# t/xmlconf/expected-fail.txt: conformance cases this build is known to fail,
# one id per line with the reason. A case listed here that passes is a test
# failure too, so the list only ever shrinks. Decision F: empty at release, or
# every line names the spec sentence that justifies the disagreement.
#
# One line, and it is here because two partitions of the suite contradict each
# other. Both cases are the same construct: a character reference to a
# restricted character inside an INTERNAL entity's literal, referenced in
# element content.
#
#   eduni rmt-054   TYPE="valid",  and its own comment says the CR grammar made
#                   it illegal, that "this is probably not intended", and that
#                   it "will be fixed in the PR version".
#   ibm02n13        TYPE="not-wf", written against that same earlier draft.
#
# The Recommendation settles it. Exactly two productions subtract restricted
# characters, and neither reaches an internal entity's replacement text:
#
#   [1]  document     ::= ( prolog element Misc* ) - ( Char* RestrictedChar Char* )
#   [78] extParsedEnt ::= ( TextDecl? content )   - ( Char* RestrictedChar Char* )
#
# So this build follows the grammar and eduni: the document entity and every
# EXTERNAL parsed entity are held to it, an internal one is not. Passing
# ibm02n13 would mean failing rmt-054, and the two cannot both be satisfied.
# See frx_lex.h at frx_lex_char for the rule as implemented.

ibm-1-1-not-wf-P02-ibm02n13.xml  XML 1.1 productions [1] and [78] subtract RestrictedChar from the document entity and external parsed entities only, never an internal entity's replacement text; eduni rmt-054 is the same construct marked valid and this build follows the Recommendation
