CVS log for xmlschema/applyschema.py

Request diff between arbitrary revisions

add schemaSpecified PSVI property,
change to nillable,
move all booleans to booleans

separate in and out maps makes refl switching easier

get namespace vars sorted out

finish separation of reflection code from main modules

restructure in preparation for allowing either normal form to be reflected

conservative about min/max types

accommodate to splitting of XMLSchema.py

accommodate to split in uid assignment,
chunkedChildren->children

accommodate changes to facet storage; implement facets on list and
union

default addrs needed a validity context

support -i -r, dump only infoset

fix reporting of builtin-root-type

typo

Remove decl/type args from validate
Change default for nullable when decl is absent to true
Use 0 for decl to indicate that decl has been looked for without success
Change reporting, aborting behaviour of failed xsi:type
Remove special search for docElt decl

catch missing eltDecl in type assignment

fix key upward merge bug

use decl when found for wildcard

restructure FSM label again

oops -- back out debugging patches

big cleanup of all tests etc. to avoid unnecessary AttributeErrors

abandon assignChildTypes in favour of doing it during parsing

Distinguish cases where we know what namespace a schema should have
  from those where we don't
Notice imports when validating schema documents
Check QName namespaces against imports when validating schema
documents

fix bug introduced in previous base fix

fix various mistakes wrt base for schema docs

todo note

remove ungrounded use of useDTD

include -D in usage

use atol, not atoi

more bug/error accounting fixup

propagate keepGoing info,
catch backtrace in right place

catch corner case of wildcard and proper elt applying in same content model

merge in key fix from main line,
add -D flag to turn on useDTD and propagate it through

provide default ofr keyTabs for no decl cases

implement upward merging of keytabs

dont go for ns attrs lookup off an attribute

accommodate to factory renaming and related changes

update schemaLocs correctly

accommodate shift to real reference to key/unique by QName
update some key(ref) messages accordingly

cover all file opening with redirect,
only check schemaLoc on root and when about to lose,
also check ns uri if new and about to lose

complete regularisation of schema doc processing

begin regularising all reading of schema files

*** empty log message ***

misc changes

Modify XML.py to be able to write plain files instead of LTXML files,
and make applyschema.py use this (so we don't rely on LTXML for output).
This loses us the ability to set the output encoding (for now; presumably
Python could do it).

handle removal of schemaNormalizedValue on error in list/union/list of
union cases

error message bug

handle attribute default/fixed better, with type and decl

protect a read

implement (most of) fixed/default for elements
support RDDL for command line, xsi:schemaLoc and namespace URIs

no namespace == no prefix == None

and again reflect2

fix reflect2 binding bug

(PSV)Infoset reorganisation

raise multiple key error

fix bug handling unknown xsi attrs

log target reading crash better

implement -i switch to do forced schema assessment independent of any instance

merge v2001 back in to main line

check for and rule out use of abstract types

handle unbound prefix in xsi:type
fix typo in assess

assignAttrTypes sets a.type to the use if there is one
so either valueConstraint can be used

fix key/keyref/unique field code

merge attr use back in to v2001 line

catch unbound prefix in xsi:type

use AttrUse to supply defaults

change NS to 2001, implement null->nil

report low-level failure correctly

accommodate to earlier XPath construction for fields/selector

merged infoset-based back to mainline

improve wildcard error msg

accommodate to change of exception with LTXMLInfoset

fix xsi:type bugs

fix spelling of whiteSpace,
add -p file switch for profiling

add whitespace processing,
fix some bugs?

accommodate to .facets and real facets

add logging of relevant declaration to elt/attr assess

fix bug in e-o error msg

add -r switch to produce reflected output

get builtin-schemas out first, sort AbInitios to the front,
more details in content-model error messages,
set null property

install schemaInformation on validation root,
assign type defn as such to typeDefinition property

put the docElt in the factory,
use it for reflection at the end,
and to implement validationContext

work around null vs "" for missing namespace name

handle xsi: attrs cleanly

make assess a method on Element,
add one on Attribute,
refer to the latter from the former

add psv infoset namespace URI to reflected docapplyschema.py

stubs for schemaNormalizedValue in place

remove convert,
accommodate change to importing XML,
put attribute verror on right item

merge in main line fixes thru 1.82

validate subordinate elements with eltdecl if available
return schema error count if not attempting instance validation

Output timing info if -t on command line

make effort to check 'fixed' attribute values

put switch on warnings, default is don't

try to catch and log all crashes
replace stale reference to atribute.characters

more infoset contributions

update values for validity property

More towards PSV infoset

correct errorCode setting

More towards PSV infoset

towards PSVI contributions

schema error count includes all errors, not just those found at prep
time

try catching and returning any crashes

add usage message
add -o outfile cmd line arg

Changes for infoset-based

Use infoset-like names for slots (now provided in XML.py) to reduce
differences with infoset-based version.

correct element.name to element.local in an error message

Oops, didn't change AbInitio to XMLSchema.AbInitio when moving methods

Move checkString methods from XMLSchema.py, because they may need to look
at *instance* in-scope namespaces

protect against undefined list itemType/union memberType

merge in CR branch

towards merge in CR branch

property name change

allow stdin,
fix stupid bug missing third schema on command line

Move to branch: more debug in vv crash

Move to branch: Direct support for validating lists and unions

handle type w/o qname as document validation type
remove special treatment for AbInitio simple types on elements,
thereby fixing list validation bug

Can now get encoding from documentEntity

try harder to always have a schema

change references to name properties to originalName

prepare for fileinfo to runit

move some methods from XMLSchema.py to applyschema.py
other fixes

changes for infoset-based version

change name to PyLTXML

bail out if textonly has elt daughter(s)
add missing import

attempt to handle interaction between xsi:type and <any>

fix bug in unqualified xsi:type

Bug in unique processing -- broke on missing field

merge xmlout branches back in to main line

cover several missing definition cases

oops, removing debugging comment

catch arity bug in xsi:schemaLoc

fix encoding bug when things break down altogether

strong enforcement of nullable
add error codes to all errors
remove remaining __class__ tests
change error reporting wrt disallowed content

make validateText a method, split across SimpleType and AbInitio

modest effort to validate list types
fix bug in noNamespaceSchemaLocation handling at validation time

use OpenStream, take more control of encoding

parameterise stylesheet,
remove formatting from xsv:xsv attributes,
add namespace decl

fix bug in handling of xsi:type

fix xsi:type validation bug

fix bug handling un-typed element declarations == urType validation

merge QName checking from main branch

add context to checkSting calls