tox project.
(+) * Update HACKING.rst documentation.
(14.06.2014.)
(+) * (Jurko) Add basic suds.sax.element unit tests.
(15.06.2014.)
(+) * (Jurko) Add tests for suds.sax.element.Element string conversion.
(+) * (Jurko) Add tests for suds.sax.document.Document string conversion.
(+) * (Jurko) Fix suds.sax.document.Document str conversion bug.
(+) * (Jurko) Fix original suds project bug with suds.cache.DocumentCache not
(+) caching suds.sax.document.Document instances correctly, even though such
(+) instances get passed to it by the suds.reader.DocumentReader.
(+) * (Jurko) Fix problems with suds silently hiding plugin exceptions.
(+) * Add plugin unit testing module
(+) * Test that calling a plugin using PluginContainer passes any
(+) exceptions raised by the plugin.
(+) * Fix the PluginContained exception passing issue.
(+) * Close issue #42 at BitBucket.
(+) * Close pull request #33 at BitBucket.
(+) * (Jurko) Fix problem with decimal zero representations having a negative
(+) exponent, e.g. as it "0.0000".
(+) * Add tests.
(+) * Fix failing assert.
(+) * Close pull request #37.
(+) * Close issue #31.
(23.06.2014.)
(+) * (Jurko) Look at issue #49 at BitBucket - generated SOAP request containing
(+) a tag with a missing namespaces identifier.
(+) * Research.
(+) * Caused by incorrect XSD schema element form attribute handling
(+) when the element is actually just a reference to another top-level
(+) element. In such cases the the referenced and not the referencing
(+) element's form attribute value should be used.
(+) * Prepare a quick-fix plugin.
(+) * Report back to the BitBucket issue tracker.
(24.06.2014.)
(+) * (Jurko) SchemaObject.form_qualified cleanup.
(+) * Add tests for the schema element's form attribute handling.
(+) * Element with form attribute set to 'qualified'.
(+) * Schema with elementFormDefault set to 'qualified'.
(+) * Schema with elementFormDefault set to other than 'qualified'.
(+) * Schema with elementFormDefault unset.
(+) * Element with form attribute set to other than 'qualified'.
(+) * Schema with elementFormDefault set to 'qualified'.
(+) * Schema with elementFormDefault set to other than 'qualified'.
(+) * Schema with elementFormDefault unset.
(+) * Element with unset form attribute.
(+) * Schema with elementFormDefault set to 'qualified'.
(+) * Schema with elementFormDefault set to other than 'qualified'.
(+) * Schema with elementFormDefault unset.
(26.06.2014.)
(+) * (Jurko) Remove unnecessary 'history' list copying in
(+) suds.xsd.sxbase.SchemaObject.content().
(+) * (Jurko) suds.xsd.deplist module cleanup.
(+) * Stylistic code & comment cleanup.
(+) * PEP8-ify.
(+) * Make pop() not ignore exceptions.
(+) * Make add() no longer return self.
(+) * Remove sorted member.
(+) * Order methods alphabetically.
(+) * Mark private operations & data as private.
(+) * Review the topological sort implementation.
(+) * `popped` list seems to be filled only for its elements to later be
(+) moved over to the `sorted` list.
(+) * Add topological sorting tests by extracting the existing embedded test
(+) into a stand-alone unit test module.
(+) * (Jurko) Look at BitBucket issue #50 - SOAP 1.2 compatibility.
(+) * (Jurko) Look at BitBucket issue #51 - complexType 'mixed' attribute.
(27.06.2014.)
(+) * (Jurko) suds.xsd.deplist module cleanup.
(+) * Reference cycle handling.
(+) * Research.
(+) * Document.
(+) * Test.
(+) * Review the topological sort implementation.
(+) * See if there is anything obvious to be made more efficient in it.
(+) * Simplify code by using a recursive implementation.
(+) * Refactor the DepList class into a single function dependency_sort()
(+) function taking a single dependency dictionary as input.
(+) * Refactor code.
(+) * Rename module to suds.xsd.depsort.
(+) * Update release notes.
(+) * 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.
(+) * (Jurko) SchemaObject.form_qualified cleanup.
(+) * Add tests for the schema element's form attribute handling.
(+) * Reference elements.
(+) * Unqualified referencing qualified.
(+) * Referencing & referenced elements should have different
(+) namespaces.
(+) * Qualified referencing unqualified.
(28.06.2014.)
(+) * (Jurko) SchemaObject.form_qualified cleanup.
(+) * Fix detected issues.
(+) * Fix failing tests.
(+) * Mark tests as no longer expected to fail.
(+) * Update project release notes.
(+) * Report back to issue #49 on BitBucket.
(29.06.2014.)
(+) * (Jurko) SchemaObject.form_qualified cleanup followup.
(+) * Research the XSD specification on the element form attribute.
(+) * Correct suds's XSD element form attribute handling.
(+) * Reference elements are always qualified.
(+) * Top-level elements are always qualified.
(+) * Non-top-level non-reference elements are qualified based on their
(+) form attribute, and if they have no form attribute, their schema's
(+) elementFormDefault attribute is used instead, and if that one is
(+) not given either - they are considered unqualified.
(28.06.2015.)
(+) * (Jurko) Clean up WSDL imports.
(+) * Remove invalid test_WSDL_import() test & related comment in
(+) `test_client.py`.
(+) * Add WSDL import tests.
(+) * Import WSDL with same target namespace.
(+) * Import WSDL with different target namespace.
(+) * (Jurko) Fix recursive WSDL import issue.
(+) * Add tests.
(29.06.2015.)
(+) * (Jurko) Fix recursive WSDL import issue.
(+) * Fix issue.
(30.06.2015.)
(+) * (Jurko) Review and commit unpublished cleanup work on Jurko's machine.
(01.07.2015.)
(+) * (Jurko) Fix recursive XSD import issue.
(+) * Add tests.
(+) * Fix issue.
(14.07.2015.)
(+) * (Jurko) Fix issue https://bitbucket.org/jurko/suds/issues/90 - code in
(+) `mx/typer.py` Typer.genprefix() broken by commit
(+) 103efca86cd07e9d7205ffb2735748f275a4b35f.
(+) * Add todo note to improve test coverage that managed to miss this
(+) problem.
(+) * Fix.
(+) * (Jurko) Add basic unit tests for Typer.genprefix().
(+) * Got broken in commit 103efca86cd07e9d7205ffb2735748f275a4b35f and
(+) fixed as part of issue https://bitbucket.org/jurko/suds/issues/90 but
(+) the bug was missed by the existing test suite.
(25.07.2015.)
(+) * (Jurko) Clean up some issues from BitBucket.
(26.07.2015.)
(+) * (Jurko) Clean up some issues from BitBucket.
(27.07.2015.)
(+) * (Jurko) Clean up some issues from BitBucket.
(28.07.2015.)
* (Jurko) Review and commit unpublished cleanup work on Jurko's machine.
* (Jurko) Clean up WSDL imports.
* Add tests for the compatibility kludge with existing WSDL schemas in
the wild that use WSDL imports to import XSD schemas.
* `Definitions` object is also used to load XSD schema content
referenced using a WSDL import element. This should be reviewed as it
seems like a quick-hack. According to the `Definitions` class comment,
its instances are intended to model data loaded from WSDL documents
only. If we want to load XSD documents and WSDL documents the same, we
should load them as generic XML documents (or at least check their
root XML element) and proceed from there with a different loading
system.
* An importing WSDL should not be allowed to reference XSD entities from
an imported WSDL. If such data needs to be referenced, it should be
imported from an external XSD schema (either using XSD import or using
the questionably standard and not widely supported WSDL import
variation importing an XSD instead of a WSDL schema).
* We could possibly retain this XSD importing feature as an optional
kludge for compatibility with broken web services.
* Makes no sense to collect all types, messages, port_types & bindings
data from imported WSDLs.
* We currently collect types, messages, port_types & bindings
entities but do not collect services or integrate schema elements.
* All this seems quite messy.
* See why this data is actually collected.
* Random thoughts.
* Possibly collect only the actually used ones, i.e. those
referenced from the importing WSDL's entities.
* If the collected data is just some internal implementation
detail, mark its member variables as private (name with a `__`
prefix) and comment on what it is they actually contain.
* (Jurko) SchemaObject.form_qualified cleanup followup.
* Correct suds's XSD element form attribute handling.
* Qualified XSD elements are connected to a namespace.
* Unqualified XSD elements are not connected to any namespace.
* XML elements matching an unqualified XSD schema element do not
necessarily need to have no namespace specified. If their parent
element's default namespace is specified, then they still need to
explicitly specify their default namespace as an empty string.
* Test SOAP XML construction for qualified/unqualified elements with
and without the `prefix` suds option enabled.
* Document breaking changes between this fork and the original suds
project.
* DocumentStore interface changes.
* Choice support.
* SchemaObject.form_qualified should not have 'form_qualified'. That
seems like something related only to elements & attributes.
* SchemaObject.dependency() method seems too complicated - its interface
allows returning a list of dependencies, with a specific one amongst
them singled out as a dependency to be merged, when in fact it always
returns no dependencies or only a single dependency with that single
dependency to be merged.
* Rename to dereference() and have it return the referenced
SchemaObject - since that is in fact what this function already
does.
* SchemaObject.merge() is misnamed - it should be named
collect_referenced_data() since it is a completely different operation
from Schema.merge() and it actually only collects data from a
referenced object, if such an object exists.
* Look at issue #46 at BitBucket - form attribute on schema attributes
does not seem to be handled correctly.
* Schema 'attributeFormDefault' attribute handling.
* Schema attribute 'form' attribute handling.
* (Jurko) Look at the Spyne project and its suds usage.
* Try running Spyne's test suite using the suds-jurko fork.
* Related links.
* 'https://spyne.ci.cloudbees.com/job/spyne/PYFLAV=2.7/
lastCompletedBuild/testReport/spyne.test.interop.test_suds/
TestSuds/'.
* 'https://github.com/arskom/spyne/blob/master/spyne/test/
interop/test_suds.py'.
* Notes from the Spyne maintainer Burak Arslan.
* You can run Spyne tests against your suds locally. Just clone
the Spyne repo and edit setup.py to use suds-jurko instead of
suds here:
'https://github.com/arskom/spyne/blob/master/setup.py#L344'.
* Send feedback back to the Spyne maintainer (burak.arslan at
arskom.com.tr).
* Try updating Spyne project test system so it can be easily configured
to use suds-jurko instead of the official suds release.
* (Jurko) Try out virtualenv to make it easier to run tests using multiple
Python versions and to allow us to automate suds installation testing in
the future.
* Compatible virtualenv versions.
* virtualenv - 1.7.2 (last before 1.8) - last on Python 2.4.
* virtualenv - 1.9.1 (last before 1.10) - last on Python 2.5.
* (Jurko) Internal project development cleanup.
* Prepare a full suds test suite subset that can be used for regular
development - select only a few platforms to test suds on regularly
and run the full test suite only 'when needed', e.g. before pushing
changes onto BitBucket or at the end of a single development session.
* (Jurko) Add an XML comparison test utility supporting raw XML data
comparison with the following features.
* Important comparison features.
* How textual data is split between child elements.
* Child element ordering.
* Element/attribute namespace + name.
* Non-namespace declaration related attribute ordering.
* Unimportant comparison features.
* Leading/trailing textual data whitespace per line for mixed
content nodes containing at least one child element.
* Element/attribute namespace prefix.
* Namespace declaration related attribute ordering.
* Namespace declarations (no prefix).
* Namespace prefix declarations.
* (Jurko) Add tests for the functionality potentially affected by a planned
_SoapClient/_SimClient refactoring.
* _SoapClient - currently can only be tested via suds.client.Client.
* 'prettyxml' option.
* 'retxml' option.
* _SimClient - currently can only be tested via suds.client.Client.
* Using Transport for sending an injected request.
* Not using Transport when a reply has been injected.
* _SimClient using injected request.
* _SimClient using injected reply.
* Success.
* Failure.
* Status & status description.
* 'nosend' option.
* With injected request data.
* With injected response data - ignored.
* 'retxml' option.
* 'prettyxml' option.
* RequestContext.
* Returned from _SoapClient.
* Returned from _SimClient.
* Processing an externally provided reply.
* Plugin processing.
* Variants.
* _SoapClient/_SimClient.
* _SimClient with injected reply/response.
* 'nosend' option + RequestContext reply.
* 'retxml'.
* 'prettyxml'.
* Fetched documents.
* document.loaded(url, bytes) --> bytes.
* document.parsed(url, document).
* Documents read from cache.
* document.parsed(url, document).
* Client initialized with WSDL.
* init.initialized(wsdl).
* SOAP requests.
* marshalled(document) - may modify document in-place.
* sending(bytes) --> bytes.
* SOAP replies.
* received(bytes) --> bytes.
* parsed(element) --> may modify document in-place.
* unmarshalled(result) --> result.
* (Jurko) Refactor _SoapClient/_SimClient related code.
* Refactor.
* Already prepared as client_1.py on Jurko's notebook but need to
add relevant tests first.
* Test.
* _SoapClient subclasses modifying _init_invocation_context(),
_get_request() & _get_reply().
* Update release notes.
* Client debug log messages changed.
* Removed the RequestContext.original_envelope attribute.
* (Jurko) See why creating two suds.client.Client instances with the same
transport instance fails. It reports some 'Duplicate domain "suds.options"
found' error.
t = suds.transport.Transport()
tests.client_from_wsdl(tests.wsdl(""), transport=t)
tests.client_from_wsdl(tests.wsdl(""), transport=t)
* (Jurko) Generic code cleanup.
* HttpTransport tests.
* Check for multiple build_opener() calls.
* Check for multiple urlopener.open() calls.
* Check what gets passed to the default built urlopener. Merge this
with externally specified urlopener testing.
* Add HttpTransport proxy configuration tests.
* See when the proxy member is set.
* See what the default proxy member value is for.
* See whether the proxy member is used at all with an externally
specified urlopener, and if so - if it is made redundant by using
such externally specified urlopeners.
* Add cachingpolicy option tests.
* Add default document store tests.
* Generic.
* suds.client.Client document store usage.
* See why exception raised when passing a transport to
suds.client.Client that is not a subclass of suds.transport.Transport
is formatted as it is.
* Class's string representation wrapped in a tuple.
* Uses class's repr() string representation.
* Double quotes used around the 'transport'.
* suds.store.defaultDocumentStore cleanup.
* See if this member should be marked as private.
* Rename to not use camelCase.
* DefinitionsReader.open() changes options on cached Definitions
instances. Make sure this does not break programs with multiple suds
Clients loading the same WSDL schema but using different options.
* Support for calling NoCache.purge() & NoCache.clear().
* Consider making FileCache.clear() remove files matching both prefix &
suffix instead of only the prefix.
* Might be ok for explicitly called clears, but not for those called
due to the cache version information getting changed. This might
need to be worked around or commented/tested in code.
* Make FileCache, DocumentCache & ObjectCache all consistently remove or
not remove cache entries on failed access. Possibly do so for the base
Cache class as well.
* Make Cache.put() not return the cached data.
* FileCache.put() operations should not leave behind invalid files in
case of failed write operations. Possibly first store a valid file
under a temporary name and then rename it.
* (Jurko) Restore last_sent()/last_received() getters for last used SOAP
request/response messages.
* Research.
* Difference between how _SoapClient & _SimClient stored their last
sent/received messages.
* Both stored them.
* _SimClient stored the simulated fault as its last received
message.
* Whether sent/received messages should be returned/logged before or
after being processed by registered plugins.
* _SoapClient stored the received reply after it got processed
by registered plugins.
* Sending with _SoapClient.
* Sending with _SimClient.
* Which plugins get called and when.
* Add back the removed code.
* Add tests.
* Notify people interested in this.
* (Jurko) Clean up input argument/parameter & wrapped/unwrapped terminology
usage in code.
* Synchronize all usage both in code and in relevant 'notes/' documents.
* (Jurko) Triage received issues & pull requests at BitBucket a bit and
prioritize them in relation to already existing todo items on this list.
* (Jurko) Process Jurko's remaining research & todo notes collected while
working on reporting extra parameter errors. Some are redundant or have
already been dealt with so the list needs to be triaged first.
* Fix non-optional choice element handling - missing values should be
used as empty strings, same as for non-choice parameters - search for
tests xfailed with reason 'non-optional choice handling buggy'.
* Add tests: which type XSD schemas do not get unwrapped automatically,
e.g. simple types such as built-in type restrictions.
* Split up test_request_construction.py tests into parameter definition,
argument parsing & possibly partial binding specific request
construction tests.
* Add test: single parameter pointing to an element - parameter name
should be ignored.
* Add test: single parameter pointing to an element - with unwrapping.
* Add test: single parameter pointing to a type - with unwrapping.
* Add test: single parameter pointing to a built-in type - no
unwrapping.
* Add test: multiple parameters pointing to types & elements.
* Add tests: WSDL tests - XSD schema content.
* Add tests: WSDL tests - recognized operations.
* Add tests: WSDL tests - recognized services.
* Add test: document/literal, document/encoded, rpc/literal &
rpc/encoded should make no difference to input processing in theory
and should all recognize the same input parameter structure.
* Add tests: WSDL tests - recognized binding - input/output - for
different methods - document binding should be the default - see
whether literal should be the default binding style.
* Add test: named sequence/choice/all - there is no such thing - their
name attributes should be ignored.
* Research: consider input parameter elements with missing type
declarations to have the type "xsd:string" or "xsd:any" or something
similar - check the XSD specification.
* '<xsd:element name="MyElement"></xsd:element>' currently used
wrapped.
* '<xsd:element name="MyElement"> </xsd:element>' currently used
unwrapped.
* Add test: unwrapped element reference input parameter.
* Add test: shema_object.resolve() &
shema_object.resolve(nobuiltin=True).
* Add test: explicitly marking an input parameter sequence as optional
should be recognized when unwrapping those input parameters - such
functions should report accepting 0 parameters.
* Add test: expected input argument count reporting with choice input
parameters explicitly marked optional.
* Add test: array parameters.
* Research: whether web service operations taking multiple input
parameter body-parts can have those body-parts simply reference other
elements, thus requiring us to resolve that reference before checking
whether the input parameter is optional (and possibly other places as
well - see Document.mkparam() which does no resolving for such web
service operations but does so for operations taking a single wrapper
input parameter structure).
* Research: see if the parameter type checked whether it is optional
inside Document.bodycontent() needs to have its ancestor elements
checked as well - this might be needed if this can actually be a
reference to some other, more deeply nested element.
* Research: can element references be chained - nope as they are allowed
to reference only top level elements and top level elements are
explicitly not allowed to use the ref attribute.
* Research: can element references be optional by themselves - yes, and
that is the only way they can be marked as optional since they may
only reference top-level elements and those are in turn not allowed to
use the minOccurs/maxOccurs attributes.
* Add test: extra array parameters - more than maxOccurs - possibly not
necessary if we want to allow suds to make such 'illegal' calls.
* Add test: missing regular parameters.
* Add test: missing non-optional choice parameters.
* Add test: missing optional choice parameters.
* Recognize missing optional values when specified as a lists/tuples
containing only values None and other such lists/tuples.
* Error when passed a list/tuple argument containing a list/tuple or
None.
* Research: web service operations taking multiple same-named parameters
- how they are and how they should be handled.
* Choice containing an empty sequence should be recognized as optional.
* Choice(a, sequence(b, c)) should set b to '' if it is not optional and
no value is given for it, but a value is given for c - same if b & c
roles are reversed.
* Research: suds does not seem to know how to handle minOccurs/maxOccurs
attributes specified on sequence & choice XSD schema XML elements - it
definitely does not recognize elements inside an optional choice or
sequence as optional - see how this affects input/output parameter
unwrapping.
* Add test: passing an array element value as an empty list.
* Add test: passing an array element value as a non-empty list.
* Add test: constructing requests with missing optional parameters -
should not be specified in the request.
* Add test: constructing requests with missing non-optional parameters
when the operation also takes some optional parameters - how this is
reported.
* Add test: constructing requests by using positional arguments when the
operation takes first optional input parameters and then non-optional
ones after that - initial positional arguments should be mapped to the
initial optional input parameters.
* Research: constructing requests with missing default parameters - see
whether this should be reported as an error or added to the request as
an empty value.
* Research: self.mkparam() call in Document.bodycontent() may return a
list in which case later setPrefix() method call on the returned
object will raise an AttributeError exception.
* Detecting and reporting multiple input parameter values specified for
a single input parameter choice group (empty choice group, simple
single choice, multiple independent choices, one choice group inside
another, one choice inside another but with additional elements
between them).
* Research: whether specifying an empty list or tuple for a parameter
should be treated the same as None - if so, add tests - possibly []/
(,) really skips the parameter while None gets mapped to an empty
value for non-optional parameters..
* Support for multiple same-named elements in a single input or output
structure - those directly back-to-back with each other can already be
set or accessed as arrays but we could need to support others as well.
* Constructing a request for a web service operation taking a single
empty all/choice/sequence parameter - should be the same as for
operations taking no parameters.
* Constructing requests for web service operations taking multiple
choice parameter groups (as positional and keyword arguments).
* Using a function with multiple choice parameter group values.
* TestExtraParameters.test_function_with_no_parameters() - why does
using WSDL '<xsd:element name="Wrapper" />' not work - seems not to be
recognized as wrapped data but if the XML element has any data (e.g. a
space or a new-line) then it gets recognized correctly - smells like
this could be a bug.
* Web service operation taking multiple input parameter bodyparts.
* None of which is an empty sequence.
* At least one of which is an empty sequence.
* Unrelated issues noticed while working on this.
* Document.document() - returned temporary can be inlined.
* Core.process() - root local variable seems redundant.
* Document.bodycontent() - mkparam() does not seem to be able to
return None but if it does, it might mess up tracking which
argument got used and we should document when it does that.
* Parsing parameter values (interpreting positional & keyword
arguments, mapping them to internal web service operation
parameters, reporting duplicate parameter values, reporting
unrecognized parameters and skipping ignored or None choice item
parameters) should not be binding class specific, i.e. it should
be shared for both document & rpc bindings.
* Code cleanup
* It seems suds ignores top-level element names & namespaces when
parsing server responses and instead just uses them by their index.
This seems wrong. Check the WSDL specification, SOAP message binding
and other related standards to see if the server must always return
its values in the correct order in which the corresponding output
message parts are defined in the WSDL. Plan updating suds to report
invalid server responses.
* Replace except: blocks with blocks catching specific exception types.
* suds.xsd.schema module imports suds.xsd.sxbuiltin.Factory as
BuiltinFactory but never used that reference. See if this interface is
perhaps documented as published and if not - remove it.
* Performance tuning - SchemaObject.str() should not use len(self) to
check whether it is empty.
* SchemaObject.str() - see whether any contained item information is
included if none of the items are instances of Content.
* TypedContent.qref() - see whether len(self) == 0 check for 'simple
types' can be improved. There should be no need to calculate the whole
length just to determine that the object is empty.
* Simple.mixed(), SimpleContent.mixed() - see if they really need to
return the current object length or if they can just check whether the
object is empty.
* suds.sudsobject.footprint() - see if the len() check is really needed
just to see if the object is empty.
* Check other len() usage in suds.sudsobject.
* Check all 'except:' blocks to see if they actually need to be such
catch-all blocks.
* (Jurko) Look into issues #7 & #13 on BitBucket - problems loading a WSDL
containing a recursively defined XML schema.
* Possibly already fixed by related work done around 01.07.2015.
NON PRIORITIZED:
=================================================
* Find and use a good spy/mock framework for our tests instead of rolling
our our own every time we need one.
* Add a functional test parsing a WSDL and running all the way down to the
Typer.genprefix() function.
* Useful to help see what else got missed in the test suite when the
suite missed this function and let through a bug reported as
https://bitbucket.org/jurko/suds/issues/90.
* Improve Typer.genprefix() implementation.
* See whether, when we have the same namespace already mapped, we should
return its existing prefix here.
* Current implementation already does that but only in some cases -
if ns1, ns2, ... ns# are # all already mapped to a different
namespace and ns#+1 is mapped to the one we want, it returns
ns#+1, making it seem like this was tacked on only as an
afterthought.
* suds.sudsobject.footprint() does some fancy recursive counting while all
it is actually used for is for detecting objects with footprint 0. This
could be optimized by replacing it with a function returning simply
whether a given object has a non-empty footprint and returning as soon as
it find something of value in it.
* Add suds.sax.element related tests.
* append().
* Add child element.
* Add child attribute.
* childAtPath().
* With namespaces.
* With prefixes.
* Empty child name - as initial child - requested.
* Empty child name - as initial child - pass through.
* Empty child name - as indirect child - requested.
* Empty child name - as indirect child - pass through.
* Handling multiple children with matching names - is there a way to
specify an index.
* Add child by constructing it with a parent argument - should
automatically add the element to the parent's children list, while the
current implementation only makes the child know about the parent and
does not let the parent know about the child.
* Constructing an element with '/' in its name should not be allowed
since that character is used as a part separator.
* See if a way to escape such character usage is needed and if so -
how to do it.
* Handling multiple children with matching names.
* childrenAtPath().
* Initializing a new element with a specific namespace.
* Add suds.sax.document unit tests similar to those already added for
suds.sax.element.
* Some ideas on what to not forget.
* See suds.sax.element related tests & todo notes.
* childAtPath().
* childrenAtPath().
* No root element.
* Look into CDATA related encoding issues reported at BitBucket (issue #14 +
pull requests #25 & #26).
* Current status: waiting for further feedback or progress on BitBucket.
* Implement a working solution satisfying the prepared tests.
* Use the prepared profiling script to make sure this change does
not degrade general suds performance.
* Get more feedback on a seemingly bad 'embedded CDATA section' test
included in the pull request #26 on BitBucket.
* Support for nested CDATA sections, in violation of the CDATA
related documentation found at
'http://en.wikipedia.org/wiki/CDATA#Nesting' and
'http://www.w3.org/TR/REC-xml/#sec-cdata-sect'.
* Decide how unclosed CDATA sections should be handled.
* Suggestion: raise an exception since this is not allowed in well
formed XML documents.
* Resolving XSD schema elements currently returns an XSD schema type object
representing the element's type. Different elements with matching types
should possibly resolve to the same XSD schema type object but currently
do not. See if this behaviour should be corrected.
* Example.
* XSD schema:
<xsd:element name="wi" type="xsd:string"/>
<xsd:element name="wo" type="xsd:string"/>
* Test code:
schema = client.wsdl.schema
element_in = schema.elements["wi", "my-namespace"]
element_out = schema.elements["wo", "my-namespace"]
assert element_in.resolve() is element_out.resolve()
* Both elements should possibly resolve to the same XString
instance, but currently they resolve to two separate ones.
* Optionally make suds check that the input parameters passed to its web
service proxy operations actually match their respective WSDL definitions.
* Current suds behaviour.
* Extra parameters are ignored.
* Parameters of invalid type are simply added into the generated
SOAP request as strings and left up to the web service
implementation to deal with and report as an error if the
constructed SOAP request turns out to be invalid.
* Having a local client-side check could make catching client side
programming bugs easier.
* Ideas.
* Make sure given values fit their respective type value domains,
e.g. integers, strings, regular expressions, restrictions, complex
types where a builtin was expected, complex type of an unexpected
structure, etc.
* Extra parameters are reported as errors.
* Missing non-optional parameters are reported as errors.
* Input message part has both element & type specified.
* See how invalid schemas containing a ref-cycle are handled.
* They should be reported as invalid either when dereferencing them
(e.g. to determine an element's target namespace) or when building the
internal schema object tree and should not cause us to go into endless
recursion.
* See how multi-occurrence input parameter elements are supposed to be
supported.
* With automated parameter unwrapping support.
* Without automated parameter unwrapping support.
* Clean up & correct the choice support implementation.
* Choice parameters seem to be supported only for document/literal style
input parameters only.
* Add tests for this.
* Fix.
* See the currently disabled
'xxxtest_choice_parameter_implementation_inconsistencies' unit test
demonstrating a problem with the current implementation.
* _SoapClient 'location' cleanup.
* URL quoting, especially if specified externally by the caller instead
of having been read from a valid WSDL schema.
* See how input parameter element for a document/literal web service
operation style gets handled when that element has 2 occurrences.
* See how this affects suds library's automated input data unwrapping.
* See how this sort of data is supposed to be entered anyway.
* FileCache class fixes.
* Remove incorrectly created cache files, e.g. if one gets created but
then writing to it fails.
* Make sure reading a cache file does not crash the whole Python
interpreter process.
* Encountered when trying to read an empty cache file using a debug
Python version 3.2.3 on Windows 7 - triggered an internal Python
assertion failure.
* Look into the 'suds.null' class.
* Research.
* See sources (grep for 'null'), old release notes (README.txt) and
commit messages.
* What it is for.
* Whether it can be replaced with None.
* If it serves some purpose see if it should be used for identifying
missing 'choice' structure members as well.
* Make it simpler to run Py3 tests.
* Current status.
* May be run without additional pytest options by running 'setup.py
test' from the root project folder.
* This method will automatically download & install pytest from
PyPI if needed, but will leave their installations behind in
the project's root folder.
* May be run by first building the project by installing the
project, e.g. by running 'setup.py develop' and then running
'pytest' from the project's 'tests' folder.
* Do some more thinking on this and, when done, update related HACKING
notes.
* Generate suds Python library documentation (epydoc).
* Research.
* HTML.
* PDF.
* Decide how to maintain, generate & distribute this documentation.
* Update project HACKING notes to note the external software required
for generating the documentation.
* Update release procedure to include releasing the documentation.
* Once the official project documentation has been integrated into this
fork, go through existing internal project notes in our fork and see
how to integrate them into the official project documentation.
* Research.
* Test how optional elements under a choice work.
* There are some comments & an additional patch related to this at
'https://fedorahosted.org/suds/ticket/342'.
* Default element values.
* What they actually mean.
* From from 'http://www.w3.org/TR/xmlschema-0'.
* If the element does not appear it is not provided; if it
does appear and it is empty, its value is the specified
default value; otherwise its value is that given.
* How elements with default values inside a choice structure should
be handled.
* See what the suds.sudsobjects.Facade class is for.
* How to implement test cases requiring a test web service.
* See how to connect to a web service via a Proxy server requiring NTLM
authentication.
* There are some projects seen on the net implementing a NTLM
authentication handler for urllib.
* Testing this will require implementing a proxy server requiring
NTLM authentication and a web service or at least a web service
requiring NTLM authentication.
* Using pylint.
* Implement an urllib connection handler allowing connecting using HTTPS
with client authentication.
* Prepare a test (will require a test web service).
* Implement.
* Prepare additional test cases.
* Loading a WSDL from a file as in suds.client.Client(
"file://localhost/Folder/aaa.wsdl").
* Path not containing spaces.
* Path containing spaces.
* Invalid element reference.
* Handling schema imports.
* With the same target namespace.
* With a different target namespace.
* Prepared SOAP operation invocation requests.
* With choice parameters.
* Calling a web service operation with no parameters.
* RPC binding style.
* Document binding style.
* Process ideas collected from external projects using suds.
* Alternative choice implementation that would not automatically expand
all choice function parameters but instead take some more generic
'choice' object parameter.
* This object would then know which of its data members is
'currently specified'.
* See whether the Marshaller class should know about choice elements and
not add XML nodes for elements with the value 'None' if they are
contained directly inside a choice.
* For some more background information on this see Marshaller
related release notes & commit messages inherited from the
original 'suds' development project.
* See if the class 'suds.xsd.sxbasic.Complex' function sequence() should
return True if its only child is a sequence node.
* Typo corrections for the original project web site.
* 'docuemnt' --> 'document'.
* 'becuase' --> 'because'.
* See whether standard and/or suds supports specifying message part
parameter element directly from an external schema.
* For example, 'integer' or 'string' elements defined in the
'http://schemas.xmlsoap.org/soap/encoding/' namespace (stored
internally by suds in the store.py module).
* See whether importing an external WSDL schema can benefit from that
schema's WSDL object already being cached.
* See if this has already been implemented, and if not - implement it.
* Additional test ideas.
* External XSD usage (including both imported and included external XSD
schemas).
* All the different external XSD usage tests should be run in
scenarios when the external XSD gets used from an imported WSDL.
* Using an external XSD schema using another external XSD schema.
* Test pluggable sax character reference encoder support.
* Have tests plug in a custom character reference encoder and try it
out.
* A HttpTransport.send() operation will return None if the web service
server responds with HTTP status code ACCEPTED (202) or NO_CONTENT (204),
but the _SoapClient method using this operation is not prepared to handle
this and will attempt to access the 'message' member on the returned
value.
* Failing tests have already been added for this.
* Alternative ideas on how to fix this.
* Continue returning None in these cases and fix up the calling
code.
* Raise a separate exception.
* Wrap this information in a regular suds.transport.Reply instance.
* Possibly handle this differently based based on whether the web
service operation invoked has any defined output message parts.
* 'soapheaders' suds option usage cleanup.
* Add tests.
* SOAP headers given as a list/tuple.
* Strings.
* Elements.
* Initial.
* Middle.
* Last.
* Multiple.
* SOAP headers given as a dictionary.
* Strings.
* Elements - error.
* Unknown header handling.
* Too many when given as list/tuple.
* Unknown key when given as dictionary.
* Plugins modifying SOAP header elements should not affect future
requests constructed using those same headers.
* Note that Elements found in the soapheaders option need to be copied
into each request so any plugins updating that request's content would
not update the original SOAP header elements and thus affect all
future requests constructed using those same SOAP headers.
* Fix bug: Elements added to a soapheaders list before all expected
headers have been added, but later ones are ignored.
* Raise an exception in case of an unrecognized header.
* Unrecognized header name in a dictionary.
* Too many non-Element headers in a list/tuple.
* Better document the 'soapheaders' suds option.
* At first glance it seems suds does not use reply headers the same as
it does its request headers. See if SOAP web service HTTP bindings
allow output headers at all, and if so, how suds should support this.
* See if header elements can be marked as required, and if so, see if
suds should report an error if a required header value has not been
provided.
* See if header elements can be marked as optional and how suds should
handle this.
* See if header elements can have multiple occurrences, and if so, how
suds should support them.
* See how header values can be specified more easily for a single method
invocation.
* See if input message parts defined using the 'type' attribute (as opposed
to those marked using the 'element' attribute) respect their optional/
required status. PartElement class used to model such part's element
information seems to always return True from its optional() method.
* If this is wrong - add a test demonstrating the problem.
* If this is correct - see what the suspicious optional() method is for.
* Look at suds.sax.element.Element.insert() - its documentation says it
accepts either an element or an attribute or a collection of elements or
attributes and adds it/them as a child/children under the current element.
* From glancing through the code it does not seem to work when passed a
collection of elements or attributes.
* Most likely it needs a check whether the passed objects parameter
is a list or a tuple before it wraps it inside a tuple.
* From glancing through the code it does not seems to work when passed
an Attribute.
* See what to do about this - either update the related function
documentation or update it to work as documented.
* See how similar functionality has been implemented in
suds.sax.element.Element.append().
* Look at suds.sax.element.Element.childAtPath() and see whether it should
return None or self if given an empty string or a string consisting only
of slashes.
* Currently it returns None in this case as if the given path could not
be found.
* Returning self or raising an exception seems consistent with the fact
that a non-empty path is interpreted relative to self.
* suds.sax.element.Element.__childrenAtPath() seems smelly. See how it can
be simplified and made to work for paths not containing slashes as well as
those containing slashes.
* This should simplify the suds.sax.element.Element.childrenAtPath()
implementation as well.
* suds.sax parser peculiarities to look into.
* Handling mixed content elements, i.e. those that have both child
elements and textual content.
* All the textual content is concatenated into a single string,
independent of how it distributed between child elements in the
original parsed XML string.
* As a consequence, converting the parsed XML document back into
an XML string will place all the element's textual content
before its child elements.
* The textual content has its whitespace trimmed.
* Note that this causes the '<a> 42 <b/> <c/> 42 </a>' XML to
be parsed so that <a> element's textual data is: '42 42'
where the whitespace outside the '42' markers is trimmed but
the whitespace between them is preserved.
* See if we should make sure any plugin exceptions get logged correctly (as
was the case when suds.plugin.Method.__call__() was eating up all such
exception in the original suds project).
* The original suds project's DocumentCache.put() implementation accepted
suds.sax.element.Element instances as input so we kept this behaviour for
now but that might have actually been a result of a bug. We should review
this and update related tests & implementation if we no longer wish to
support this.
* See the TestDocumentCache.test_cache_element() test in
'test_sax_element.py'.
* Improve the XML comparison test utilities.
* Add test utility tests.
* Make SAX comparison compare XML attributes - order is irrelevant.
* Support for comparing two XML data strings in more detail, with
explicitly enabling/disabling which differences to consider important.
* Some ideas on what might or might not be considered important.
* Leading/trailing textual data whitespace.
* Per element.
* Per block between child elements.
* Per input line.
* All whitespace blocks considered equivalent.
* Possibly consider leading/trailing textual data whitespace
unimportant only inside mixed content nodes containing at
least one child element.
* How a mixed content node's textual content is split between
its child elements.
* Attribute declaration order.
* Child element ordering.
* Namespace declarations (no prefix).
* Namespace prefix declarations.
* External references - to parse them or not.
* Extra elements on either side.
* XML node namespace prefixes.
* XML node namespaces.
* Input code-page.
* Extra data.
* Before XML.
* Surrounding top level XML nodes.
* Each feature importance might be enabled/disabled globally or
for a specific XML part only.
* Add SOAP 1.2 support.
* Related links.
* Project's issue #50 on the BitBucket issue tracker:
https://bitbucket.org/jurko/suds/issue/50#comment-10940316
* http://stackoverflow.com/questions/2370573/soap-1-2-python-client
* Add support for the `mixed` attribute on `complexType` XSD schema
elements.
* They should allow the use of textual data, possibly in addition to
other child elements & attributes.
* Related link:
https://bitbucket.org/jurko/suds/issue/51#comment-10941116
View Git Blame Copy Permalink