Support new IGNORE-INVALID-SEQUENCESP argument, #'query-coding-region. lisp/ChangeLog addition: 2009-02-07 Aidan Kehoe <kehoea@parhasard.net> * coding.el (query-coding-clear-highlights): Rename the BUFFER argument to BUFFER-OR-STRING, describe it as possibly being a string in its documentation. (default-query-coding-region): Add a new IGNORE-INVALID-SEQUENCESP argument, document that this function does not support it. Bind case-fold-search to nil, we don't want this to influence what the function thinks is encodable or not. (query-coding-region): Add a new IGNORE-INVALID-SEQUENCESP argument, document what it does; reflect this new argument in the associated compiler macro. (query-coding-string): Add a new IGNORE-INVALID-SEQUENCESP argument, document what it does. Support the HIGHLIGHT argument correctly. * unicode.el (unicode-query-coding-region): Add a new IGNORE-INVALID-SEQUENCESP argument, document what it does, implement this. Document a potential problem. Use #'query-coding-clear-highlights instead of reimplementing it ourselves. Remove some debugging messages. * mule/arabic.el (iso-8859-6): * mule/cyrillic.el (iso-8859-5): * mule/greek.el (iso-8859-7): * mule/hebrew.el (iso-8859-8): * mule/latin.el (iso-8859-2): * mule/latin.el (iso-8859-3): * mule/latin.el (iso-8859-4): * mule/latin.el (iso-8859-14): * mule/latin.el (iso-8859-15): * mule/latin.el (iso-8859-16): * mule/latin.el (iso-8859-9): * mule/latin.el (windows-1252): * mule/mule-coding.el (iso-8859-1): Avoid the assumption that characters not given an explicit mapping in these coding systems map to the ISO 8859-1 characters corresponding to the octets on disk; this makes it much more reasonable to implement the IGNORE-INVALID-SEQUENCESP argument to query-coding-region. * mule/mule-cmds.el (set-language-info): Correct the docstring. * mule/mule-cmds.el (finish-set-language-environment): Treat invalid Unicode sequences produced from invalid-sequence-coding-system and corresponding to control characters the same as control characters in redisplay. * mule/mule-cmds.el: Document that encode-coding-char is available in coding.el * mule/mule-coding.el (make-8-bit-generate-helper): Change to return the both the encode-program generated and the relevant non-ASCII charset; update the docstring to reflect this. * mule/mule-coding.el (make-8-bit-generate-encode-program-and-skip-chars-strings): Rename this function; have it return skip-chars-strings as well as the encode program. Have these skip-chars-strings use ranges for charsets, where possible. * mule/mule-coding.el (make-8-bit-create-decode-encode-tables): Revise this to allow people to specify explicitly characters that should be undefined (= corresponding to keys in unicode-error-default-translation-table), and treating unspecified octets above #x7f as undefined by default. * mule/mule-coding.el (8-bit-fixed-query-coding-region): Add a new IGNORE-INVALID-SEQUENCESP argument, implement support for it using the 8-bit-fixed-invalid-sequences-skip-chars coding system property; remove some debugging messages. * mule/mule-coding.el (make-8-bit-coding-system): This function is dumped, autoloading it makes no sense. Document what happens when characters above #x7f are not specified, implement this. * mule/vietnamese.el: Correct spelling. tests/ChangeLog addition: 2009-02-07 Aidan Kehoe <kehoea@parhasard.net> * automated/query-coding-tests.el: Add FAILING-CASE arguments to the Assert calls, making #'q-c-debug mostly unnecessary. Remove #'q-c-debug. Add new tests that use the IGNORE-INVALID-SEQUENCESP argument to #'query-coding-region; rework the existing ones to respect it.

#ifndef INCLUDED_lwlib_internal_h_
#define INCLUDED_lwlib_internal_h_

#include "lwlib.h"

#ifdef USE_ASSERTIONS
/* Highly dubious kludge */
/*   (thanks, Jamie, I feel better now -- ben) */
EXTERN_C void assert_failed (const char *, int, const char *);
# define abort() (assert_failed (__FILE__, __LINE__, "abort()"))
# define assert(x) ((x) ? (void) 0 : assert_failed (__FILE__, __LINE__, #x))
#else
# ifdef DEBUG_XEMACS
#  define assert(x) ((x) ? (void) 0 : (void) abort ())
# else
#  define assert(x) ((void) 0)
# endif
#endif

/* This represents a single widget within a widget tree.  All the
   widgets in a widget tree are chained through the `next' field.
   `info' is a back pointer to the widget tree. */

typedef struct _widget_instance
{
  Widget		widget;
  Widget		parent;
  Boolean		pop_up_p;
  struct _widget_info*		info;
  struct _widget_instance*	next;
} widget_instance;

/* This represents a single widget tree, such as a single menubar.
   The global variable `all_widget_info' lists all widget trees,
   chained through the `next' field of this structure. */

typedef struct _widget_info
{
  char*			type;
  char*			name;
  LWLIB_ID		id;
  widget_value*		val;
  Boolean		busy;
  lw_callback		pre_activate_cb;
  lw_callback		selection_cb;
  lw_callback		post_activate_cb;
  struct _widget_instance*	instances;
  struct _widget_info*		next;
} widget_info;

typedef Widget
(*widget_creation_function) (widget_instance* instance);

typedef struct _widget_creation_entry
{
  const char*			type;
  widget_creation_function	function;
} widget_creation_entry;

/* update all other instances of a widget.  Can be used in a callback when
   a widget has been used by the user */
void
lw_internal_update_other_instances (Widget widget, XtPointer closure,
				    XtPointer call_data);

/* get the widget_value for a widget in a given instance */
widget_value*
lw_get_widget_value_for_widget (widget_instance* instance, Widget w);

widget_info *lw_get_widget_info (LWLIB_ID id);

#endif /* INCLUDED_lwlib_internal_h_ */