GitLab

(lexer.__parse_userdefined_types): Handle @anchor.
(lexer.__parse_type): Requre @lt markup.
(lexer.__get_lt_token): New method.
(lexer.__get_token): Handle the new tokens '{' and '}'.

(LysKOM Data Types): Likewise.
(@lt): New macro.  Use it in all requests and asynchronous
	messages to mark up the types.

	incompatibilities regarding Latin-1 characters without having
	to clutter the document with @iftex constructs.  Removed all
	related @iftex constructs in the body of the document.  @ae{}
	can be used in both Info and TeX mode; do so.

	usage.  Previously, many but not all examples were introduced
	by a @i{Example:} line.  That line is now included in the
	@reqexample macro, together with a "@need 2000" statement.

(lexer.toplevel_asyncdlink): New method.

	where appropriate.
(Asynchronous Messages): Use @reqlink instead of @req.

	numbered-requests.tmp.  Create asyncs-numbered.tmp.  Omit
	DEBUG_CALLS-only asynchronous messages from "async.h".

@aarg markup for all asynchronous messages, and remove the trailing
semicolon in the argument list.  Merge all indices into one.
(add-member): Malformed sentence fixed.
(async-i-am-off, async-i-am-on-obsolete): Mark as obsolete, and
	tell why they are no longer used.
(async-broadcast): Mark as obsolete.
(Type Index, Request Index): Removed.
(Index): Added.

	__implemented_asyncs.
(lexer): Ignore @defcodeindex and @syncodeindex.
(lexer.toplevel_node): Check for undocumentet arguments to
	asynchronous messages.
(lexer.toplevel_findex): Check for @amindex entries in @findex nodes.
(lexer.toplevel_amindex): New method.
(lexer.__parse_async): New method.
(lexer.toplevel_aarg): Implemented.
(lexer.toplevel_bye): Check for unterminated @amindex nodes and
	undocumented asynchronous messages.

(lexer.toplevel_findex): Check that the same request name isn't used twice.
(lexer.toplevel_bye): Check that all implemented requests are documented.
(lexer.__parse_request): Check that the proper request number is
	used, and that the request is really implemented.

stuff.  Renamed ENUMREATION_OF to ENUMERATION-OF.  Minor fixes to the
protocol specification.  @reqdlink now works when using TeX.

(number_suffixed): New function.
(prot_a_type): New class, with several derived classes.
(reader.error): Allow the line number to be a string, such as '*builtin*'.
(lexer.__init__): Added 'ENUMERATION-OF' to the set of builtin
	aggregates.  Add builtin types to defined_types.
(lexer): Ignore @dots{}.
(lexer.toplevel_bye): Check that all defined types are used.
(lexer.toplevel_tindex): Pass a list of the @tindex entries to
	__parse_userdefined_types, so that that function can check that
	they are all defined.
(lexer.__parse_userdefined_types): Implemented.  Several helper
	functions introduced.
(lexer.__parse_type): Simplified.  Do some error checking and
	bookkeeping.  Return a tuple, so that it is easier for the caller
	to see if an array was used.  All callers updated.
(lexer.__bad_arg): Accept "reservedX" and "flgX" for numerical values of X.
(lexer.__get_token): Handle the new tokens "::=", "|" and "=".
	Ignore comments (introduced with "!").

	Nothing else uses "_".  The only user updated.
(LysKOM Data Types): Missing semicolon added.

(Simple Data Types): Use "::=" instead of ":" in two places.
	Added missing @example...@end example markup.  Fixed the
	SELECTION example -- the tail is not optional.

	__builtin_typelike, __tindex_seen and __types_used.
(lexer.__seen_type): New method.
(lexer.toplevel_tindex): Implement.
(lexer.__parse_userdefined_types): New method, not yet implemented.
(lexer.__bad_type): Check that the casing rules for user-defined
	types are adhered to.

	callers updated.  Clear __parenstack, so that errors are
	reported only once.
(lexer.__init__): Initialize __linkhere.
(lexer.toplevel_node): Check that no @linkhere or closing
	parenthesis are pending.
(lexer.toplevel_reqdlink, lexer.__assert_no_linkhere)
(lexer.toplevel_linkhere): New methods.

Protocol-A.texi.  Mark up all requests and fix all references in
Protocol-A.texi.  Check for balanced parenthesis.  Don't omit stuff
when formatting Protocol-A.texi as HTML.

(reader.__init__): Initialise __parenstack, __errfound and __quoted.
(reader.error, reader.errfound): New methods.
(reader.ungetc, reader.getc_eofok): Handle quoting and parens nesting.
(reader.check_paren_eof): Report unbalanced parens at EOF.
(lexer.__init__): Moved __errfound to the reader class.
(lexer.run): Check for unbalanced parens at EOF.
(lexer.toplevel_reqlink): Ignore.
(lexer.toplevel_reqdlink, lexer.toplevel_linkhere): Ignore, for
	now.  We should check nesting et c of these.
(lexer.toplevel_unmacro): Ignore.
(lexer.error): Moved the implementation to the reader class.  This
	is only a wrapper.

@ref, @xref and @pxref usages.  Also:
(@reqlink, @reqdlink, @linkhere): New markup macros.
(Document Edition History): Added missing @async markup.
(About Aux-Items): Added missing @field markup.
(Simple Data Types): Removed an unbalanced close parenthesis.
(LysKOM Data Types): Added missing @field markup. Removed an
	unbalanced close parenthesis.
(lookup-name): Added a missing close parenthesis.
(get-info-old): Added a missing close parenthesis.  This returns
	Info-Old, not Info.
(create-anonymous-text-old): Added missing @field markup.
(login): Added missing close parenthesis.
(local-to-global): Don't use fancy mathemtical notations for
	half-closed intervals.  Added missing @type markup.

(re-z-lookup): Added missing spaces in the example.
(lookup-z-name): Likewise.
(get-members): Likewise.
(Protocol Requests): Removed text about "upright text" in
	the replies in the examples, since no such text exists.

(Protocol Requests): Mention that extra newlines sometimes are used in
	the examples.
(create-person): Language fix.