* testapi.c tree.c: fixing a leak detected by testapi in
xmlDOMWrapAdoptNode, and fixing another side effect in testapi
seems to pass tests fine now.
* include/libxml/parser.h parser.c: xmlStopParser() is no more limited
to push mode
* error.c: remove a warning
* runtest.c xmllint.c: avoid compilation errors if only some parts
of the library are compiled in.
Daniel
* xmlschemas.c: Fixed NULL values to be validated as empty
string values (reported by Dhyanesh to the mailing list).
Adjusted text concatenation of mixed content.
* xmlschemas.c: Fixed SAX2 validation: grow of internal
namespace list, appending of element character content.
* xstc/xstc.py: Added "--sax" option for SAX2 validation.
* xmllint.c: if sax1 is used and input is a file use the old
API xmlParseFile()
* xmlschemas.c: cleanups
* doc/* testapi.c elfgcchack.h: rebuilt to add python bindings
for the new functions in Schemas.
Daniel
* pattern.c: Some changes/fixes to the streaming evaluation.
* xmlschemas.c: A bit of support for parsing the schema for
schema. Fixed attribute derivation when the use is
"prohibited" and was "optional". Fixed an attribute construction
bug, a left-over from the time, where <complexContent>,
<extension>, etc. where created as structs.
* xmlschemas.c xmlschemastypes.c include/libxml/schemasInternals.h
include/libxml/xmlschemastypes.h: Changed the validation process
to be able to work in streaming mode. Some datatype fixes,
especially for list and union types. Due to the changes the
error report output has changed in most cases. Initial migration to
functions usable by both, the parser and the validator. This should
ease a yet-to-come XS construction API in the long term as well.
* xmlschemas.c: Fixed facet errors to be channelled back for
union type members; facet-validation will stop now on the
first error. Reported by GUY Fabrice to the mailing-list.
* xmlschemastypes.c: Changed to ignore lengh-related facet
validation for QNames and NOTATIONs as proposed by the
schema people.
* test/schemas/union2* result/schemas/union2*: Added
regression tests for union types (by GUY Fabrice).
* xmlschemas.c: Removed workaround for bug #172215, since it
does more harm in some scenarios. Added some of the
"Particle Restriction OK" constraints - not yet enabled.
* xmlschemas.c: Added output of canonical values in
identity-constraint error messages.
* xmlschemastypes.c include/libxml/xmlschemastypes.h:
Added xmlSchemaGetCanonValueWhtsp() to the API.
Further enhancement of the canonical value
conversion.
* test/schemas/changelog093_0.*: Added test with an XSD
submitted by Randy J. Ray.
* xmlschemas.c: Added substitution group constraints; changed
the build of the pre-computed substitution groups. Channeled
errors during xsi assembling of schemas to the validation
context. Fixed a big memory leak, which occured when using
IDCs: the precomputed value of attributes was not freed if
the attribute did not resolve to an IDC field (discovered
with the help of Randy J. Ray's schema, posted to the
xmlschema-dev maling list).
* xmlschemas.c: Added "Particle correct 2" to parsing of model groups.
Added handling substitution groups inside <choice> and <sequence>;
for <all> this is not supported yet. Changed circular checks for
model groups definitions. "memberTypes" are processed at different
levels now: component resolution first, construction later; this
goes hand in hand with a global change to handle component
resolution in a distinct phase. Fixed invalid default values for
elements to mark the schema as invalid; this just resulted in an
error report previously, but the schema was handled as valid.
Separated the assignment of the model groups to referencing
model group definition references (i.e. particles); this was
needed to perform the circularity check for model group definitions.
Added "Element Declaration Properties Correct (e-props-correct)"
constraints. Separated component resolution for simple/complex
types.
* include/libxml/schemasInternals.h: Added a flag for substitution
group heads.
* xmlschemas.c: Added "Type Derivation OK (Complex)" constraints
and anchored them in the "Element Locally Valid (Element)"
constraints. This restricts the type substitution via "xsi:type".
* xmlschemastypes.c: Corrected 'length' facet validation for
QNames and notations. Corrected xmlSchemaGetCanonValue: some
data types did not return a value if already in normalized
form.
* xmlschemas.c include/libxml/schemasInternals.h:
Eliminated creation of structs for <restriction>, <extension>,
<simpleContent>, <complexContent>, <list> and <union>: the
information is now set directly on the corresponding simple/
complex type. Added some more complex type constraints.
Added facet derivation constraints. Introduced "particle"
components, which seem to be really needed if applying
constraints. Corrected/change some of the parsing functions.
This is all a bit scary, since a significant change to the code.
* result/schemas/src-element2-*.err result/schemas/element-*.err:
Adapted regression test results.
* doc/apibuild.py doc/elfgcchack.xsl: revamped the elfgcchack.h
format to cope with gcc4 change of aliasing allowed scopes, had
to add extra informations to doc/libxml2-api.xml to separate
the header from the c module source.
* *.c: updated all c library files to add a #define bottom_xxx
and reimport elfgcchack.h thereafter, and a bit of cleanups.
* doc//* testapi.c: regenerated when rebuilding the API
Daniel
* python/generator.py python/libxml.py: another patch from Brent
Hendricks to add new handlers with the various validity contexts
* python/tests/Makefile.am python/tests/validDTD.py
python/tests/validRNG.py python/tests/validSchemas.py: also
added the regression tests he provided
Daniel
* xmlschemas.c: Removed a stupid bug in xmlSchemaValidateAttributes,
failing to build a linked list correctly (bug #169898, reported
by bing song, hmm...).
* xmlschemas.c: Fixed a segfault, which occured during bubbling
of IDC nodes (bug #170779 and #170778, reported by GUY Fabrice):
a variable was missed to be reset in a loop. Deactivated bubbling,
if not referenced by a keyref.
* xmlschemas.c: Removed inheritance of "mixed" content type for
short-hand restrictions of "anyType" (reported by Guy Fabrice
to the mailing list). Added the namespace conversion (chameleon
includes) for the base type reference of <restriction> and
<extension>.
* test/schemas/bug152470_1.xsd: Adapted due to the above change
of "mixed" inheritance.
* xmlschemas.c xmlschemastypes.c include/libxml/xmlschemastypes.h:
Hopefully finished validation against facets to use the normalized
value of both, the facets and instance values. Added
xmlSchemaValidateLengthFacetWhtsp(), xmlSchemaValidateFacetWhtsp()
and xmlSchemaGetValType() to the schema API.
* encoding.c: removed a static buffer in xmlByteConsumed(),
as pointed by Ben Maurer, fixes#170086
* xmlschemas.c: remove a potentially uninitialized pointer warning
Daniel
* xmlschemas.c: The schema parser will stop if components could
not be resolved. This is not conforming to the spec but for now
will avoid internal errors during type fixup and content model
creation. Restructured inclusion/import of schemata: this avoids
duplicate, self and circular inclusion. Chameleon includes are
still workarounded. Added restriction to disallow references to
non-imported namespaces. Corrected parsing of <group>.
* result/schemas/bug167754_0_0*: Added a missing test result.
* xmlschemas.c: Tiny restructuring of the validation start-up
functions. Added cleanup of the validation context at the
end of validation. This takes care of the validation context
being reused.
* xmlschemas.c: Added IDC validation of anySimpleType attribute
values; anyway the IDC key comparison is restricted to
anySimpleType <--> string based types. For other types we
would possibly need the canonical lexical representation of
them; this sounds not performant, since we would need to
build such a representation each time we want to compare against
anySimpleType. TODO: think about buffering the canonical values
somewhere. Fixed error reports for default attributes to work
without a node being specified. This all and the enabling of IDC
validation fixes bug #165346 (reported by Benoit Gr?goire - could
not read his last name correctly from bugzilla).
* xmlschemas.c: Enabled IDC parsing and validation.
* xmlschemastypes.c include/libxml/xmlschemastypes.h:
Added xmlSchemaCopyValue to the API; this was done due to
validation of default attributes against IDCs: since IDC keys
consume the precomputed value, one needs a copy.
* pattern.c: Enabled IDC support; this is currently done
via calling xmlPatterncompile with a flag arg of 1.
* xmlschemas.c: Started support for IDC resolution to default
attributes. If building the content model for <all>: ensured
to put element declarations and not the particles into the
content model automaton (this was bug #167754, reported by
Frans Englich).
* xmlschemas.c: Added validation for the attribute "id" in the
schemata; doing this needed error report fixes for notations,
facets and group. Changed NOTATION validation to work with the
declared NOTATIONs in the schema; this does have no impact on
the validation via the relaxng module.
* xmlschemastypes.c include/libxml/xmlschemastypes.h:
Added xmlSchemaNewNOTATIONValue to the API to be able to do
the NOTATION validation described above.
* test/schemas/element-err_0.xsd test/schemas/annot-err_0.xsd:
Fixed the values of the "id" attributes, which were not validated
previously.
* xmlschemas.c: Fixed comparison for default/fixed attribute
values, if the type was 'xsd:string'. Changed the comparison
for IDCs to use the whitespace aware comparison function.
* xmlschemastypes.c include/libxml/xmlschemastypes.h:
Added xmlSchemaGetCanonValue, xmlSchemaNewStringValue and
xmlSchemaCompareValuesWhtsp to the API. Added functions
to compare strings with whitespace combinations of "preserve",
"replace" and "collapse".
* xmlschemas.c: Added IDC evaluation for attribute nodes.
Made 'nil'ed elements work. Added a specific error message
for 'strict' attribute wildcards.
* include/libxml/xmlerror.h: Added an error code for
wildcards.
* result/schemas/anyAttr-processContents-err1_0_0.err: Adapted.
* xmlschemas.c: fixed one internal function
* doc/Makefile.am doc/wiki.xsl: applied patch from Joel Reed
* testapi.c doc/libxml2-api.xml doc/libxml2-refs.xml: regenerated
daniel
* xmlschemas.c: Integrated the streaming pattern from the
pattern module. Fixed some IDC code bugs. Changed
fallback for attribute declaration addition to work like for
element declarations.
* xmlschemas.c: Corrected an ambigious symbol-space for
local attribute declarations. IFDEFed more IDC code to
surpress compiler warnings.
* xmlschemas.c: tiny enhancement for content model error reports
(#157190, #143948). Removed abbreviations: CT, ST and WC
(#157190, reported by Frans Englich).
* result/schemas/all* result/schemas/any3_0_0.err
result/schemas/choice*
result/schemas/cos-st-restricts-1-2-err_0_0.err
result/schemas/derivation-ok-extension-err_0_0.err
result/schemas/derivation-ok-extension_0_0.err
result/schemas/derivation-ok-restriction-2-1-1_0_0.err
result/schemas/derivation-ok-restriction-4-1-err_0_0.err
result/schemas/deter0_0_0.err result/schemas/extension1_0_2.err
result/schemas/facet-unionST-err1_0_0.err
result/schemas/hexbinary_0_1.err
result/schemas/list* result/schemas/restriction-attr1_0_0.err
result/schemas/vdv-first4_0_1.err result/schemas/vdv-first4_0_2.err:
Adapted output.
xmlSchemaTypeFixup; this ensures facets of inherited types to be
checked prior to facets of derived types - which caused a seg
fault otherwise (bug #158216, reported by Frans Englich).
* xmlschemas.c include/libxml/schemasInternals.h:
Added an initial skeleton for indentity-constraints. This is all
defined out, since not complete, plus it needs support from other
modules.
Added machanism to store element information for the
ancestor-or-self axis; this is needed for identity-constraints
and should be helpfull for a future streamable validation.
* include/libxml/xmlerror.h: Added an error code for
identity-constraints.
* include/libxml/schemasInternals.h xmlschemas.c:
Exposed targetNamespace for simple/complex types, model groups,
attribute groups and notations (reported by Michael Hewarth
to the mailing list). Added targetNamespace to xmlSchemaType,
xmlSchemaAttributeGroup and xmlSchemaNotation.
Tiny cosmetic change to the content model error report output.
* result//all_*.err result//any3_0_0.err result//choice_*.err
result//list0_0_1.err result//list0_1_1.err: Adapted output
of regression tests.
* xmlschemas.c: Integrated xmlRegExecErrInfo and xmlRegExecNextValues
from xmlregexp.c to report expected elements on content model errors.
* all_*.err any3_0_0.err choice_*.err list0_0_1.err list0_1_1.err:
Adapted output of regression tests.
Synchronized the header files with the library code in order
to assure that all the various conditionals (LIBXML_xxxx_ENABLED)
were the same in both. Modified the API database content to more
accurately reflect the conditionals. Enhanced the generation
of that database. Although there was no substantial change to
any of the library code's logic, a large number of files were
modified to achieve the above, and the configuration script
was enhanced to do some automatic enabling of features (e.g.
--with-xinclude forces --with-xpath). Additionally, all the format
errors discovered by apibuild.py were corrected.
* configure.in: enhanced cross-checking of options
* doc/apibuild.py, doc/elfgcchack.xsl, doc/libxml2-refs.xml,
doc/libxml2-api.xml, gentest.py: changed the usage of the
<cond> element in module descriptions
* elfgcchack.h, testapi.c: regenerated with proper conditionals
* HTMLparser.c, SAX.c, globals.c, tree.c, xmlschemas.c, xpath.c,
testSAX.c: cleaned up conditionals
* include/libxml/[SAX.h, SAX2.h, debugXML.h, encoding.h, entities.h,
hash.h, parser.h, parserInternals.h, schemasInternals.h, tree.h,
valid.h, xlink.h, xmlIO.h, xmlautomata.h, xmlreader.h, xpath.h]:
synchronized the conditionals with the corresponding module code
* doc/examples/tree2.c, doc/examples/xpath1.c, doc/examples/xpath2.c:
added additional conditions required for compilation
* doc/*.html, doc/html/*.html: rebuilt the docs
* xmlschemas.c: Fixed strict/lax element wildcards: the children
of elements for which a declaration existed were still processed
by the wildcard mechanism (reported by philippe ventrillon to the
mailing list).
Changed the import and include machanism to share dictionaries.
