From 173e68ccfa768342ddc7073efaaa2a1bc5f5690b Mon Sep 17 00:00:00 2001
From: Petr Baudis <pasky@ucw.cz>
Date: Fri, 6 Jan 2006 12:00:30 +0100
Subject: [PATCH 01/64] Support for pasting from GNU screen clipboard

---
 src/osdep/osdep.c | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

diff --git a/src/osdep/osdep.c b/src/osdep/osdep.c
index 9e1af2801..96091075f 100644
--- a/src/osdep/osdep.c
+++ b/src/osdep/osdep.c
@@ -347,10 +347,21 @@ exe(unsigned char *path)
 unsigned char *
 get_clipboard_text(void)	/* !!! FIXME */
 {
-	unsigned char *ret = mem_alloc(1);
+	unsigned char *ret;
 
+	/* GNU Screen's clipboard */
+	if (is_gnuscreen()) {
+		struct string str;
+
+		if (!init_string(&str)) return;
+
+		add_to_string(&str, "screen -X paste .");
+		if (str.length) exe(str.source);
+		if (str.source) done_string(&str);
+	}
+
+	ret = mem_alloc(1);
 	if (ret) ret[0] = 0;
-
 	return ret;
 }
 

From f18de1d3f3c2e96aba5bb6791a663a40be402fc1 Mon Sep 17 00:00:00 2001
From: Petr Baudis <pasky@ucw.cz>
Date: Fri, 6 Jan 2006 12:02:51 +0100
Subject: [PATCH 02/64] Internal clipboard support

---
 src/osdep/osdep.c | 15 +++++++--------
 1 file changed, 7 insertions(+), 8 deletions(-)

diff --git a/src/osdep/osdep.c b/src/osdep/osdep.c
index 96091075f..d22ace621 100644
--- a/src/osdep/osdep.c
+++ b/src/osdep/osdep.c
@@ -344,11 +344,11 @@ exe(unsigned char *path)
 
 #endif
 
-unsigned char *
-get_clipboard_text(void)	/* !!! FIXME */
-{
-	unsigned char *ret;
+static unsigned char *clipboard;
 
+unsigned char *
+get_clipboard_text(void)
+{
 	/* GNU Screen's clipboard */
 	if (is_gnuscreen()) {
 		struct string str;
@@ -360,9 +360,7 @@ get_clipboard_text(void)	/* !!! FIXME */
 		if (str.source) done_string(&str);
 	}
 
-	ret = mem_alloc(1);
-	if (ret) ret[0] = 0;
-	return ret;
+	return stracpy(clipboard ? clipboard : "");
 }
 
 void
@@ -381,7 +379,8 @@ set_clipboard_text(unsigned char *data)
 		if (str.source) done_string(&str);
 	}
 
-	/* TODO: internal clipboard */
+	if (clipboard) mem_free(clipboard);
+	clipboard = stracpy(data);
 }
 
 /* Set xterm-like term window's title. */

From 0c40ff05c94cb581763cd0e0c13e7279c736344c Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Fri, 6 Jan 2006 15:48:15 +0100
Subject: [PATCH 03/64] Fix compilation of get_clipboard_text()

---
 src/osdep/osdep.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/src/osdep/osdep.c b/src/osdep/osdep.c
index d22ace621..14710a7fa 100644
--- a/src/osdep/osdep.c
+++ b/src/osdep/osdep.c
@@ -353,14 +353,14 @@ get_clipboard_text(void)
 	if (is_gnuscreen()) {
 		struct string str;
 
-		if (!init_string(&str)) return;
+		if (!init_string(&str)) return NULL;
 
 		add_to_string(&str, "screen -X paste .");
 		if (str.length) exe(str.source);
 		if (str.source) done_string(&str);
 	}
 
-	return stracpy(clipboard ? clipboard : "");
+	return stracpy(empty_string_or_(clipboard));
 }
 
 void

From a637349b324f2763d20956388e2dc5073bb50c60 Mon Sep 17 00:00:00 2001
From: Petr Baudis <pasky@ucw.cz>
Date: Fri, 6 Jan 2006 17:08:38 +0100
Subject: [PATCH 04/64] Prevent complaining about leaking clipboard

---
 src/osdep/osdep.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/src/osdep/osdep.c b/src/osdep/osdep.c
index 14710a7fa..1503f84db 100644
--- a/src/osdep/osdep.c
+++ b/src/osdep/osdep.c
@@ -379,8 +379,9 @@ set_clipboard_text(unsigned char *data)
 		if (str.source) done_string(&str);
 	}
 
-	if (clipboard) mem_free(clipboard);
-	clipboard = stracpy(data);
+	/* Shouldn't complain about leaks. */
+	if (clipboard) free(clipboard);
+	clipboard = strdup(data);
 }
 
 /* Set xterm-like term window's title. */

From b1514c7a3573509ea03db83aaedd844f3fc0f4a0 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Fri, 6 Jan 2006 17:46:55 +0100
Subject: [PATCH 05/64] Check if the test 'trash' directory could be entered

Original patch by Alex Riesen <raa.lkml@gmail.com> for GIT.
---
 test/libtest.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/test/libtest.sh b/test/libtest.sh
index 2d40a9162..12de5c920 100755
--- a/test/libtest.sh
+++ b/test/libtest.sh
@@ -162,4 +162,4 @@ test_done () {
 test=trash
 rm -fr "$test"
 mkdir "$test"
-cd "$test"
+cd "$test" || error "Cannot setup test environment"

From 5f5c78a87ff5feab5252537e8d376e6b8a75fb94 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Fri, 6 Jan 2006 18:32:22 +0100
Subject: [PATCH 06/64] Realign the test docs with reality

---
 src/dom/test/README | 27 +++++++++++++++++----------
 1 file changed, 17 insertions(+), 10 deletions(-)

diff --git a/src/dom/test/README b/src/dom/test/README
index a980f16ed..2c4bcd008 100644
--- a/src/dom/test/README
+++ b/src/dom/test/README
@@ -1,7 +1,7 @@
-Core GIT Tests
-==============
+ELinks testing infrastructure
+=============================
 
-This directory holdstest scripts for the DOM implementation. The first part of
+This directory holds test scripts for the DOM implementation. The first part of
 this short document describes how to run the tests and read their output.
 
 When fixing the tools or adding enhancements, you are strongly encouraged to
@@ -27,7 +27,7 @@ The easiest way to run tests is to say "make test". This runs all the tests.
 
 Or you can run each test individually from command line, like this:
 
-    $ sh ./test-sgml-parser-basic
+    $ TEST_LIB=${path_to_top_srcdir}/test/libtest.sh sh ./test-sgml-parser-basic
     *   ok 1: parse a small document.
     ...
     *   ok 23: parse a CDATA section.
@@ -49,6 +49,10 @@ command line argument to the test.
 	This causes the test to immediately exit upon the first
 	failed test.
 
+Note, these options can be passed indirectly to all tests when running test using
+make by setting TEST_OPTS, like this:
+
+	make test TEST_OPTS=--immediate
 
 Naming Tests
 ------------
@@ -91,20 +95,23 @@ The test script is written as a shell script. It should start with the standard
 
 
 Source 'libtest'
---------------------
+----------------
 
-After assigning test_description, the test script should source test-lib.sh
-like this:
+After assigning test_description, the test script should source the shell test
+library like this:
 
-	. ./libtest
+	. "$TEST_LIB"
+
+This assumes that the TEST_LIB environment variable has been set and is needed
+for test to run from out of tree builds.
 
 This test harness library does the following things:
 
  - If the script is invoked with command line argument --help (or -h), it shows
    the test_description and exits.
 
- - Creates an empty test directory. This directory is 'test/trash' if you must
-   know, but I do not think you care.
+ - Creates an empty test directory. This directory is 'trash' if you must know,
+   but I do not think you care.
 
  - Defines standard test helper functions for your scripts to use.  These
    functions are designed to make all scripts behave consistently when command

From c08d932433a0baf8b61cc39fd9f9216a1a4126e0 Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Fri, 6 Jan 2006 21:34:57 +0100
Subject: [PATCH 07/64] Merge is_hotkey(), is_not_so_hotkey() and
 check_hotkeys_common() to reduce code redundancy and speed up things.

---
 src/bfu/hotkey.c | 79 ++++++++++++++++++++----------------------------
 1 file changed, 32 insertions(+), 47 deletions(-)

diff --git a/src/bfu/hotkey.c b/src/bfu/hotkey.c
index 42222c449..107fa5a4b 100644
--- a/src/bfu/hotkey.c
+++ b/src/bfu/hotkey.c
@@ -121,52 +121,9 @@ refresh_hotkeys(struct terminal *term, struct menu *menu)
 #endif
 }
 
-/* Returns true if key (upcased) matches one of the hotkeys in menu */
-static int
-is_hotkey(struct menu_item *item, unsigned char key, struct terminal *term)
-{
-	unsigned char *text;
-	int key_pos;
-
-	assert(item);
-	if_assert_failed return 0;
-
-	if (!mi_has_left_text(item)) return 0;
-
-	text = item->text;
-	if (mi_text_translate(item)) text = _(text, term);
-	if (!text || !*text) return 0;
-
-	key_pos = item->hotkey_pos;
-
-#ifdef CONFIG_DEBUG
-	if (key_pos < 0) key_pos = -key_pos;
-#endif
-
-	return (key_pos && (toupper(text[key_pos]) == key));
-}
-
-/* Returns true if key (upcased) matches first letter of menu item left text. */
-static int
-is_not_so_hotkey(struct menu_item *item, unsigned char key, struct terminal *term)
-{
-	unsigned char *text;
-
-	assert(item);
-	if_assert_failed return 0;
-
-	if (!mi_has_left_text(item)) return 0;
-
-	text = item->text;
-	if (mi_text_translate(item)) text = _(text, term);
-	if (!text || !*text) return 0;
-
-	return (toupper(*text) == key);
-}
-
 static int
 check_hotkeys_common(struct menu *menu, unsigned char hotkey, struct terminal *term,
-		     int (*func)(struct menu_item *, unsigned char, struct terminal *))
+		     int check_mode)
 {
 	unsigned char key = toupper(hotkey);
 	int i = menu->selected;
@@ -179,9 +136,37 @@ check_hotkeys_common(struct menu *menu, unsigned char hotkey, struct terminal *t
 
 	start = i;
 	do {
+		struct menu_item *item;
+		unsigned char *text;
+		int res;
+
 		if (++i == menu->size) i = 0;
 
-		if (func(&menu->items[i], key, term)) {
+		item = &menu->items[i];
+
+		if (!mi_has_left_text(item)) continue;
+
+		text = item->text;
+		if (mi_text_translate(item)) text = _(text, term);
+		if (!text || !*text) continue;
+
+		if (check_mode == 0) {
+			/* Does the key (upcased) matches one of the
+			 * hotkeys in menu ? */
+			int key_pos = item->hotkey_pos;
+
+#ifdef CONFIG_DEBUG
+			if (key_pos < 0) key_pos = -key_pos;
+#endif
+			res = (key_pos && (toupper(text[key_pos]) == key));
+
+		} else {
+			/* Does the key (upcased) matches first letter
+			 * of menu item left text ? */
+			res = (toupper(*text) == key);
+		}
+
+		if (res) {
 			menu->selected = i;
 			return 1;
 		}
@@ -195,7 +180,7 @@ check_hotkeys_common(struct menu *menu, unsigned char hotkey, struct terminal *t
 int
 check_hotkeys(struct menu *menu, unsigned char key, struct terminal *term)
 {
-	return check_hotkeys_common(menu, key, term, is_hotkey);
+	return check_hotkeys_common(menu, key, term, 0);
 }
 
 /* Search if first letter of an entry in menu matches the key (caseless comp.).
@@ -205,5 +190,5 @@ check_hotkeys(struct menu *menu, unsigned char key, struct terminal *term)
 int
 check_not_so_hot_keys(struct menu *menu, unsigned char key, struct terminal *term)
 {
-	return check_hotkeys_common(menu, key, term, is_not_so_hotkey);
+	return check_hotkeys_common(menu, key, term, 1);
 }

From d767650585952d5c0d478992e7ba90569c6c51c9 Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Fri, 6 Jan 2006 21:37:36 +0100
Subject: [PATCH 08/64] check_hotkeys_common(): @res -> @found.

---
 src/bfu/hotkey.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/src/bfu/hotkey.c b/src/bfu/hotkey.c
index 107fa5a4b..0c3078dec 100644
--- a/src/bfu/hotkey.c
+++ b/src/bfu/hotkey.c
@@ -138,7 +138,7 @@ check_hotkeys_common(struct menu *menu, unsigned char hotkey, struct terminal *t
 	do {
 		struct menu_item *item;
 		unsigned char *text;
-		int res;
+		int found;
 
 		if (++i == menu->size) i = 0;
 
@@ -158,15 +158,15 @@ check_hotkeys_common(struct menu *menu, unsigned char hotkey, struct terminal *t
 #ifdef CONFIG_DEBUG
 			if (key_pos < 0) key_pos = -key_pos;
 #endif
-			res = (key_pos && (toupper(text[key_pos]) == key));
+			found = (key_pos && (toupper(text[key_pos]) == key));
 
 		} else {
 			/* Does the key (upcased) matches first letter
 			 * of menu item left text ? */
-			res = (toupper(*text) == key);
+			found = (toupper(*text) == key);
 		}
 
-		if (res) {
+		if (found) {
 			menu->selected = i;
 			return 1;
 		}

From 84f8973720c21699caf781290ab9244dc6d8e7cd Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Fri, 6 Jan 2006 22:08:49 +0100
Subject: [PATCH 09/64] Make the compile macro use -o and append any 2nd
 argument

---
 Makefile.lib | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/Makefile.lib b/Makefile.lib
index 7e102c3d4..a56d45cd1 100644
--- a/Makefile.lib
+++ b/Makefile.lib
@@ -33,8 +33,8 @@ mcmd = @$(if $($(mquiet)cmd_$(1)),echo $($(mquiet)cmd_$(1)) &&) $(cmd_$(1))
 ecmd = @$(if $($(mquiet)cmd_$(1)),printf "%-38s " $($(mquiet)cmd_$(1)) &&) $(cmd_$(1))
 
 quiet_cmd_compile = '      [$(CC_COLOR)CC$(END_COLOR)]   $(RELPATH)$@'
- masq_cmd_compile = $(COMPILE) -c $<
-      cmd_compile = $(COMPILE) -Wp,-MD,.deps/$(*F).pp -c $<
+ masq_cmd_compile = $(COMPILE) -o $(@) -c $< $(2)
+      cmd_compile = $(COMPILE) -o $(@) -Wp,-MD,.deps/$(*F).pp -c $< $(2)
 
 # Rule to compile a set of .o files into one .o file
 quiet_cmd_ld_objs = "      [$(LD_COLOR)LD$(END_COLOR)]   $(RELPATH)$@"

From 215d7ec1589dee752b9e66a52c467227f492ff17 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Fri, 6 Jan 2006 22:11:45 +0100
Subject: [PATCH 10/64] Append memdebug to test dependencies in Makefile.lib

---
 Makefile.lib              | 1 +
 src/dom/test/Makefile     | 2 --
 src/protocol/ftp/Makefile | 2 --
 3 files changed, 1 insertion(+), 4 deletions(-)

diff --git a/Makefile.lib b/Makefile.lib
index a56d45cd1..4cedf9124 100644
--- a/Makefile.lib
+++ b/Makefile.lib
@@ -153,6 +153,7 @@ clean-test:
 test-default:
 
 ifdef TEST_PROGS
+TESTDEPS-$(CONFIG_DEBUG) += $(top_builddir)/src/util/memdebug.o
 TESTDEPS += $(TESTDEPS-yes)
 
 TESTS = $(wildcard $(srcdir)test-*)
diff --git a/src/dom/test/Makefile b/src/dom/test/Makefile
index 3b7bd24d5..2bf9957e9 100644
--- a/src/dom/test/Makefile
+++ b/src/dom/test/Makefile
@@ -14,6 +14,4 @@ TESTDEPS = \
  $(top_builddir)/src/util/string.o \
  $(top_builddir)/src/util/memory.o
 
-TESTDEPS-$(CONFIG_DEBUG) += $(top_builddir)/src/util/memdebug.o \
-
 include $(top_srcdir)/Makefile.lib
diff --git a/src/protocol/ftp/Makefile b/src/protocol/ftp/Makefile
index 9aa1dc759..115744b79 100644
--- a/src/protocol/ftp/Makefile
+++ b/src/protocol/ftp/Makefile
@@ -16,6 +16,4 @@ TESTDEPS = \
  $(top_builddir)/src/util/string.o \
  $(top_builddir)/src/util/time.o
 
-TESTDEPS-$(CONFIG_DEBUG) += $(top_builddir)/src/util/memdebug.o
-
 include $(top_srcdir)/Makefile.lib

From 3fd2828fb0dc3c389dea810eeec583f32e1e106d Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Fri, 6 Jan 2006 22:18:11 +0100
Subject: [PATCH 11/64] Add a simple test script of the mailcap cache

It has a few nasty hacks to reduce the dependencies on defined symbols.
---
 src/mime/backend/Makefile           |  22 ++++++
 src/mime/backend/mailcap.c          | 111 ++++++++++++++++++++++++++++
 src/mime/backend/test-mailcap-cache |  95 ++++++++++++++++++++++++
 3 files changed, 228 insertions(+)
 create mode 100755 src/mime/backend/test-mailcap-cache

diff --git a/src/mime/backend/Makefile b/src/mime/backend/Makefile
index b1e61a0c0..eb9300cc5 100644
--- a/src/mime/backend/Makefile
+++ b/src/mime/backend/Makefile
@@ -6,4 +6,26 @@ OBJS-$(CONFIG_MIMETYPES)	+= mimetypes.o
 
 OBJS = common.o default.o
 
+TEST_PROGS = \
+ mailcap-cache
+
+# The dependencies are a bit funny here! I don't know why. Just remember to
+# make clean before making the test. --jonas
+mailcap-cache.o: mailcap.c
+	$(call cmd,compile,-DTEST_MAILCAP)
+
+TESTDEPS = \
+ common.o \
+ $(top_builddir)/src/osdep/osdep.o \
+ $(top_builddir)/src/osdep/stub.o \
+ $(top_builddir)/src/util/conv.o \
+ $(top_builddir)/src/util/error.o \
+ $(top_builddir)/src/util/file.o \
+ $(top_builddir)/src/util/hash.o \
+ $(top_builddir)/src/util/memory.o \
+ $(top_builddir)/src/util/string.o \
+ $(top_builddir)/src/util/time.o
+
+TESTDEPS-$(CONFIG_DEBUG) += $(top_builddir)/src/util/memdebug.o
+
 include $(top_srcdir)/Makefile.lib
diff --git a/src/mime/backend/mailcap.c b/src/mime/backend/mailcap.c
index bf9daec38..e3eccec55 100644
--- a/src/mime/backend/mailcap.c
+++ b/src/mime/backend/mailcap.c
@@ -443,6 +443,8 @@ done_mailcap(struct module *module)
 	mailcap_map_size = 0;
 }
 
+#ifndef TEST_MAILCAP
+
 static int
 change_hook_mailcap(struct session *ses, struct option *current, struct option *changed)
 {
@@ -469,6 +471,10 @@ init_mailcap(struct module *module)
 		get_mailcap_enable() = 0;
 }
 
+#else
+#define init_mailcap NULL
+#endif /* TEST_MAILCAP */
+
 /* The command semantics include the following:
  *
  * %s		is the filename that contains the mail body data
@@ -673,3 +679,108 @@ struct module mailcap_mime_module = struct_module(
 	/* init: */		init_mailcap,
 	/* done: */		done_mailcap
 );
+
+#ifdef TEST_MAILCAP
+/* Some ugly shortcuts for getting defined symbols to work. */
+int default_mime_backend,
+    install_signal_handler,
+    mimetypes_mime_backend;
+struct list_head terminals;
+
+void die(const char *msg, ...)
+{
+	va_list args;
+
+	if (msg) {
+		va_start(args, msg);
+		vfprintf(stderr, msg, args);
+		fputs("\n", stderr);
+		va_end(args);
+	}
+
+	exit(1);
+}
+
+int
+main(int argc, char *argv[])
+{
+	unsigned char *format = "description,ask,block,program";
+	int has_gotten = 0;
+	int i;
+
+	for (i = 1; i < argc; i++) {
+		char *arg = argv[i];
+
+		if (strncmp(arg, "--", 2))
+			break;
+
+		arg += 2;
+
+		if (!strncmp(arg, "path", 4)) {
+			arg += 4;
+			if (*arg == '=') {
+				arg++;
+				get_mailcap_path() = arg;
+			} else {
+				i++;
+				if (i >= argc)
+					die("--path expects a parameter");
+				get_mailcap_path() = argv[i];
+			}
+			done_mailcap(NULL);
+
+		} else if (!strncmp(arg, "format", 6)) {
+			arg += 6;
+			if (*arg == '=') {
+				arg++;
+				format = arg;
+			} else {
+				i++;
+				if (i >= argc)
+					die("--format expects a parameter");
+				format = argv[i];
+			}
+
+		} else if (!strncmp(arg, "get", 3)) {
+			struct mime_handler *handler;
+
+			arg += 3;
+			if (*arg == '=') {
+				arg++;
+			} else {
+				i++;
+				if (i >= argc)
+					die("--get expects a parameter");
+				arg = argv[i];
+			}
+
+			if (has_gotten)
+				printf("\n");
+			has_gotten = 1;
+			printf("type: %s\n", arg);
+			handler = get_mime_handler_mailcap(arg, 0);
+			if (!handler) continue;
+
+			if (strstr(format, "description"))
+				printf("description: %s\n", handler->description);
+
+			if (strstr(format, "ask"))
+				printf("ask: %d\n", handler->ask);
+
+			if (strstr(format, "block"))
+				printf("block: %d\n", handler->block);
+
+			if (strstr(format, "program"))
+				printf("program: %s\n", handler->program);
+
+		} else {
+			die("Unknown argument '%s'", arg - 2);
+		}
+	}
+
+	done_mailcap(NULL);
+
+	return 0;
+}
+
+#endif /* TEST_MAILCAP */
diff --git a/src/mime/backend/test-mailcap-cache b/src/mime/backend/test-mailcap-cache
new file mode 100755
index 000000000..6bd9bbf07
--- /dev/null
+++ b/src/mime/backend/test-mailcap-cache
@@ -0,0 +1,95 @@
+#!/bin/sh
+#
+# Copyright (c) 2005 Jonas Fonseca
+#
+
+test_description='Test mailcap parsing and querying
+
+This tests the parsing of various mailcap files, if they are
+"prioritised" correctly, if the test are run correctly and
+if querying returns the expected mailcap entry.
+'
+
+. "$TEST_LIB"
+
+# Set PAGER to something recognisable since it gets appended as
+# "|copiousoutput_handler" to entries with copiousoutput.
+export PAGER=copiousoutput_handler
+
+################################################################
+# Parse a simple mailcap file
+
+cat > mailcap-basic <<EOF
+# Mimetype		Handler
+text/html;		elinks --force-html %s; needsterminal
+text/enriched;		richtext ; copiousoutput
+text/*;			view %s; needsterminal
+application/postscript;	ps2ascii %s ; copiousoutput
+
+# Convert images to text using the netpbm tools
+image/*;		(anytopnm %s | pnmscale -xysize 200 150 | \
+                         pnminvert | ppmtopgm | pgmtopbm | \
+			 pbmtoascii -1x2 ) 2>&1 ; copiousoutput
+EOF
+
+mailcap-cache \
+	--path "mailcap-basic" \
+	--format "block,program" \
+	--get "text/html" \
+	--get "text/x-csh" \
+	--get "application/postscript" \
+	--get "application/foo" \
+	> output
+
+cat > expected <<EOF
+type: text/html
+block: 1
+program: elinks --force-html %
+
+type: text/x-csh
+block: 1
+program: view %
+
+type: application/postscript
+block: 1
+program: ps2ascii %|copiousoutput_handler
+
+type: application/foo
+EOF
+
+test_expect_success \
+	'Parse simple mailcap file.' \
+	'cmp output expected' \
+
+################################################################
+# Parse a two simple mailcap file; first one with tests
+
+touch DISPLAY
+
+cat > mailcap-simple-with-test <<EOF
+application/postscript;	gv %s ; test=test -e "DISPLAY" ;
+image/*;		xzgv %s ; test=test -e "DISPLAY";
+EOF
+
+mailcap-cache \
+	--path "mailcap-simple-with-test:mailcap-simple" \
+	--format "block,program" \
+	--get "image/jpeg" \
+	--get "application/postscript" \
+	> output
+
+cat > expected <<EOF
+type: image/jpeg
+block: 0
+program: xzgv %
+
+type: application/postscript
+block: 0
+program: gv %
+EOF
+
+test_expect_success \
+	'Parse two simple mailcap file; first one with tests.' \
+	'cmp output expected' \
+
+test_done

From 2a11525d1aaf5ca9e3ca63616bc182897dd93c0f Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Fri, 6 Jan 2006 22:32:07 +0100
Subject: [PATCH 12/64] Fix a few typos

---
 src/mime/backend/test-mailcap-cache | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/src/mime/backend/test-mailcap-cache b/src/mime/backend/test-mailcap-cache
index 6bd9bbf07..5440066c3 100755
--- a/src/mime/backend/test-mailcap-cache
+++ b/src/mime/backend/test-mailcap-cache
@@ -62,7 +62,7 @@ test_expect_success \
 	'cmp output expected' \
 
 ################################################################
-# Parse a two simple mailcap file; first one with tests
+# Parse a two simple mailcap files; first one with tests
 
 touch DISPLAY
 
@@ -89,7 +89,7 @@ program: gv %
 EOF
 
 test_expect_success \
-	'Parse two simple mailcap file; first one with tests.' \
+	'Parse two simple mailcap files; first one with tests.' \
 	'cmp output expected' \
 
 test_done

From bca330fcbdd20c4ee1475b92cac36089b37f381e Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 01:22:14 +0100
Subject: [PATCH 13/64] Simplify incomplete test helper and fix quoting problem

---
 src/dom/test/test-sgml-parser-basic | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/src/dom/test/test-sgml-parser-basic b/src/dom/test/test-sgml-parser-basic
index 4a379c265..28daae50e 100755
--- a/src/dom/test/test-sgml-parser-basic
+++ b/src/dom/test/test-sgml-parser-basic
@@ -43,8 +43,9 @@ test_expect_incomplete () {
 		y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/;
 		s/[^a-zA-Z0-9-]//g;')"
 
-	test_expect_code 1 "$desc" \
-		"sgml-parser --uri '$URI' --src '$src' --incomplete"
+	sgml-parser --uri "$URI" --src "$src" --incomplete
+	test_expect_success "$desc" \
+		"test $? = 1"
 }
 
 
@@ -354,7 +355,7 @@ test_expect_incomplete \
 
 test_expect_incomplete \
 'Check incomplete attribute quoted value. (II)' \
-'<element attr='\''...'
+"<element attr='..."
 
 test_expect_incomplete \
 'Check incomplete processing instruction. (I)' \

From f8d44ffe326ec32f85309b091da71f761a1d2610 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 01:25:42 +0100
Subject: [PATCH 14/64] scan_sgml_tokens(): Drop local variable and use
 scanner->current

... so lower level scanners can change the next token to use.
---
 src/dom/sgml/scanner.c | 19 +++++++++----------
 1 file changed, 9 insertions(+), 10 deletions(-)

diff --git a/src/dom/sgml/scanner.c b/src/dom/sgml/scanner.c
index 6399b4395..7396b7933 100644
--- a/src/dom/sgml/scanner.c
+++ b/src/dom/sgml/scanner.c
@@ -622,35 +622,34 @@ static struct dom_scanner_token *
 scan_sgml_tokens(struct dom_scanner *scanner)
 {
 	struct dom_scanner_token *table_end = scanner->table + DOM_SCANNER_TOKENS;
-	struct dom_scanner_token *current;
 
 	if (!begin_dom_token_scanning(scanner))
 		return get_dom_scanner_token(scanner);
 
 	/* Scan tokens until we fill the table */
-	for (current = scanner->table + scanner->tokens;
-	     current < table_end && scanner->position < scanner->end;
-	     current++) {
+	for (scanner->current = scanner->table + scanner->tokens;
+	     scanner->current < table_end && scanner->position < scanner->end;
+	     scanner->current++) {
 		if (scanner->state == SGML_STATE_ELEMENT
 		    || (*scanner->position == '<'
 			&& scanner->state != SGML_STATE_PROC_INST)) {
 			skip_sgml_space(scanner, &scanner->position);
 			if (scanner->position >= scanner->end) break;
 
-			scan_sgml_element_token(scanner, current);
+			scan_sgml_element_token(scanner, scanner->current);
 
 			/* Shall we scratch this token? */
-			if (current->type == SGML_TOKEN_SKIP) {
-				current--;
+			if (scanner->current->type == SGML_TOKEN_SKIP) {
+				scanner->current--;
 			}
 
 		} else if (scanner->state == SGML_STATE_TEXT) {
-			scan_sgml_text_token(scanner, current);
+			scan_sgml_text_token(scanner, scanner->current);
 
 		} else {
-			scan_sgml_proc_inst_token(scanner, current);
+			scan_sgml_proc_inst_token(scanner, scanner->current);
 		}
 	}
 
-	return end_dom_token_scanning(scanner, current);
+	return end_dom_token_scanning(scanner, scanner->current);
 }

From c9c41e38a2d3b8bfa21cda74949062c2dc96f27a Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 01:27:48 +0100
Subject: [PATCH 15/64] test_expect_incomplete(): Put sgml-parser output to
 /dev/null

---
 src/dom/test/test-sgml-parser-basic | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/dom/test/test-sgml-parser-basic b/src/dom/test/test-sgml-parser-basic
index 28daae50e..c3a5e2e8a 100755
--- a/src/dom/test/test-sgml-parser-basic
+++ b/src/dom/test/test-sgml-parser-basic
@@ -43,7 +43,7 @@ test_expect_incomplete () {
 		y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/;
 		s/[^a-zA-Z0-9-]//g;')"
 
-	sgml-parser --uri "$URI" --src "$src" --incomplete
+	sgml-parser --uri "$URI" --src "$src" --incomplete >/dev/null
 	test_expect_success "$desc" \
 		"test $? = 1"
 }

From 450e44a598d394e4090fad8256fd03afc127ddc4 Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Sat, 7 Jan 2006 01:28:54 +0100
Subject: [PATCH 16/64] exec_on_terminal(): simplify code.

---
 src/terminal/terminal.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/src/terminal/terminal.c b/src/terminal/terminal.c
index 895e89806..7b9241eb8 100644
--- a/src/terminal/terminal.c
+++ b/src/terminal/terminal.c
@@ -201,17 +201,17 @@ void
 exec_on_terminal(struct terminal *term, unsigned char *path,
 		 unsigned char *delete, int fg)
 {
-	int plen;
-	int dlen = strlen(delete);
+	int plen, dlen;
 
-	if (path && !*path) return;
-	if (!path) {
-		path = "";
-		plen = 0;
+	if (path) {
+		if (!*path) return;
 	} else {
-		plen = strlen(path);
+		path = "";
 	}
-
+	
+	plen = strlen(path);
+	dlen = strlen(delete);
+		
 #ifdef NO_FG_EXEC
 	fg = 0;
 #endif

From 7c65c06b413ac81eb62b67dcde0f246e46b0380f Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 01:29:44 +0100
Subject: [PATCH 17/64] Move up enum sgml_parser_code declaration

---
 src/dom/sgml/parser.h | 20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

diff --git a/src/dom/sgml/parser.h b/src/dom/sgml/parser.h
index de836a9f8..4764c91b7 100644
--- a/src/dom/sgml/parser.h
+++ b/src/dom/sgml/parser.h
@@ -38,6 +38,16 @@ struct sgml_parser_state {
 	struct dom_scanner_token end_token;
 };
 
+enum sgml_parser_code {
+	SGML_PARSER_CODE_OK,		/* The parsing was successful */
+	SGML_PARSER_CODE_INCOMPLETE,	/* The parsing could not be completed */
+	SGML_PARSER_CODE_MEM_ALLOC,	/* Failed to allocate memory */
+
+	/* FIXME: For when we will add support for requiring stricter parsing
+	 * or even a validator. */
+	SGML_PARSER_CODE_ERROR,
+};
+
 struct sgml_parser {
 	enum sgml_parser_type type;	/* Stream or tree */
 	enum sgml_parser_flag flags;	/* Flags that control the behaviour */
@@ -57,16 +67,6 @@ init_sgml_parser(enum sgml_parser_type type, enum sgml_document_type doctype,
 
 void done_sgml_parser(struct sgml_parser *parser);
 
-enum sgml_parser_code {
-	SGML_PARSER_CODE_OK,		/* The parsing was successful */
-	SGML_PARSER_CODE_INCOMPLETE,	/* The parsing could not be completed */
-	SGML_PARSER_CODE_MEM_ALLOC,	/* Failed to allocate memory */
-
-	/* FIXME: For when we will add support for requiring stricter parsing
-	 * or even a validator. */
-	SGML_PARSER_CODE_ERROR,
-};
-
 enum sgml_parser_code
 parse_sgml(struct sgml_parser *parser, struct dom_string *buffer, int complete);
 

From 7ff2cb26074e35cdf21f5681bc1f639673d80eba Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 01:41:07 +0100
Subject: [PATCH 18/64] Improve a comment a bit

---
 src/dom/sgml/scanner.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/src/dom/sgml/scanner.c b/src/dom/sgml/scanner.c
index 7396b7933..041d1b57d 100644
--- a/src/dom/sgml/scanner.c
+++ b/src/dom/sgml/scanner.c
@@ -599,7 +599,8 @@ scan_sgml_proc_inst_token(struct dom_scanner *scanner, struct dom_scanner_token
 	}
 
 	if (!string) {
-		/* Makes the next succeed when checking for incompletion. */
+		/* Makes the next succeed when checking for incompletion, and
+		 * puts the rest of the text within the token. */
 		string = scanner->end;
 
 		if (check_sgml_incomplete(scanner, string)) {

From a2b9d8bb780777728b5624185b8ab86ef77dbffb Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Sat, 7 Jan 2006 01:47:36 +0100
Subject: [PATCH 19/64] Explode exec_on_terminal() to new
 exec_on_master_terminal() and exec_on_slave_terminal().

---
 src/terminal/terminal.c | 155 ++++++++++++++++++++++------------------
 1 file changed, 85 insertions(+), 70 deletions(-)

diff --git a/src/terminal/terminal.c b/src/terminal/terminal.c
index 7b9241eb8..8dd536319 100644
--- a/src/terminal/terminal.c
+++ b/src/terminal/terminal.c
@@ -197,91 +197,106 @@ unblock_terminal(struct terminal *term)
 		textarea_edit(1, NULL, NULL, NULL, NULL);
 }
 
+
+static void
+exec_on_master_terminal(struct terminal *term,
+			unsigned char *path, int plen,
+		 	unsigned char *delete, int dlen,
+			int fg)
+{
+	int blockh;
+	unsigned char *param;
+	int param_size;
+
+	if (is_blocked() && fg) {
+		unlink(delete);
+		return;
+	}
+
+	param_size = plen + dlen + 2 /* 2 null char */ + 1 /* fg */;
+	param = mem_alloc(param_size);
+	if (!param) return;
+
+	param[0] = fg;
+	memcpy(param + 1, path, plen + 1);
+	memcpy(param + 1 + plen + 1, delete, dlen + 1);
+
+	if (fg == 1) block_itrm(term->fdin);
+
+	blockh = start_thread((void (*)(void *, int)) exec_thread,
+			      param, param_size);
+	if (blockh == -1) {
+		if (fg == 1) unblock_itrm(term->fdin);
+		mem_free(param);
+		return;
+	}
+
+	mem_free(param);
+	if (fg == 1) {
+		term->blocked = blockh;
+		set_handlers(blockh,
+			     (select_handler_T) unblock_terminal,
+			     NULL,
+			     (select_handler_T) unblock_terminal,
+			     term);
+		set_handlers(term->fdin, NULL, NULL,
+			     (select_handler_T) destroy_terminal,
+			     term);
+		/* block_itrm(term->fdin); */
+	} else {
+		set_handlers(blockh, close_handle, NULL,
+			     close_handle, (void *) (long) blockh);
+	}
+}
+
+static void
+exec_on_slave_terminal( struct terminal *term,
+		 	unsigned char *path, int plen,
+		 	unsigned char *delete, int dlen,
+			int fg)
+{
+	int data_size = plen + dlen + 1 /* 0 */ + 1 /* fg */ + 2 /* 2 null char */;
+	unsigned char *data = mem_alloc(data_size);
+
+	if (data) {
+		data[0] = 0;
+		data[1] = fg;
+		memcpy(data + 2, path, plen + 1);
+		memcpy(data + 2 + plen + 1, delete, dlen + 1);
+		hard_write(term->fdout, data, data_size);
+		mem_free(data);
+	}
+}
+
 void
 exec_on_terminal(struct terminal *term, unsigned char *path,
 		 unsigned char *delete, int fg)
 {
-	int plen, dlen;
-
 	if (path) {
 		if (!*path) return;
 	} else {
 		path = "";
 	}
-	
-	plen = strlen(path);
-	dlen = strlen(delete);
-		
+
 #ifdef NO_FG_EXEC
 	fg = 0;
 #endif
+
 	if (term->master) {
-		if (!*path) dispatch_special(delete);
-		else {
-			int blockh;
-			unsigned char *param;
-			int param_size;
-
-			if (is_blocked() && fg) {
-				unlink(delete);
-				return;
-			}
-
-			param_size = plen + dlen + 2 /* 2 null char */ + 1 /* fg */;
-			param = mem_alloc(param_size);
-			if (!param) return;
-
-			param[0] = fg;
-			memcpy(param + 1, path, plen + 1);
-			memcpy(param + 1 + plen + 1, delete, dlen + 1);
-
-			if (fg == 1) block_itrm(term->fdin);
-
-			blockh = start_thread((void (*)(void *, int)) exec_thread,
-					      param, param_size);
-			if (blockh == -1) {
-				if (fg == 1) unblock_itrm(term->fdin);
-				mem_free(param);
-				return;
-			}
-
-			mem_free(param);
-			if (fg == 1) {
-				term->blocked = blockh;
-				set_handlers(blockh,
-					     (select_handler_T) unblock_terminal,
-					     NULL,
-					     (select_handler_T) unblock_terminal,
-					     term);
-				set_handlers(term->fdin, NULL, NULL,
-					     (select_handler_T) destroy_terminal,
-					     term);
-				/* block_itrm(term->fdin); */
-			} else {
-				set_handlers(blockh, close_handle, NULL,
-					     close_handle, (void *) (long) blockh);
-			}
+		if (!*path) {
+			dispatch_special(delete);
+			return;
 		}
+
+		exec_on_master_terminal(term,
+					path, strlen(path),
+		 			delete, strlen(delete),
+					fg);
 	} else {
-		int data_size = plen + dlen + 1 /* 0 */ + 1 /* fg */ + 2 /* 2 null char */;
-		unsigned char *data = mem_alloc(data_size);
-
-		if (data) {
-			data[0] = 0;
-			data[1] = fg;
-			memcpy(data + 2, path, plen + 1);
-			memcpy(data + 2 + plen + 1, delete, dlen + 1);
-			hard_write(term->fdout, data, data_size);
-			mem_free(data);
-		}
-#if 0
-		char x = 0;
-		hard_write(term->fdout, &x, 1);
-		x = fg;
-		hard_write(term->fdout, &x, 1);
-		hard_write(term->fdout, path, strlen(path) + 1);
-		hard_write(term->fdout, delete, strlen(delete) + 1);
-#endif
+		exec_on_slave_terminal( term,
+					path, strlen(path),
+		 			delete, strlen(delete),
+					fg);
 	}
 }
 

From dee8ac5b45f4e863bb4207c1568aa2f234ffcc0f Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 01:48:51 +0100
Subject: [PATCH 20/64] Move test for incompleteness to own file

---
 src/dom/test/test-sgml-parser-basic      |  97 -------------------
 src/dom/test/test-sgml-parser-incomplete | 115 +++++++++++++++++++++++
 2 files changed, 115 insertions(+), 97 deletions(-)
 create mode 100755 src/dom/test/test-sgml-parser-incomplete

diff --git a/src/dom/test/test-sgml-parser-basic b/src/dom/test/test-sgml-parser-basic
index c3a5e2e8a..742e81280 100755
--- a/src/dom/test/test-sgml-parser-basic
+++ b/src/dom/test/test-sgml-parser-basic
@@ -31,23 +31,6 @@ test_output_equals () {
 	test_expect_success "$desc" 'cmp output expected' 
 }
 
-test_expect_incomplete () {
-	desc="$1"; shift
-	src="$1"; shift
-
-	URI="test:$(echo "$desc" | sed '
-		s/^[ \t]*\[[^]]*\][ \t]*//;
-		s/[:., \t][:., \t]*/-/g;
-		s/_/-/g;
-		# *cough*
-		y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/;
-		s/[^a-zA-Z0-9-]//g;')"
-
-	sgml-parser --uri "$URI" --src "$src" --incomplete >/dev/null
-	test_expect_success "$desc" \
-		"test $? = 1"
-}
-
 
 ################################################################
 # Parse various SGML node types.
@@ -300,85 +283,5 @@ test_output_equals \
 '
 8' \
 --print-lines
- 
-test_expect_incomplete \
-'Check incomplete comment. (I)' \
-'<!-'
-
-test_expect_incomplete \
-'Check incomplete comment. (II)' \
-'<!-- ... '
-
-test_expect_incomplete \
-'Check incomplete notation. (I)' \
-'<!'
-
-test_expect_incomplete \
-'Check incomplete notation. (II)' \
-'<!D'
-
-test_expect_incomplete \
-'Check incomplete cdata section. (I)' \
-'<![CDATA[ ... '
-
-test_expect_incomplete \
-'Check incomplete cdata section. (II)' \
-'<![CDAT'
-
-test_expect_incomplete \
-'Check incomplete element. (I)' \
-'<elem...'
-
-test_expect_incomplete \
-'Check incomplete element. (II)' \
-'<'
-
-test_expect_incomplete \
-'Check incomplete element end. (I)' \
-'<a></a'
-
-test_expect_incomplete \
-'Check incomplete element end. (II)' \
-'<a></'
-
-test_expect_incomplete \
-'Check incomplete attribute.' \
-'<element attr...'
-
-test_expect_incomplete \
-'Check incomplete attribute value.' \
-'<element attr=...'
-
-test_expect_incomplete \
-'Check incomplete attribute quoted value. (I)' \
-'<element attr="...'
-
-test_expect_incomplete \
-'Check incomplete attribute quoted value. (II)' \
-"<element attr='..."
-
-test_expect_incomplete \
-'Check incomplete processing instruction. (I)' \
-'<?xml'
-
-test_expect_incomplete \
-'Check incomplete processing instruction. (II)' \
-'<?xml attr...'
-
-test_expect_incomplete \
-'Check incomplete notation.' \
-'<!DOCTYPE html PUBLIC ...'
-
-test_expect_incomplete \
-'Check incomplete reference. (I)' \
-'&#123456789'
-
-test_expect_incomplete \
-'Check incomplete reference. (II)' \
-'&amp'
-
-test_expect_incomplete \
-'Check incomplete text.' \
-'plain text is always incomplete (if incomplete)'
 
 test_done
diff --git a/src/dom/test/test-sgml-parser-incomplete b/src/dom/test/test-sgml-parser-incomplete
new file mode 100755
index 000000000..697c951d7
--- /dev/null
+++ b/src/dom/test/test-sgml-parser-incomplete
@@ -0,0 +1,115 @@
+#!/bin/sh
+#
+# Copyright (c) 2005 Jonas Fonseca
+#
+
+test_description='Test that incompleteness is correctly detected
+
+This test checks that the SGML parser correctly finds and reports
+when a part of the source is incomplete.
+'
+
+. "$TEST_LIB"
+
+test_expect_incomplete () {
+	desc="$1"; shift
+	src="$1"; shift
+
+	URI="test:$(echo "$desc" | sed '
+		s/^[ \t]*\[[^]]*\][ \t]*//;
+		s/[:., \t][:., \t]*/-/g;
+		s/_/-/g;
+		# *cough*
+		y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/;
+		s/[^a-zA-Z0-9-]//g;')"
+
+	sgml-parser --uri "$URI" --src "$src" --incomplete >/dev/null
+	test_expect_success "$desc" \
+		"test $? = 1"
+}
+
+
+################################################################
+# Check for incompleteness
+
+test_expect_incomplete \
+'Check incomplete comment. (I)' \
+'<!-'
+
+test_expect_incomplete \
+'Check incomplete comment. (II)' \
+'<!-- ... '
+
+test_expect_incomplete \
+'Check incomplete notation. (I)' \
+'<!'
+
+test_expect_incomplete \
+'Check incomplete notation. (II)' \
+'<!D'
+
+test_expect_incomplete \
+'Check incomplete cdata section. (I)' \
+'<![CDATA[ ... '
+
+test_expect_incomplete \
+'Check incomplete cdata section. (II)' \
+'<![CDAT'
+
+test_expect_incomplete \
+'Check incomplete element. (I)' \
+'<elem...'
+
+test_expect_incomplete \
+'Check incomplete element. (II)' \
+'<'
+
+test_expect_incomplete \
+'Check incomplete element end. (I)' \
+'<a></a'
+
+test_expect_incomplete \
+'Check incomplete element end. (II)' \
+'<a></'
+
+test_expect_incomplete \
+'Check incomplete attribute.' \
+'<element attr...'
+
+test_expect_incomplete \
+'Check incomplete attribute value.' \
+'<element attr=...'
+
+test_expect_incomplete \
+'Check incomplete attribute quoted value. (I)' \
+'<element attr="...'
+
+test_expect_incomplete \
+'Check incomplete attribute quoted value. (II)' \
+"<element attr='..."
+
+test_expect_incomplete \
+'Check incomplete processing instruction. (I)' \
+'<?xml'
+
+test_expect_incomplete \
+'Check incomplete processing instruction. (II)' \
+'<?xml attr...'
+
+test_expect_incomplete \
+'Check incomplete notation.' \
+'<!DOCTYPE html PUBLIC ...'
+
+test_expect_incomplete \
+'Check incomplete reference. (I)' \
+'&#123456789'
+
+test_expect_incomplete \
+'Check incomplete reference. (II)' \
+'&amp'
+
+test_expect_incomplete \
+'Check incomplete text.' \
+'plain text is always incomplete (if incomplete)'
+
+test_done

From b09d2c7a3ff94b54fc16976020edeec30056925f Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Sat, 7 Jan 2006 01:49:12 +0100
Subject: [PATCH 21/64] exec_on_slave_terminal(): reduce indentation level and
 use fmem_alloc() instead of mem_alloc().

---
 src/terminal/terminal.c | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/src/terminal/terminal.c b/src/terminal/terminal.c
index 8dd536319..205f148d9 100644
--- a/src/terminal/terminal.c
+++ b/src/terminal/terminal.c
@@ -256,16 +256,16 @@ exec_on_slave_terminal( struct terminal *term,
 			int fg)
 {
 	int data_size = plen + dlen + 1 /* 0 */ + 1 /* fg */ + 2 /* 2 null char */;
-	unsigned char *data = mem_alloc(data_size);
+	unsigned char *data = fmem_alloc(data_size);
 
-	if (data) {
-		data[0] = 0;
-		data[1] = fg;
-		memcpy(data + 2, path, plen + 1);
-		memcpy(data + 2 + plen + 1, delete, dlen + 1);
-		hard_write(term->fdout, data, data_size);
-		mem_free(data);
-	}
+	if (!data) return;
+	
+	data[0] = 0;
+	data[1] = fg;
+	memcpy(data + 2, path, plen + 1);
+	memcpy(data + 2 + plen + 1, delete, dlen + 1);
+	hard_write(term->fdout, data, data_size);
+	fmem_free(data);
 }
 
 void

From df138c775abff99d1fcb9b124c4779fb7f975208 Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Sat, 7 Jan 2006 01:52:35 +0100
Subject: [PATCH 22/64] exec_on_master_terminal(): reduce code redundancy a bit
 and use fmem_alloc() instead of mem_alloc().

---
 src/terminal/terminal.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/src/terminal/terminal.c b/src/terminal/terminal.c
index 205f148d9..6922ed690 100644
--- a/src/terminal/terminal.c
+++ b/src/terminal/terminal.c
@@ -214,7 +214,7 @@ exec_on_master_terminal(struct terminal *term,
 	}
 
 	param_size = plen + dlen + 2 /* 2 null char */ + 1 /* fg */;
-	param = mem_alloc(param_size);
+	param = fmem_alloc(param_size);
 	if (!param) return;
 
 	param[0] = fg;
@@ -225,13 +225,12 @@ exec_on_master_terminal(struct terminal *term,
 
 	blockh = start_thread((void (*)(void *, int)) exec_thread,
 			      param, param_size);
+	fmem_free(param);
 	if (blockh == -1) {
 		if (fg == 1) unblock_itrm(term->fdin);
-		mem_free(param);
 		return;
 	}
 
-	mem_free(param);
 	if (fg == 1) {
 		term->blocked = blockh;
 		set_handlers(blockh,

From 16e2724c12a5987cfeaf6a5cef94d68a55752084 Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Sat, 7 Jan 2006 01:54:44 +0100
Subject: [PATCH 23/64] exec_on_master_terminal(): move code to caller instead.

---
 src/terminal/terminal.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/src/terminal/terminal.c b/src/terminal/terminal.c
index 6922ed690..2657be6c4 100644
--- a/src/terminal/terminal.c
+++ b/src/terminal/terminal.c
@@ -208,11 +208,6 @@ exec_on_master_terminal(struct terminal *term,
 	unsigned char *param;
 	int param_size;
 
-	if (is_blocked() && fg) {
-		unlink(delete);
-		return;
-	}
-
 	param_size = plen + dlen + 2 /* 2 null char */ + 1 /* fg */;
 	param = fmem_alloc(param_size);
 	if (!param) return;
@@ -287,6 +282,11 @@ exec_on_terminal(struct terminal *term, unsigned char *path,
 			return;
 		}
 
+		if (fg && is_blocked()) {
+			unlink(delete);
+			return;
+		}
+
 		exec_on_master_terminal(term,
 					path, strlen(path),
 		 			delete, strlen(delete),

From 2e67ca1df9e415617ef353809951a6a4533b0090 Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Sat, 7 Jan 2006 01:55:18 +0100
Subject: [PATCH 24/64] exec_on_master_terminal(): drop old commented code.

---
 src/terminal/terminal.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/terminal/terminal.c b/src/terminal/terminal.c
index 2657be6c4..528c31cc5 100644
--- a/src/terminal/terminal.c
+++ b/src/terminal/terminal.c
@@ -236,7 +236,7 @@ exec_on_master_terminal(struct terminal *term,
 		set_handlers(term->fdin, NULL, NULL,
 			     (select_handler_T) destroy_terminal,
 			     term);
-		/* block_itrm(term->fdin); */
+
 	} else {
 		set_handlers(blockh, close_handle, NULL,
 			     close_handle, (void *) (long) blockh);

From ee0dc57771ebc11b477ea0934763aef0833ff4e0 Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Sat, 7 Jan 2006 01:57:11 +0100
Subject: [PATCH 25/64] exec_on_master_terminal(): tidy up.

---
 src/terminal/terminal.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/src/terminal/terminal.c b/src/terminal/terminal.c
index 528c31cc5..51ae23790 100644
--- a/src/terminal/terminal.c
+++ b/src/terminal/terminal.c
@@ -205,11 +205,9 @@ exec_on_master_terminal(struct terminal *term,
 			int fg)
 {
 	int blockh;
-	unsigned char *param;
-	int param_size;
+	int param_size = plen + dlen + 2 /* 2 null char */ + 1 /* fg */;
+	unsigned char *param = fmem_alloc(param_size);
 
-	param_size = plen + dlen + 2 /* 2 null char */ + 1 /* fg */;
-	param = fmem_alloc(param_size);
 	if (!param) return;
 
 	param[0] = fg;
@@ -253,7 +251,7 @@ exec_on_slave_terminal( struct terminal *term,
 	unsigned char *data = fmem_alloc(data_size);
 
 	if (!data) return;
-	
+
 	data[0] = 0;
 	data[1] = fg;
 	memcpy(data + 2, path, plen + 1);

From f1c3c90a4f2f224367ff58b001acc1d753662f3c Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 02:02:21 +0100
Subject: [PATCH 26/64] Move line counting tests to own file; simplifies a few
 things

---
 src/dom/test/sgml-parser.c          |  3 +-
 src/dom/test/test-sgml-parser-basic | 39 +----------------
 src/dom/test/test-sgml-parser-lines | 67 +++++++++++++++++++++++++++++
 3 files changed, 70 insertions(+), 39 deletions(-)
 create mode 100755 src/dom/test/test-sgml-parser-lines

diff --git a/src/dom/test/sgml-parser.c b/src/dom/test/sgml-parser.c
index 01d4c95ca..642548ae8 100644
--- a/src/dom/test/sgml-parser.c
+++ b/src/dom/test/sgml-parser.c
@@ -104,7 +104,8 @@ sgml_parser_test_tree(struct dom_stack *stack, struct dom_node *node, void *data
 	struct dom_string *name = get_dom_node_name(node);
 
 	/* Always print the URI for identification. */
-	update_number_of_lines(stack);
+	if (update_number_of_lines(stack))
+		return;
 
 	print_indent(stack);
 	printf("%.*s: %.*s\n",
diff --git a/src/dom/test/test-sgml-parser-basic b/src/dom/test/test-sgml-parser-basic
index 742e81280..a203a78ac 100755
--- a/src/dom/test/test-sgml-parser-basic
+++ b/src/dom/test/test-sgml-parser-basic
@@ -24,7 +24,7 @@ test_output_equals () {
 		y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/;
 		s/[^a-zA-Z0-9-]//g;')"
 
-	sgml-parser --uri "$URI" --src "$src" $@ | sed 's/^  //' > output
+	sgml-parser --uri "$URI" --src "$src" | sed 's/^  //' > output
 	echo "#document: $URI" > expected
 	echo "$out" | sed -n '2,$p' >> expected
 
@@ -247,41 +247,4 @@ element: root
   attribute: ns:attr -> value
   proc-instruction: target -> data'
 
-test_output_equals \
-'Check line numbers. (I)' \
-'<!-- line --> number <one />' \
-'
-1' \
---print-lines
- 
-test_output_equals \
-'Check line numbers. (II)' \
-'<
-line:2
-line:3                      
-=
-"line:5"
-><?xml	
-	line:7="..."
- line:8
- =
- '\''...'\''></line:10>' \
-'
-10' \
---print-lines
- 
-test_output_equals \
-'Check line numbers. (III)' \
-'1
-2
-3
-4
-5
-6
-7
-8' \
-'
-8' \
---print-lines
-
 test_done
diff --git a/src/dom/test/test-sgml-parser-lines b/src/dom/test/test-sgml-parser-lines
new file mode 100755
index 000000000..fac0a433a
--- /dev/null
+++ b/src/dom/test/test-sgml-parser-lines
@@ -0,0 +1,67 @@
+#!/bin/sh
+#
+# Copyright (c) 2005 Jonas Fonseca
+#
+
+test_description='Test the SGML parsers counting of lines
+
+Checks that the SGML parser correctly reports how many lines the
+source has.
+'
+
+. "$TEST_LIB"
+
+test_output_line_numbers () {
+	desc="$1"; shift
+	src="$1"; shift
+	expected_lines="$1"; shift
+
+	URI="test:$(echo "$desc" | sed '
+		s/^[ \t]*\[[^]]*\][ \t]*//;
+		s/[:., \t][:., \t]*/-/g;
+		s/_/-/g;
+		# *cough*
+		y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/;
+		s/[^a-zA-Z0-9-]//g;')"
+
+	lines=$(sgml-parser --uri "$URI" --src "$src" --print-lines)
+
+	test_expect_success "$desc" "test $lines = $expected_lines"
+}
+
+
+################################################################
+# Check line numbers
+
+test_output_line_numbers \
+'Check line numbers. (I)' \
+'<!-- line --> number <one />' \
+1
+ 
+test_output_line_numbers \
+'Check line numbers. (II)' \
+'<
+line:2
+line:3                      
+=
+"line:5"
+><?xml	
+	line:7="..."
+ line:8
+ =
+ '\''...'\''></line:10>' \
+10
+ 
+test_output_line_numbers \
+'Check line numbers. (III)' \
+'1
+2
+3
+4
+5
+6
+7
+8' \
+8
+
+test_done

From a1e512218338713813239244fd5e75229af09d52 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 02:14:45 +0100
Subject: [PATCH 27/64] Drop unnneeded URL argument and simplify test helpers

---
 src/dom/test/test-sgml-parser-incomplete | 10 +---------
 src/dom/test/test-sgml-parser-lines      | 10 +---------
 2 files changed, 2 insertions(+), 18 deletions(-)

diff --git a/src/dom/test/test-sgml-parser-incomplete b/src/dom/test/test-sgml-parser-incomplete
index 697c951d7..2135e94b9 100755
--- a/src/dom/test/test-sgml-parser-incomplete
+++ b/src/dom/test/test-sgml-parser-incomplete
@@ -15,15 +15,7 @@ test_expect_incomplete () {
 	desc="$1"; shift
 	src="$1"; shift
 
-	URI="test:$(echo "$desc" | sed '
-		s/^[ \t]*\[[^]]*\][ \t]*//;
-		s/[:., \t][:., \t]*/-/g;
-		s/_/-/g;
-		# *cough*
-		y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/;
-		s/[^a-zA-Z0-9-]//g;')"
-
-	sgml-parser --uri "$URI" --src "$src" --incomplete >/dev/null
+	sgml-parser --src "$src" --incomplete >/dev/null
 	test_expect_success "$desc" \
 		"test $? = 1"
 }
diff --git a/src/dom/test/test-sgml-parser-lines b/src/dom/test/test-sgml-parser-lines
index fac0a433a..e021d59ff 100755
--- a/src/dom/test/test-sgml-parser-lines
+++ b/src/dom/test/test-sgml-parser-lines
@@ -16,15 +16,7 @@ test_output_line_numbers () {
 	src="$1"; shift
 	expected_lines="$1"; shift
 
-	URI="test:$(echo "$desc" | sed '
-		s/^[ \t]*\[[^]]*\][ \t]*//;
-		s/[:., \t][:., \t]*/-/g;
-		s/_/-/g;
-		# *cough*
-		y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/;
-		s/[^a-zA-Z0-9-]//g;')"
-
-	lines=$(sgml-parser --uri "$URI" --src "$src" --print-lines)
+	lines=$(sgml-parser --src "$src" --print-lines)
 
 	test_expect_success "$desc" "test $lines = $expected_lines"
 }

From 5defc48eb3fe50ce0378ef9dbd5e0b7541aeb2ad Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 04:21:39 +0100
Subject: [PATCH 28/64] Add basic support for requesting error detection; SGML
 scanner part missing

---
 src/dom/scanner.c      |  3 ++-
 src/dom/scanner.h      |  8 ++++++-
 src/dom/select.c       |  2 +-
 src/dom/sgml/parser.c  | 47 ++++++++++++++++++++++++++++++++++++++++--
 src/dom/sgml/parser.h  |  9 ++++++++
 src/dom/sgml/scanner.h |  4 ++++
 6 files changed, 68 insertions(+), 5 deletions(-)

diff --git a/src/dom/scanner.c b/src/dom/scanner.c
index aea8e2d18..c3cd1baa4 100644
--- a/src/dom/scanner.c
+++ b/src/dom/scanner.c
@@ -155,7 +155,7 @@ init_dom_scanner_info(struct dom_scanner_info *scanner_info)
 void
 init_dom_scanner(struct dom_scanner *scanner, struct dom_scanner_info *scanner_info,
 		 struct dom_string *string, int state, int count_lines, int complete,
-		 int check_complete)
+		 int check_complete, int detect_errors)
 {
 	if (!scanner_info->initialized) {
 		init_dom_scanner_info(scanner_info);
@@ -173,6 +173,7 @@ init_dom_scanner(struct dom_scanner *scanner, struct dom_scanner_info *scanner_i
 	scanner->count_lines = !!count_lines;
 	scanner->incomplete = !complete;
 	scanner->check_complete = !!check_complete;
+	scanner->detect_errors = !!detect_errors;
 	scanner->lineno = scanner->count_lines;
 	scanner->info->scan(scanner);
 }
diff --git a/src/dom/scanner.h b/src/dom/scanner.h
index cb9906d44..eabe95746 100644
--- a/src/dom/scanner.h
+++ b/src/dom/scanner.h
@@ -20,6 +20,9 @@ struct dom_scanner_token {
 	/* Some precedence value */
 	int precedence;
 
+	/* The line number; used for error tokens */
+	unsigned int lineno;
+
 	/* The start of the token string and the token length */
 	struct dom_string string;
 };
@@ -93,7 +96,7 @@ struct dom_scanner_info {
 /* Initializes the scanner. */
 void init_dom_scanner(struct dom_scanner *scanner, struct dom_scanner_info *scanner_info,
 		      struct dom_string *string, int state, int count_lines, int complete,
-		      int check_complete);
+		      int check_complete, int detect_error);
 
 /* The number of tokens in the scanners token table:
  * At best it should be big enough to contain properties with space separated
@@ -130,6 +133,9 @@ struct dom_scanner {
 	unsigned int check_complete:1;	/* Only generate complete tokens */
 	unsigned int incomplete:1;	/* The scanned string is incomplete */
 
+	unsigned int detect_errors:1;	/* Check for markup errors */
+	unsigned int found_error;	/* Did we already report this error? */
+
 	unsigned int count_lines:1;	/* Is line counting enbaled? */
 	unsigned int lineno;		/* Line # of the last scanned token */
 
diff --git a/src/dom/select.c b/src/dom/select.c
index 1c53f61c3..a0e5d7644 100644
--- a/src/dom/select.c
+++ b/src/dom/select.c
@@ -391,7 +391,7 @@ parse_dom_select(struct dom_select *select, struct dom_stack *stack,
 	struct dom_scanner scanner;
 	struct dom_select_node sel;
 
-	init_dom_scanner(&scanner, &dom_css_scanner_info, string, 0, 0, 1, 0);
+	init_dom_scanner(&scanner, &dom_css_scanner_info, string, 0, 0, 1, 0, 0);
 
 	memset(&sel, 0, sizeof(sel));
 
diff --git a/src/dom/sgml/parser.c b/src/dom/sgml/parser.c
index 0c5b4a3b4..81227d09a 100644
--- a/src/dom/sgml/parser.c
+++ b/src/dom/sgml/parser.c
@@ -154,6 +154,17 @@ add_sgml_node(struct dom_stack *stack, enum dom_node_type type, struct dom_scann
 
 /* SGML parser main handling: */
 
+static enum sgml_parser_code
+call_sgml_error_function(struct dom_stack *stack, struct dom_scanner_token *token)
+{
+	struct sgml_parser *parser = get_sgml_parser(stack);
+	unsigned int line = get_sgml_parser_line_number(parser);
+
+	assert(parser->error_func);
+
+	return parser->error_func(parser, &token->string, line);
+}
+
 static inline enum sgml_parser_code
 parse_sgml_attributes(struct dom_stack *stack, struct dom_scanner *scanner)
 {
@@ -217,6 +228,17 @@ parse_sgml_attributes(struct dom_stack *stack, struct dom_scanner *scanner)
 		case SGML_TOKEN_INCOMPLETE:
 			return SGML_PARSER_CODE_INCOMPLETE;
 
+		case SGML_TOKEN_ERROR:
+		{
+			enum sgml_parser_code code;
+
+			code = call_sgml_error_function(stack, token);
+			if (code != SGML_PARSER_CODE_OK)
+				return code;
+
+			skip_dom_scanner_token(scanner);
+			break;
+		}
 		default:
 			skip_dom_scanner_token(scanner);
 		}
@@ -326,7 +348,7 @@ parse_sgml_plain(struct dom_stack *stack, struct dom_scanner *scanner)
 				/* The attribute souce is complete. */
 				init_dom_scanner(&attr_scanner, &sgml_scanner_info,
 						 &token->string, SGML_STATE_ELEMENT,
-						 scanner->count_lines, 1, 0);
+						 scanner->count_lines, 1, 0, 0);
 
 				if (dom_scanner_has_tokens(&attr_scanner)) {
 					/* Ignore parser codes from this
@@ -350,6 +372,17 @@ parse_sgml_plain(struct dom_stack *stack, struct dom_scanner *scanner)
 		case SGML_TOKEN_INCOMPLETE:
 			return SGML_PARSER_CODE_INCOMPLETE;
 
+		case SGML_TOKEN_ERROR:
+		{
+			enum sgml_parser_code code;
+
+			code = call_sgml_error_function(stack, token);
+			if (code != SGML_PARSER_CODE_OK)
+				return code;
+
+			skip_dom_scanner_token(scanner);
+			break;
+		}
 		case SGML_TOKEN_SPACE:
 		case SGML_TOKEN_TEXT:
 		default:
@@ -403,11 +436,13 @@ sgml_parsing_push(struct dom_stack *stack, struct dom_node *node, void *data)
 	int count_lines = !!(parser->flags & SGML_PARSER_COUNT_LINES);
 	int complete = !!(parser->flags & SGML_PARSER_COMPLETE);
 	int incremental = !!(parser->flags & SGML_PARSER_INCREMENTAL);
+	int detect_errors = !!(parser->flags & SGML_PARSER_DETECT_ERRORS);
 
 	parsing->depth = parser->stack.depth;
 	get_dom_stack_top(&parser->stack)->immutable = 1;
 	init_dom_scanner(&parsing->scanner, &sgml_scanner_info, &node->string,
-			 SGML_STATE_TEXT, count_lines, complete, incremental);
+			 SGML_STATE_TEXT, count_lines, complete, incremental,
+			 detect_errors);
 }
 
 static void
@@ -494,6 +529,11 @@ get_sgml_parser_line_number(struct sgml_parser *parser)
 
 	assert(pstate->scanner.count_lines && pstate->scanner.lineno);
 
+	if (pstate->scanner.current
+	    && pstate->scanner.current < pstate->scanner.table + DOM_SCANNER_TOKENS
+	    && pstate->scanner.current->type == SGML_TOKEN_ERROR)
+		return pstate->scanner.current->lineno;
+
 	return pstate->scanner.lineno;
 }
 
@@ -553,6 +593,9 @@ init_sgml_parser(enum sgml_parser_type type, enum sgml_document_type doctype,
 		return NULL;
 	}
 
+	if (flags & SGML_PARSER_DETECT_ERRORS)
+		flags |= SGML_PARSER_COUNT_LINES;
+
 	parser->type  = type;
 	parser->flags = flags;
 	parser->info  = get_sgml_info(doctype);
diff --git a/src/dom/sgml/parser.h b/src/dom/sgml/parser.h
index 4764c91b7..b81c6a933 100644
--- a/src/dom/sgml/parser.h
+++ b/src/dom/sgml/parser.h
@@ -7,6 +7,7 @@
 #include "dom/sgml/sgml.h"
 #include "dom/scanner.h"
 
+struct sgml_parser;
 struct string;
 struct uri;
 
@@ -26,6 +27,7 @@ enum sgml_parser_flag {
 	SGML_PARSER_COUNT_LINES	= 1,
 	SGML_PARSER_COMPLETE = 2,
 	SGML_PARSER_INCREMENTAL = 4,
+	SGML_PARSER_DETECT_ERRORS = 8,
 };
 
 struct sgml_parser_state {
@@ -48,6 +50,11 @@ enum sgml_parser_code {
 	SGML_PARSER_CODE_ERROR,
 };
 
+/* If the return code is not SGML_PARSER_CODE_OK the parsing will be ended and
+ * that code will be returned. */
+typedef enum sgml_parser_code
+(*sgml_error_T)(struct sgml_parser *, struct dom_string *, unsigned int);
+
 struct sgml_parser {
 	enum sgml_parser_type type;	/* Stream or tree */
 	enum sgml_parser_flag flags;	/* Flags that control the behaviour */
@@ -57,6 +64,8 @@ struct sgml_parser {
 	struct dom_string uri;		/* The URI of the DOM document */
 	struct dom_node *root;		/* The document root node */
 
+	sgml_error_T error_func;
+
 	struct dom_stack stack;		/* A stack for tracking parsed nodes */
 	struct dom_stack parsing;	/* Used for tracking parsing states */
 };
diff --git a/src/dom/sgml/scanner.h b/src/dom/sgml/scanner.h
index 3857ad78e..fc58b0aa4 100644
--- a/src/dom/sgml/scanner.h
+++ b/src/dom/sgml/scanner.h
@@ -52,6 +52,10 @@ enum sgml_token_type {
 	 * not complete. Only meaningful if scanner->complete is incomplete. */
 	SGML_TOKEN_INCOMPLETE,
 
+	/* A special token for reporting that an error in the markup was found.
+	 * Only in effect when error checking has been requested. */
+	SGML_TOKEN_ERROR,
+
 	/* Token type used internally when scanning to signal that the token
 	 * should not be recorded in the scanners token table. */
 	SGML_TOKEN_SKIP,

From c993a0012e28a957fd68d148809101944f211bf8 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 04:26:08 +0100
Subject: [PATCH 29/64] Add basic support for detection errors while scanning

It mostly uses the checking for incompleteness already in place. Tested
lightly so it will definately need some more work.
---
 src/dom/sgml/scanner.c | 90 +++++++++++++++++++++++++++++++++++++++---
 1 file changed, 84 insertions(+), 6 deletions(-)

diff --git a/src/dom/sgml/scanner.c b/src/dom/sgml/scanner.c
index 041d1b57d..b8544a0c1 100644
--- a/src/dom/sgml/scanner.c
+++ b/src/dom/sgml/scanner.c
@@ -117,6 +117,35 @@ set_sgml_incomplete(struct dom_scanner *scanner, struct dom_scanner_token *token
 	scanner->position = scanner->end;
 }
 
+
+static inline int
+check_sgml_error(struct dom_scanner *scanner)
+{
+	unsigned int found_error = scanner->found_error;
+
+	/* Toggle if we found an error previously. */
+	scanner->found_error = 0;
+
+	return scanner->detect_errors && !found_error;
+}
+
+static struct dom_scanner_token *
+set_sgml_error(struct dom_scanner *scanner, unsigned char *end)
+{
+	struct dom_scanner_token *token = scanner->current;
+
+	assert(!scanner->found_error);
+
+	scanner->found_error = 1;
+
+	token->type = SGML_TOKEN_ERROR;
+	token->lineno = scanner->lineno;
+	set_dom_string(&token->string, scanner->position, end - scanner->position);
+
+	return scanner->found_error ? NULL : scanner->current;
+}
+
+
 /* Text token scanning */
 
 /* I think it is faster to not check the table here --jonas */
@@ -155,9 +184,17 @@ scan_sgml_text_token(struct dom_scanner *scanner, struct dom_scanner_token *toke
 		}
 
 		/* We want the biggest possible text token. */
-		if (check_sgml_incomplete(scanner, string) && !complete) {
-			set_sgml_incomplete(scanner, token);
-			return;
+		if (!complete) {
+			if (check_sgml_incomplete(scanner, string)) {
+				set_sgml_incomplete(scanner, token);
+				return;
+			}
+
+			if (check_sgml_error(scanner)) {
+				token = set_sgml_error(scanner, string);
+				if (!token)
+					return;
+			}
 		}
 
 	} else {
@@ -564,9 +601,44 @@ scan_sgml_element_token(struct dom_scanner *scanner, struct dom_scanner_token *t
 		}
 	}
 
-	if (possibly_incomplete && check_sgml_incomplete(scanner, string)) {
-		set_sgml_incomplete(scanner, token);
-		return;
+	if (possibly_incomplete) {
+		if (check_sgml_incomplete(scanner, string)) {
+			set_sgml_incomplete(scanner, token);
+			return;
+		}
+
+		if (check_sgml_error(scanner)) {
+			unsigned char *end = string;
+
+			DBG("%d %d", type, SGML_TOKEN_NOTATION_DOCTYPE);
+			switch (type) {
+			case SGML_TOKEN_CDATA_SECTION:
+			case SGML_TOKEN_NOTATION_ATTLIST:
+			case SGML_TOKEN_NOTATION_DOCTYPE:
+			case SGML_TOKEN_NOTATION_ELEMENT:
+				if (scanner->position + 9 < end)
+					end = scanner->position + 9;
+				break;
+
+			case SGML_TOKEN_NOTATION_COMMENT:
+				/* Just include the '<!--' part. */
+				if (scanner->position + 4 < end)
+					end = scanner->position + 4;
+				break;
+
+			case SGML_TOKEN_NOTATION_ENTITY:
+				if (scanner->position + 6 < end)
+					end = scanner->position + 6;
+				break;
+
+			default:
+				break;
+			}
+
+			token = set_sgml_error(scanner, end);
+			if (!token)
+				return;
+		}
 	}
 
 	token->type = type;
@@ -607,6 +679,12 @@ scan_sgml_proc_inst_token(struct dom_scanner *scanner, struct dom_scanner_token
 			set_sgml_incomplete(scanner, token);
 			return;
 		}
+
+		if (check_sgml_error(scanner)) {
+			token = set_sgml_error(scanner, scanner->position + 2);
+			if (!token)
+				return;
+		}
 	}
 
 	token->type = SGML_TOKEN_PROCESS_DATA;

From 03ee543e213e0562fde3bb0bc3f274b98d9ec396 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 04:27:08 +0100
Subject: [PATCH 30/64] Make sgml-parser request error detection when passed
 --error option

---
 src/dom/test/sgml-parser.c | 16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)

diff --git a/src/dom/test/sgml-parser.c b/src/dom/test/sgml-parser.c
index 642548ae8..5a4e8cf1a 100644
--- a/src/dom/test/sgml-parser.c
+++ b/src/dom/test/sgml-parser.c
@@ -178,7 +178,8 @@ sgml_parser_test_end(struct dom_stack *stack, struct dom_node *node, void *data)
 {
 	struct sgml_parser *parser = stack->contexts[0]->data;
 
-	if (parser->flags & SGML_PARSER_COUNT_LINES) {
+	if ((parser->flags & SGML_PARSER_COUNT_LINES)
+	    && !(parser->flags & SGML_PARSER_DETECT_ERRORS)) {
 		printf("%d\n", number_of_lines);
 	}
 }
@@ -219,6 +220,15 @@ struct dom_stack_context_info sgml_parser_test_context_info = {
 	}
 };
 
+static enum sgml_parser_code
+sgml_error_function(struct sgml_parser *parser, struct dom_string *string,
+		    unsigned int line_number)
+{
+	printf("error on line %d: %.*s\n",
+	       line_number, string->length, string->string);
+
+	return SGML_PARSER_CODE_OK;
+}
 
 void die(const char *msg, ...)
 {
@@ -285,6 +295,9 @@ main(int argc, char *argv[])
 			flags |= SGML_PARSER_INCREMENTAL;
 			complete = 0;
 
+		} else if (!strcmp(arg, "error")) {
+			flags |= SGML_PARSER_DETECT_ERRORS;
+
 		} else if (!strcmp(arg, "help")) {
 			die(NULL);
 
@@ -296,6 +309,7 @@ main(int argc, char *argv[])
 	parser = init_sgml_parser(SGML_PARSER_STREAM, doctype, &uri, flags);
 	if (!parser) return 1;
 
+	parser->error_func = sgml_error_function;
 	add_dom_stack_context(&parser->stack, NULL, &sgml_parser_test_context_info);
 
 	code = parse_sgml(parser, &source, complete);

From 97f403a9d9b201d079c487bd81fe56306948936e Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 05:15:16 +0100
Subject: [PATCH 31/64] Add a test file for checking detection of errors by the
 SGML parser

---
 src/dom/test/test-sgml-parser-error | 166 ++++++++++++++++++++++++++++
 1 file changed, 166 insertions(+)
 create mode 100755 src/dom/test/test-sgml-parser-error

diff --git a/src/dom/test/test-sgml-parser-error b/src/dom/test/test-sgml-parser-error
new file mode 100755
index 000000000..fe83da3ea
--- /dev/null
+++ b/src/dom/test/test-sgml-parser-error
@@ -0,0 +1,166 @@
+#!/bin/sh
+#
+# Copyright (c) 2005 Jonas Fonseca
+#
+
+test_description='Test SGML parser error reporting
+
+This test checks that the SGML parser will report errors in the source
+given to it.
+'
+
+. "$TEST_LIB"
+
+test_output_error () {
+	desc="$1"; shift
+	src="$1"; shift
+	out="$1"; shift
+
+	sgml-parser --src "$src" --error > output
+	echo "$out" | sed -n '2,$p' > expected
+
+	test_expect_success "$desc" 'cmp output expected' 
+}
+
+
+################################################################
+# Check parsing errors
+
+test_output_error \
+'Check an element error.' \
+'<html' \
+'
+error on line 1: <html'
+
+test_output_error \
+'Check an entity reference error.' \
+'a
+b
+&c' \
+'
+error on line 3: &c'
+
+test_output_error \
+'Check multiple entity reference errors.' \
+'a
+&b&amp;
+c
+&d' \
+'
+error on line 2: &b
+error on line 4: &d'
+
+test_output_error \
+'Check incomplete comment. (I)' \
+'<!-' \
+'
+error on line 1: <!-'
+
+test_output_error \
+'Check incomplete comment. (II)' \
+'<!-- ... ' \
+"
+error on line 1: <!--"
+
+test_output_error \
+'Check incomplete notation. (I)' \
+'<!' \
+'
+error on line 1: <!'
+
+test_output_error \
+'Check incomplete notation. (II)' \
+'<!DOCTYPE ...' \
+'
+error on line 1: <!DOCTYPE'
+
+
+test_output_error \
+'Check incomplete cdata section. (I)' \
+'<![CDATA[ ... ' \
+'
+error on line 1: <![CDATA['
+
+
+test_output_error \
+'Check incomplete cdata section. (II)' \
+'<![CDAT' \
+'
+error on line 1: <![CDAT'
+
+test_output_error \
+'Check incomplete processing instruction. (I)' \
+'<?xml ' \
+'
+error on line 1: <?xml'
+
+test_output_error \
+'Check incomplete processing instruction. (II)' \
+'<?xml-stylesheet attr...' \
+'
+error on line 1: attr...'
+
+test_output_error \
+'Check incomplete reference. (I)' \
+'&#123456789' \
+'
+error on line 1: &#123456789'
+
+test_output_error \
+'Check incomplete reference. (II)' \
+'&amp' \
+'
+error on line 1: &amp'
+
+test_output_error \
+'Check incomplete element. (I)' \
+'<elem...' \
+'
+error on line 1: <elem...'
+
+
+test_output_error \
+'Check incomplete element. (II)' \
+'<' \
+'
+error on line 1: <'
+
+test_output_error \
+'Check incomplete element end. (I)' \
+'<a></a' \
+'
+error on line 1: </a'
+
+test_output_error \
+'Check incomplete element end. (II)' \
+'<a></' \
+'
+error on line 1: </'
+
+test_output_error \
+'Check incomplete attribute.' \
+'<element attr...' \
+'
+error on line 1: attr...'
+
+cat >/dev/null <<FIXME
+test_output_error \
+'Check incomplete attribute value.' \
+'<element attr=...' \
+'
+error on line 1: ...'
+
+test_output_error \
+'Check incomplete attribute quoted value. (I)' \
+'<element attr="...' \
+'
+error on line 1: "...' 
+
+test_output_error \
+'Check incomplete attribute quoted value. (II)' \
+"<element attr='..." \
+"
+error on line 1: '..."
+FIXME
+
+test_done

From 3835bf84494967290c7d5ade71ba1bb0ddcb529a Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 05:18:43 +0100
Subject: [PATCH 32/64] A handful of fixes related to error detection

 - Fix assertion failure by breaking the switch if an error token is next
   when previous was a processing instruction.
 - Fix <!notation parsing by skipping ident chars instead of spaces.
 - Improve checking of processing instruction 'target'-end and what error
   string is generated.
 - For now put all of the processing instruction data in the error token.
 - Remove a DBG()-print.
---
 src/dom/sgml/parser.c  |  7 ++++++-
 src/dom/sgml/scanner.c | 19 +++++++++++++++----
 2 files changed, 21 insertions(+), 5 deletions(-)

diff --git a/src/dom/sgml/parser.c b/src/dom/sgml/parser.c
index 81227d09a..565177e00 100644
--- a/src/dom/sgml/parser.c
+++ b/src/dom/sgml/parser.c
@@ -336,8 +336,13 @@ parse_sgml_plain(struct dom_stack *stack, struct dom_scanner *scanner)
 			if (!token || token->type == SGML_TOKEN_INCOMPLETE)
 				return SGML_PARSER_CODE_INCOMPLETE;
 
-			assert(token->type == SGML_TOKEN_PROCESS_DATA);
+			if (token->type == SGML_TOKEN_ERROR)
+				break;
 
+			assert(token->type == SGML_TOKEN_PROCESS_DATA);
+			/* Fall-through */
+
+		case SGML_TOKEN_PROCESS_DATA:
 			if (add_sgml_proc_instruction(stack, &target, token)
 			    && (target.type == SGML_TOKEN_PROCESS_XML
 			        || target.type == SGML_TOKEN_PROCESS_XML_STYLESHEET)
diff --git a/src/dom/sgml/scanner.c b/src/dom/sgml/scanner.c
index b8544a0c1..875938ba2 100644
--- a/src/dom/sgml/scanner.c
+++ b/src/dom/sgml/scanner.c
@@ -449,7 +449,7 @@ scan_sgml_element_token(struct dom_scanner *scanner, struct dom_scanner_token *t
 				assert(real_length >= 0);
 
 			} else {
-				skip_sgml_space(scanner, &string);
+				scan_sgml(scanner, string, SGML_CHAR_IDENT);
 				type = map_dom_scanner_string(scanner, ident, string, base);
 				if (skip_sgml(scanner, &string, '>', 0)) {
 					/* We found the end. */
@@ -473,7 +473,9 @@ scan_sgml_element_token(struct dom_scanner *scanner, struct dom_scanner_token *t
 			real_length = string - token->string.string;
 			skip_sgml_space(scanner, &string);
 
-			if (is_sgml_space(string[-1])) {
+			/* Make '<?xml ' cause the right kind of error. */
+			if (is_sgml_space(string[-1])
+			    && string < scanner->end) {
 				/* We found the end. */
 				possibly_incomplete = 0;
 			}
@@ -610,7 +612,6 @@ scan_sgml_element_token(struct dom_scanner *scanner, struct dom_scanner_token *t
 		if (check_sgml_error(scanner)) {
 			unsigned char *end = string;
 
-			DBG("%d %d", type, SGML_TOKEN_NOTATION_DOCTYPE);
 			switch (type) {
 			case SGML_TOKEN_CDATA_SECTION:
 			case SGML_TOKEN_NOTATION_ATTLIST:
@@ -631,6 +632,16 @@ scan_sgml_element_token(struct dom_scanner *scanner, struct dom_scanner_token *t
 					end = scanner->position + 6;
 				break;
 
+			case SGML_TOKEN_PROCESS_XML:
+				if (scanner->position + 5 < end)
+					end = scanner->position + 5;
+				break;
+
+			case SGML_TOKEN_PROCESS_XML_STYLESHEET:
+				if (scanner->position + 16 < end)
+					end = scanner->position + 16;
+				break;
+
 			default:
 				break;
 			}
@@ -681,7 +692,7 @@ scan_sgml_proc_inst_token(struct dom_scanner *scanner, struct dom_scanner_token
 		}
 
 		if (check_sgml_error(scanner)) {
-			token = set_sgml_error(scanner, scanner->position + 2);
+			token = set_sgml_error(scanner, string);
 			if (!token)
 				return;
 		}

From 21f9b5913022904d7c11f78ffa1a5b6e70fe3371 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 19:16:51 +0100
Subject: [PATCH 33/64] Use check_whether_file_exists() when encoding a file
 URI string

The file URI encoding is done in translate_url() immediately after the call
to find_uri_protocol() [which also uses check_whether_file_exists()]. So it
only makes sense to use check_whether_file_exists() also here since this is
a potential ambiguous URL from the user.

This fixes handling of files like 'a?b' given the command line, reported by
rrowan. It broke (for the n-th time ;) in CVS version 1.317->1.318 back in
August. The fix preserves the intend of that patch, namely the intend to
optimize the number of calls to file_exists().
---
 src/protocol/uri.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/protocol/uri.c b/src/protocol/uri.c
index 833da11e9..897e4b8e0 100644
--- a/src/protocol/uri.c
+++ b/src/protocol/uri.c
@@ -169,7 +169,7 @@ check_uri_file(unsigned char *name)
 static void
 encode_file_uri_string(struct string *string, unsigned char *uristring)
 {
-	int filenamelen = check_uri_file(uristring);
+	int filenamelen = check_whether_file_exists(uristring);
 
 	encode_uri_string(string, uristring, filenamelen, 0);
 }

From d0be89a16c13d0b228c9dae40bd6cb4137f75b57 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 20:19:37 +0100
Subject: [PATCH 34/64] When requested to open local files with a handler use
 the file in place

Everything else is really stupid, i.e. to go through the callback-infested
download code. A parly fix of bug 238 (caching of local files).
---
 src/session/download.c | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/src/session/download.c b/src/session/download.c
index fa968a762..69db67369 100644
--- a/src/session/download.c
+++ b/src/session/download.c
@@ -985,6 +985,25 @@ tp_open(struct type_query *type_query)
 		return;
 	}
 
+	if (type_query->uri->protocol == PROTOCOL_FILE) {
+		unsigned char *file = get_uri_string(type_query->uri, URI_PATH);
+		unsigned char *handler = NULL;
+
+		if (file) {
+			handler = subst_file(type_query->external_handler, file);
+			mem_free(file);
+		}
+
+		if (handler) {
+			exec_on_terminal(type_query->ses->tab->term,
+					 handler, "", !!type_query->block);
+			mem_free(handler);
+		}
+
+		done_type_query(type_query);
+		return;
+	}
+
 	continue_download(type_query, "");
 }
 

From 9183e0cf04549888980ef64f4ca82b7607f3649d Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Sat, 7 Jan 2006 20:59:34 +0100
Subject: [PATCH 35/64] check_whether_file_exists(): optimize by not using
 strlen() more than needed.

---
 src/protocol/uri.c | 19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)

diff --git a/src/protocol/uri.c b/src/protocol/uri.c
index 897e4b8e0..ad10ec8d7 100644
--- a/src/protocol/uri.c
+++ b/src/protocol/uri.c
@@ -115,28 +115,31 @@ end_with_known_tld(unsigned char *s, int slen)
 	return -1;
 }
 
+/* XXX: this function writes to @name. */
 static int
 check_whether_file_exists(unsigned char *name)
 {
 	/* Check POST_CHAR etc ... */
 	static const unsigned char chars[] = POST_CHAR_S "#?";
 	int i;
+	int namelen = strlen(name);
 
 	if (file_exists(name))
-		return strlen(name);
+		return namelen;
 
 	for (i = 0; i < sizeof(chars) - 1; i++) {
-		unsigned char *pos = strchr(name, chars[i]);
-		int namelen = -1;
-
+		unsigned char *pos = memchr(name, chars[i], namelen);
+		int exists;
+		
 		if (!pos) continue;
 
 		*pos = 0;
-		if (file_exists(name))
-			namelen = strlen(name);
+		exists = file_exists(name);
 		*pos = chars[i];
-
-		if (namelen >= 0) return namelen;
+		
+		if (exists) {
+			return pos - name;
+		}
 	}
 
 	return -1;

From eda6529cedb3948b00ca06d4eddc8d9b62ff3569 Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Sat, 7 Jan 2006 21:00:51 +0100
Subject: [PATCH 36/64] check_uri_file(): optimize by not using strlen() more
 than needed.

---
 src/protocol/uri.c | 14 +++++---------
 1 file changed, 5 insertions(+), 9 deletions(-)

diff --git a/src/protocol/uri.c b/src/protocol/uri.c
index ad10ec8d7..7538ca92e 100644
--- a/src/protocol/uri.c
+++ b/src/protocol/uri.c
@@ -151,21 +151,17 @@ check_uri_file(unsigned char *name)
 	/* Check POST_CHAR etc ... */
 	static const unsigned char chars[] = POST_CHAR_S "#?";
 	int i;
+	int namelen = strlen(name);
 
 	for (i = 0; i < sizeof(chars) - 1; i++) {
-		unsigned char *pos = strchr(name, chars[i]);
-		int namelen;
-
+		unsigned char *pos = memchr(name, chars[i], namelen);
+	
 		if (!pos) continue;
 
-		*pos = 0;
-		namelen = strlen(name);
-		*pos = chars[i];
-
-		return namelen;
+		return pos - name;
 	}
 
-	return strlen(name);
+	return namelen;
 }
 
 /* Encodes URIs without encoding stuff like fragments and query separators. */

From 8f2d4860bc938683f59df72026c97c97596668c8 Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Sat, 7 Jan 2006 21:02:51 +0100
Subject: [PATCH 37/64] Trim trailing whitespaces.

---
 src/protocol/uri.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/src/protocol/uri.c b/src/protocol/uri.c
index 7538ca92e..bbd54c501 100644
--- a/src/protocol/uri.c
+++ b/src/protocol/uri.c
@@ -130,13 +130,13 @@ check_whether_file_exists(unsigned char *name)
 	for (i = 0; i < sizeof(chars) - 1; i++) {
 		unsigned char *pos = memchr(name, chars[i], namelen);
 		int exists;
-		
+
 		if (!pos) continue;
 
 		*pos = 0;
 		exists = file_exists(name);
 		*pos = chars[i];
-		
+
 		if (exists) {
 			return pos - name;
 		}
@@ -155,7 +155,7 @@ check_uri_file(unsigned char *name)
 
 	for (i = 0; i < sizeof(chars) - 1; i++) {
 		unsigned char *pos = memchr(name, chars[i], namelen);
-	
+
 		if (!pos) continue;
 
 		return pos - name;

From 534a16fff14e44de335040f070034e42c298a0d2 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 23:40:21 +0100
Subject: [PATCH 38/64] Improve error detection

---
 src/dom/sgml/scanner.c              | 15 ++++++++++++---
 src/dom/test/test-sgml-parser-error |  2 --
 2 files changed, 12 insertions(+), 5 deletions(-)

diff --git a/src/dom/sgml/scanner.c b/src/dom/sgml/scanner.c
index 875938ba2..3b887dcba 100644
--- a/src/dom/sgml/scanner.c
+++ b/src/dom/sgml/scanner.c
@@ -133,16 +133,25 @@ static struct dom_scanner_token *
 set_sgml_error(struct dom_scanner *scanner, unsigned char *end)
 {
 	struct dom_scanner_token *token = scanner->current;
+	struct dom_scanner_token *next;
 
 	assert(!scanner->found_error);
 
-	scanner->found_error = 1;
+	if (scanner->current >= scanner->table + DOM_SCANNER_TOKENS) {
+		scanner->found_error = 1;
+		next = NULL;
+
+	} else {
+		scanner->current++;
+		next = scanner->current;
+		copy_struct(next, token);
+	}
 
 	token->type = SGML_TOKEN_ERROR;
 	token->lineno = scanner->lineno;
 	set_dom_string(&token->string, scanner->position, end - scanner->position);
 
-	return scanner->found_error ? NULL : scanner->current;
+	return next;
 }
 
 
@@ -609,7 +618,7 @@ scan_sgml_element_token(struct dom_scanner *scanner, struct dom_scanner_token *t
 			return;
 		}
 
-		if (check_sgml_error(scanner)) {
+		if (check_sgml_error(scanner) && string == scanner->end) {
 			unsigned char *end = string;
 
 			switch (type) {
diff --git a/src/dom/test/test-sgml-parser-error b/src/dom/test/test-sgml-parser-error
index fe83da3ea..bcf1fb469 100755
--- a/src/dom/test/test-sgml-parser-error
+++ b/src/dom/test/test-sgml-parser-error
@@ -143,7 +143,6 @@ test_output_error \
 '
 error on line 1: attr...'
 
-cat >/dev/null <<FIXME
 test_output_error \
 'Check incomplete attribute value.' \
 '<element attr=...' \
@@ -161,6 +160,5 @@ test_output_error \
 "<element attr='..." \
 "
 error on line 1: '..."
-FIXME
 
 test_done

From acb1f7e74dfda745d5f9538418838db42435eb63 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sat, 7 Jan 2006 23:51:19 +0100
Subject: [PATCH 39/64] Refactor computation of scanner error string length to
 get_sgml_error_end()

---
 src/dom/sgml/scanner.c | 79 +++++++++++++++++++++++-------------------
 1 file changed, 44 insertions(+), 35 deletions(-)

diff --git a/src/dom/sgml/scanner.c b/src/dom/sgml/scanner.c
index 3b887dcba..a372c9a59 100644
--- a/src/dom/sgml/scanner.c
+++ b/src/dom/sgml/scanner.c
@@ -129,6 +129,48 @@ check_sgml_error(struct dom_scanner *scanner)
 	return scanner->detect_errors && !found_error;
 }
 
+static unsigned char *
+get_sgml_error_end(struct dom_scanner *scanner, enum sgml_token_type type,
+		   unsigned char *end)
+{
+	switch (type) {
+	case SGML_TOKEN_CDATA_SECTION:
+	case SGML_TOKEN_NOTATION_ATTLIST:
+	case SGML_TOKEN_NOTATION_DOCTYPE:
+	case SGML_TOKEN_NOTATION_ELEMENT:
+		if (scanner->position + 9 < end)
+			end = scanner->position + 9;
+		break;
+
+	case SGML_TOKEN_NOTATION_COMMENT:
+		/* Just include the '<!--' part. */
+		if (scanner->position + 4 < end)
+			end = scanner->position + 4;
+		break;
+
+	case SGML_TOKEN_NOTATION_ENTITY:
+		if (scanner->position + 6 < end)
+			end = scanner->position + 6;
+		break;
+
+	case SGML_TOKEN_PROCESS_XML:
+		if (scanner->position + 5 < end)
+			end = scanner->position + 5;
+		break;
+
+	case SGML_TOKEN_PROCESS_XML_STYLESHEET:
+		if (scanner->position + 16 < end)
+			end = scanner->position + 16;
+		break;
+
+	default:
+		break;
+	}
+
+	return end;
+}
+
+
 static struct dom_scanner_token *
 set_sgml_error(struct dom_scanner *scanner, unsigned char *end)
 {
@@ -619,42 +661,9 @@ scan_sgml_element_token(struct dom_scanner *scanner, struct dom_scanner_token *t
 		}
 
 		if (check_sgml_error(scanner) && string == scanner->end) {
-			unsigned char *end = string;
-
-			switch (type) {
-			case SGML_TOKEN_CDATA_SECTION:
-			case SGML_TOKEN_NOTATION_ATTLIST:
-			case SGML_TOKEN_NOTATION_DOCTYPE:
-			case SGML_TOKEN_NOTATION_ELEMENT:
-				if (scanner->position + 9 < end)
-					end = scanner->position + 9;
-				break;
-
-			case SGML_TOKEN_NOTATION_COMMENT:
-				/* Just include the '<!--' part. */
-				if (scanner->position + 4 < end)
-					end = scanner->position + 4;
-				break;
-
-			case SGML_TOKEN_NOTATION_ENTITY:
-				if (scanner->position + 6 < end)
-					end = scanner->position + 6;
-				break;
-
-			case SGML_TOKEN_PROCESS_XML:
-				if (scanner->position + 5 < end)
-					end = scanner->position + 5;
-				break;
-
-			case SGML_TOKEN_PROCESS_XML_STYLESHEET:
-				if (scanner->position + 16 < end)
-					end = scanner->position + 16;
-				break;
-
-			default:
-				break;
-			}
+			unsigned char *end;
 
+			end = get_sgml_error_end(scanner, type, string);
 			token = set_sgml_error(scanner, end);
 			if (!token)
 				return;

From 2f9c406ef1a062dae4b74076bbe9a9a40aab243e Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sun, 8 Jan 2006 03:40:54 +0100
Subject: [PATCH 40/64] Introduce add_to_dom_string() and turn
 init_dom_string() into its user

---
 src/dom/string.h | 21 ++++++++++++++-------
 1 file changed, 14 insertions(+), 7 deletions(-)

diff --git a/src/dom/string.h b/src/dom/string.h
index b4a237800..15a15cfde 100644
--- a/src/dom/string.h
+++ b/src/dom/string.h
@@ -39,21 +39,28 @@ dom_string_ncasecmp(struct dom_string *string1, struct dom_string *string2, size
 	set_dom_string(string1, (string2)->string, (string2)->length)
 
 static inline struct dom_string *
-init_dom_string(struct dom_string *string, unsigned char *str, size_t len)
+add_to_dom_string(struct dom_string *string, unsigned char *str, size_t len)
 {
-	string->string = mem_alloc(len + 1);
-	if (!string->string)
+	unsigned char *newstring;
+
+	newstring = mem_realloc(string->string, string->length + len + 1);
+	if (!newstring)
 		return NULL;
 
-	memcpy(string->string, str, len);
-	string->string[len] = 0;
-	string->length = len;
+	string->string = newstring;
+	memcpy(string->string + string->length, str, len);
+	string->length += len;
+	string->string[string->length] = 0;
+
 	return string;
 }
 
+#define init_dom_string(string, str, len) add_to_dom_string(string, str, len)
+
 #define is_dom_string_set(str) ((str)->string && (str)->length)
 
-#define done_dom_string(str) mem_free((str)->string);
+#define done_dom_string(str) \
+	do { mem_free_set(&(str)->string, NULL); (str)->length = 0; } while (0)
 
 #define isquote(c)	((c) == '"' || (c) == '\'')
 

From bc55cd55cb5c647a50866ea2a020261376268a68 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sun, 8 Jan 2006 03:44:23 +0100
Subject: [PATCH 41/64] Add the basic support for rendering RSS documents

It is very simple for now, so it only displays headlines and doesn't
support a whole lot of RSS flavours.
---
 src/document/dom/renderer.c | 307 ++++++++++++++++++++++++++++++++++--
 src/document/renderer.c     |   8 +-
 src/session/download.c      |   2 +-
 3 files changed, 304 insertions(+), 13 deletions(-)

diff --git a/src/document/dom/renderer.c b/src/document/dom/renderer.c
index 0415d3516..70b022b1d 100644
--- a/src/document/dom/renderer.c
+++ b/src/document/dom/renderer.c
@@ -24,6 +24,7 @@
 #include "document/renderer.h"
 #include "dom/scanner.h"
 #include "dom/sgml/parser.h"
+#include "dom/sgml/rss/rss.h"
 #include "dom/node.h"
 #include "dom/stack.h"
 #include "intl/charsets.h"
@@ -54,6 +55,13 @@ struct dom_renderer {
 	unsigned int find_url:1;
 #endif
 	struct screen_char styles[DOM_NODES];
+
+	/* RSS renderer variables */
+	struct dom_node *channel;
+	struct dom_node_list *items;
+	struct dom_node *item;
+	struct dom_node *node;
+	struct dom_string text;
 };
 
 #define URL_REGEX "(file://|((f|ht|nt)tp(s)?|smb)://[[:alnum:]]+([-@:.]?[[:alnum:]])*\\.[[:alpha:]]{2,4}(:[[:digit:]]+)?)(/(%[[:xdigit:]]{2}|[-_~&=;?.a-z0-9])*)*"
@@ -334,7 +342,8 @@ render_dom_text(struct dom_renderer *renderer, struct screen_char *template,
 	ALIGN_LINK(&(doc)->links, (doc)->nlinks, size)
 
 static inline struct link *
-add_dom_link(struct dom_renderer *renderer, unsigned char *string, int length)
+add_dom_link(struct dom_renderer *renderer, unsigned char *string, int length,
+	     unsigned char *uristring, int urilength)
 {
 	struct document *document = renderer->document;
 	int x = renderer->canvas_x;
@@ -343,7 +352,6 @@ add_dom_link(struct dom_renderer *renderer, unsigned char *string, int length)
 	struct link *link;
 	struct point *point;
 	struct screen_char template;
-	unsigned char *uristring;
 	color_T fgcolor;
 
 	if (!realloc_document_links(document, document->nlinks + 1))
@@ -355,7 +363,7 @@ add_dom_link(struct dom_renderer *renderer, unsigned char *string, int length)
 		return NULL;
 
 	uristring = convert_string(renderer->convert_table,
-				   string, length, document->options.cp,
+				   uristring, urilength, document->options.cp,
 	                           CSM_DEFAULT, NULL, NULL, NULL);
 	if (!uristring) return NULL;
 
@@ -479,7 +487,7 @@ render_dom_node_enhanced_text(struct dom_renderer *renderer, struct dom_node *no
 		string += offset;
 		length -= offset;
 
-		add_dom_link(renderer, string, matchlen);
+		add_dom_link(renderer, string, matchlen, string, matchlen);
 
 		length -= matchlen;
 		string += matchlen;
@@ -601,7 +609,8 @@ render_dom_attribute_source(struct dom_stack *stack, struct dom_node *node, void
 				break;
 			}
 
-			add_dom_link(renderer, value, valuelen - skips);
+			add_dom_link(renderer, value, valuelen - skips,
+				     value, valuelen - skips);
 
 			if (skips > 0) {
 				value += valuelen - skips;
@@ -682,6 +691,272 @@ static struct dom_stack_context_info dom_source_renderer_context_info = {
 };
 
 
+/* DOM RSS Renderer */
+
+static void
+dom_rss_push_element(struct dom_stack *stack, struct dom_node *node, void *data)
+{
+	struct dom_renderer *renderer = stack->current->data;
+
+	assert(node && renderer && renderer->document);
+
+	switch (node->data.element.type) {
+	case RSS_ELEMENT_CHANNEL:
+		/* The stack should have: #document * channel */
+		if (stack->depth != 3)
+			break;
+
+		if (!renderer->channel) {
+			renderer->channel = node;
+		}
+		break;
+
+	case RSS_ELEMENT_ITEM:
+		/* The stack should have: #document * channel item */
+#if 0
+		/* Don't be so strict ... */
+		if (stack->depth != 4)
+			break;
+#endif
+		/* ... but be exclusive. */
+		if (renderer->item)
+			break;
+		add_to_dom_node_list(&renderer->items, node, -1);
+		renderer->item = node;
+		break;
+
+	case RSS_ELEMENT_LINK:
+	case RSS_ELEMENT_DESCRIPTION:
+	case RSS_ELEMENT_TITLE:
+	case RSS_ELEMENT_AUTHOR:
+	case RSS_ELEMENT_PUBDATE:
+		if (!node->parent || renderer->node != node->parent)
+			break;
+
+		renderer->node = node;
+	}
+}
+
+static void
+dom_rss_pop_element(struct dom_stack *stack, struct dom_node *node, void *data)
+{
+	struct dom_renderer *renderer = stack->current->data;
+	struct dom_node_list **list;
+
+	assert(node && renderer && renderer->document);
+
+	switch (node->data.element.type) {
+	case RSS_ELEMENT_ITEM:
+		if (is_dom_string_set(&renderer->text))
+			done_dom_string(&renderer->text);
+		renderer->item = NULL;
+		break;
+
+	case RSS_ELEMENT_LINK:
+	case RSS_ELEMENT_DESCRIPTION:
+	case RSS_ELEMENT_TITLE:
+	case RSS_ELEMENT_AUTHOR:
+	case RSS_ELEMENT_PUBDATE:
+		if (!is_dom_string_set(&renderer->text)
+		    || !node->parent
+		    || renderer->item != node->parent
+		    || renderer->node != node)
+			break;
+
+		/* Replace any child nodes with the normalized text node. */
+		list = get_dom_node_list(node->parent, node);
+		done_dom_node_list(*list);
+		if (is_dom_string_set(&renderer->text)) {
+			if (!add_dom_node(node, DOM_NODE_TEXT, &renderer->text))
+				done_dom_string(&renderer->text);
+		}
+		renderer->node = NULL;
+		break;
+
+	default:
+		break;
+	}
+}
+
+static void
+dom_rss_push_content(struct dom_stack *stack, struct dom_node *node, void *data)
+{
+	struct dom_renderer *renderer = stack->current->data;
+	unsigned char *string = node->string.string;
+	int length = node->string.length;
+
+	assert(node && renderer && renderer->document);
+
+	if (!renderer->node)
+		return;
+
+	if (node->type == DOM_NODE_ENTITY_REFERENCE) {
+		string -= 1;
+		length += 2;
+	}
+
+	if (!is_dom_string_set(&renderer->text)) {
+		init_dom_string(&renderer->text, string, length);
+	} else {
+		add_to_dom_string(&renderer->text, string, length);
+	}
+}
+
+static struct dom_string *
+get_rss_node_text(struct dom_node *node)
+{
+	struct dom_node *child;
+	int index;
+
+	if (!node->data.element.children)
+		return NULL;
+
+	foreach_dom_node (node->data.element.children, child, index) {
+		if (child->type == DOM_NODE_TEXT)
+			return &child->string;
+	}
+
+	return NULL;
+}
+
+static struct dom_node *
+get_rss_child(struct dom_node *parent, enum rss_element_type type)
+{
+	struct dom_node *node;
+	int index;
+
+	if (!parent->data.element.children)
+		return NULL;
+
+	foreach_dom_node (parent->data.element.children, node, index) {
+		if (node->type == DOM_NODE_ELEMENT
+		    && type == node->data.element.type)
+			return node;
+	}
+
+	return NULL;
+}
+
+
+static struct dom_string *
+get_rss_text(struct dom_node *node, enum rss_element_type type)
+{
+	node = get_rss_child(node, type);
+
+	return node ? get_rss_node_text(node) : NULL;
+}
+
+static void
+render_rss_item(struct dom_renderer *renderer, struct dom_node *item)
+{
+	struct dom_string *title  = get_rss_text(item, RSS_ELEMENT_TITLE);
+	struct dom_string *link   = get_rss_text(item, RSS_ELEMENT_LINK);
+	struct dom_string *author = get_rss_text(item, RSS_ELEMENT_AUTHOR);
+	struct dom_string *date   = get_rss_text(item, RSS_ELEMENT_PUBDATE);
+
+	if (title && is_dom_string_set(title)) {
+		render_dom_text(renderer, &renderer->styles[DOM_NODE_ELEMENT],
+				title->string, title->length);
+	}
+
+	if (link && is_dom_string_set(link)) {
+		X(renderer)++;
+		add_dom_link(renderer, "[link]", 6, link->string, link->length);
+	}
+
+	/* New line, and indent */
+	Y(renderer)++;
+	X(renderer) = 0;
+
+	if (author && is_dom_string_set(author)) {
+		render_dom_text(renderer, &renderer->styles[DOM_NODE_COMMENT],
+				author->string, author->length);
+	}
+
+	if (date && is_dom_string_set(date)) {
+		if (author && is_dom_string_set(author)) {
+			render_dom_text(renderer, &renderer->styles[DOM_NODE_COMMENT],
+					" - ", 3);
+		}
+
+		render_dom_text(renderer, &renderer->styles[DOM_NODE_COMMENT],
+				date->string, date->length);
+	}
+
+	if ((author && is_dom_string_set(author))
+	    || (date && is_dom_string_set(date))) {
+		/* New line, and indent */
+		Y(renderer)++;
+		X(renderer) = 0;
+	}
+}
+
+static void
+dom_rss_pop_document(struct dom_stack *stack, struct dom_node *root, void *data)
+{
+	struct dom_renderer *renderer = stack->current->data;
+
+	if (!renderer->channel)
+		return;
+
+	render_rss_item(renderer, renderer->channel);
+
+	if (renderer->items) {
+		struct dom_node *node;
+		int index;
+
+		foreach_dom_node (renderer->items, node, index) {
+			Y(renderer)++;
+			X(renderer) = 0;
+			render_rss_item(renderer, node);
+		}
+	}
+
+	if (is_dom_string_set(&renderer->text))
+		done_dom_string(&renderer->text);
+	mem_free_if(renderer->items);
+
+	done_dom_node(root);
+}
+
+
+static struct dom_stack_context_info dom_rss_renderer_context_info = {
+	/* Object size: */			0,
+	/* Push: */
+	{
+		/*				*/ NULL,
+		/* DOM_NODE_ELEMENT		*/ dom_rss_push_element,
+		/* DOM_NODE_ATTRIBUTE		*/ NULL,
+		/* DOM_NODE_TEXT		*/ dom_rss_push_content,
+		/* DOM_NODE_CDATA_SECTION	*/ dom_rss_push_content,
+		/* DOM_NODE_ENTITY_REFERENCE	*/ dom_rss_push_content,
+		/* DOM_NODE_ENTITY		*/ NULL,
+		/* DOM_NODE_PROC_INSTRUCTION	*/ NULL,
+		/* DOM_NODE_COMMENT		*/ NULL,
+		/* DOM_NODE_DOCUMENT		*/ NULL,
+		/* DOM_NODE_DOCUMENT_TYPE	*/ NULL,
+		/* DOM_NODE_DOCUMENT_FRAGMENT	*/ NULL,
+		/* DOM_NODE_NOTATION		*/ NULL,
+	},
+	/* Pop: */
+	{
+		/*				*/ NULL,
+		/* DOM_NODE_ELEMENT		*/ dom_rss_pop_element,
+		/* DOM_NODE_ATTRIBUTE		*/ NULL,
+		/* DOM_NODE_TEXT		*/ NULL,
+		/* DOM_NODE_CDATA_SECTION	*/ NULL,
+		/* DOM_NODE_ENTITY_REFERENCE	*/ NULL,
+		/* DOM_NODE_ENTITY		*/ NULL,
+		/* DOM_NODE_PROC_INSTRUCTION	*/ NULL,
+		/* DOM_NODE_COMMENT		*/ NULL,
+		/* DOM_NODE_DOCUMENT		*/ dom_rss_pop_document,
+		/* DOM_NODE_DOCUMENT_TYPE	*/ NULL,
+		/* DOM_NODE_DOCUMENT_FRAGMENT	*/ NULL,
+		/* DOM_NODE_NOTATION		*/ NULL,
+	}
+};
+
+
 /* Shared multiplexor between renderers */
 void
 render_dom_document(struct cache_entry *cached, struct document *document,
@@ -692,14 +967,13 @@ render_dom_document(struct cache_entry *cached, struct document *document,
 	struct conv_table *convert_table;
 	struct sgml_parser *parser;
 	enum sgml_document_type doctype;
+ 	enum sgml_parser_type parser_type;
 	unsigned char *string = struri(cached->uri);
 	size_t length = strlen(string);
 	struct dom_string uri = INIT_DOM_STRING(string, length);
 	struct dom_string source = INIT_DOM_STRING(buffer->source, buffer->length);
 	enum sgml_parser_code code;
 
-	assert(document->options.plain);
-
 	convert_table = get_convert_table(head, document->options.cp,
 					  document->options.assume_cp,
 					  &document->cp,
@@ -710,6 +984,11 @@ render_dom_document(struct cache_entry *cached, struct document *document,
 
 	document->bgcolor = document->options.default_bg;
 
+	if (document->options.plain)
+		parser_type = SGML_PARSER_STREAM;
+	else
+		parser_type = SGML_PARSER_TREE;
+
 	/* FIXME: Refactor the doctype lookup. */
 	if (!strcasecmp("application/rss+xml", cached->content_type)) {
 		doctype = SGML_DOCTYPE_RSS;
@@ -730,11 +1009,17 @@ render_dom_document(struct cache_entry *cached, struct document *document,
 		doctype = SGML_DOCTYPE_HTML;
 	}
 
-	parser = init_sgml_parser(SGML_PARSER_STREAM, doctype, &uri, 0);
-	if (!parser) return;
+	parser = init_sgml_parser(parser_type, doctype, &uri, 0);
+  	if (!parser) return;
+  
+	if (document->options.plain) {
+		add_dom_stack_context(&parser->stack, &renderer,
+				      &dom_source_renderer_context_info);
 
-	add_dom_stack_context(&parser->stack, &renderer,
-			      &dom_source_renderer_context_info);
+	} else if (doctype == SGML_DOCTYPE_RSS) {
+		add_dom_stack_context(&parser->stack, &renderer,
+				      &dom_rss_renderer_context_info);
+	}
 
 	/* FIXME: When rendering this way we don't really care about the code.
 	 * However, it will be useful when we will be able to also
diff --git a/src/document/renderer.c b/src/document/renderer.c
index 6c3a1d81a..ad458caa4 100644
--- a/src/document/renderer.c
+++ b/src/document/renderer.c
@@ -254,7 +254,13 @@ render_encoded_document(struct cache_entry *cached, struct document *document)
 			render_plain_document(cached, document, &buffer);
 
 	} else {
-		render_html_document(cached, document, &buffer);
+#ifdef CONFIG_DOM
+		if (cached->content_type
+		    && (!strlcasecmp("application/rss+xml", 19, cached->content_type, -1)))
+			render_dom_document(cached, document, &buffer);
+		else
+#endif
+			render_html_document(cached, document, &buffer);
 	}
 
 	if (encoding != ENCODING_NONE) {
diff --git a/src/session/download.c b/src/session/download.c
index 69db67369..b0c0ef741 100644
--- a/src/session/download.c
+++ b/src/session/download.c
@@ -1171,7 +1171,7 @@ struct {
 	{ "application/xhtml+xml",	0 }, /* RFC 3236 */
 #if CONFIG_DOM
 	{ "application/docbook+xml",	1 },
-	{ "application/rss+xml",	1 },
+	{ "application/rss+xml",	0 },
 	{ "application/xbel+xml",	1 },
 	{ "application/xbel",		1 },
 	{ "application/x-xbel",		1 },

From 38fe0f729299362db7f506e29ab460d081d389b7 Mon Sep 17 00:00:00 2001
From: Miciah Dashiel Butler Masters <miciah@myrealbox.com>
Date: Sun, 8 Jan 2006 07:34:59 +0000
Subject: [PATCH 42/64] Fix add_document_to_string to actually add stuff

Revision 1.147 (14 months, 2 weeks ago) broke the logic so that only
frame characters were added. This fixes that change, getting us halfway
back to a working current_document_formatted function for Lua.
---
 src/viewer/dump/dump.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/src/viewer/dump/dump.c b/src/viewer/dump/dump.c
index 3e038502c..e88115fbf 100644
--- a/src/viewer/dump/dump.c
+++ b/src/viewer/dump/dump.c
@@ -335,8 +335,9 @@ add_document_to_string(struct string *string, struct document *document)
 			if (!isscreensafe(data)) {
 				white++;
 				continue;
-			} else if (frame && data >= 176 && data < 224) {
-				data = frame_dumb[data - 176];
+			} else {
+				if (frame && data >= 176 && data < 224)
+					data = frame_dumb[data - 176];
 
 				if (data <= ' ') {
 					/* Count spaces. */

From 55253399013924e6d8a2ddc8589341b80496e093 Mon Sep 17 00:00:00 2001
From: Miciah Dashiel Butler Masters <miciah@myrealbox.com>
Date: Sun, 8 Jan 2006 07:43:44 +0000
Subject: [PATCH 43/64] Fix add_document_to_string to add all chars in a line,
 not just the first

Revision 1.82 of src/scripting/lua/core.c and revision 1.42
of src/viewer/dump/dump.c (2 years, 2 months ago), which moved
add_document_to_string from the former to the latter, introduced
a logic error: the local variable pos was set in the outer loop,
which iterates thru the lines in the document, whereas it should be
(and had been) set in the inner loop, which iterates thru the columns.

This brings us all the way back to a working current_document_formatted
for Lua scripts.
---
 src/viewer/dump/dump.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/viewer/dump/dump.c b/src/viewer/dump/dump.c
index e88115fbf..9b38ab2de 100644
--- a/src/viewer/dump/dump.c
+++ b/src/viewer/dump/dump.c
@@ -324,11 +324,11 @@ add_document_to_string(struct string *string, struct document *document)
 	if_assert_failed return NULL;
 
 	for (y = 0; y < document->height; y++) {
-		struct screen_char *pos = document->data[y].chars;
 		int white = 0;
 		int x;
 
 		for (x = 0; x < document->data[y].length; x++) {
+			struct screen_char *pos = &document->data[y].chars[x];
 			unsigned char data = pos->data;
 			unsigned int frame = (pos->attr & SCREEN_ATTR_FRAME);
 

From 67a89e1454303623c5025c85d9f884d8c639c5e6 Mon Sep 17 00:00:00 2001
From: Russ Rowan <neuromancer@HAL-9000.(none)>
Date: Sun, 8 Jan 2006 05:15:08 -0500
Subject: [PATCH 44/64] Added a document helper script: code2doc.  It's ugly
 and mean, but it gets the job done.

---
 doc/tools/code2doc | 43 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)
 create mode 100755 doc/tools/code2doc

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
new file mode 100755
index 000000000..f5865b338
--- /dev/null
+++ b/doc/tools/code2doc
@@ -0,0 +1,43 @@
+#!/usr/bin/perl -w
+use strict;
+use warnings;
+use diagnostics;
+
+print "Usage: $0 [FILE]\n\tParses [FILE], outputing the result to [FILE].out.\n"
+	and exit if not @ARGV;
+
+my ($input)  = @ARGV;
+my ($output) = $input . '.out';
+my ($found, $start);
+print "CopyleftŠ 2006, Russ Rowan (See `COPYING')\n" and exit if $input eq '-v';
+open FILEIN,  "<$input"  or print "File `$input' was not found.\n"     and exit;
+open FILEOUT, ">$output" or print "Cannot open $output for writing.\n" and exit;
+print "Processing $input...\n";
+while (<FILEIN>)
+{
+	if ($found)
+	{
+		next if $_ =~ /^\s+\*\s$/;
+		if ($_ =~ /^\s\*+\/$/ or $_ !~ /^\s/) { $found = undef; next; }
+		$_ =~ s/^(\s*)\s\*\s/$1/;
+		$found = 'sorta' if $_ =~ s/\s*\*\/$//;
+	}
+	elsif ($_ =~ /^\s*\/\*\*\s(.*)/)
+	{
+		print FILEOUT "\n\n" if $start;
+		$_ = $1;
+		if ($_ =~ s/\s*\*\/$//) { $found = 'sorta'; } else { $found = $.; }
+		my $dash;  for (my $x = 0; $x < length($_); $x++) { $dash .= '-'; }
+		$_ = "$_\n$dash\n";
+	}
+	elsif ($_ =~ /\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s*\*\/$/)
+	{
+		$_ =~ s/\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s*\*\/$/id[$1]::\n  $2/;
+		$found = 'sorta';
+	}
+	next if not $found;  $found = undef if $found eq 'sorta';
+	print FILEOUT $_ and $start = 1;
+}
+close FILEOUT;
+close FILEIN;
+print "$output completed.\n";

From 92be45d1be3633533971a8a6a292876d7d4b3cd3 Mon Sep 17 00:00:00 2001
From: Russ Rowan <neuromancer@HAL-9000.(none)>
Date: Sun, 8 Jan 2006 12:56:15 -0500
Subject: [PATCH 45/64] Don't remove all blank lines.  Just the first of a
 comment block.

---
 doc/tools/code2doc | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index f5865b338..ff22bb7fc 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -8,7 +8,7 @@ print "Usage: $0 [FILE]\n\tParses [FILE], outputing the result to [FILE].out.\n"
 
 my ($input)  = @ARGV;
 my ($output) = $input . '.out';
-my ($found, $start);
+my ($found, $start, $first);
 print "CopyleftŠ 2006, Russ Rowan (See `COPYING')\n" and exit if $input eq '-v';
 open FILEIN,  "<$input"  or print "File `$input' was not found.\n"     and exit;
 open FILEOUT, ">$output" or print "Cannot open $output for writing.\n" and exit;
@@ -17,10 +17,11 @@ while (<FILEIN>)
 {
 	if ($found)
 	{
-		next if $_ =~ /^\s+\*\s$/;
+		if ($_ =~ /^\s+\*\s$/) { next if $first; $_ =~ s/\*// if not $first; }
 		if ($_ =~ /^\s\*+\/$/ or $_ !~ /^\s/) { $found = undef; next; }
 		$_ =~ s/^(\s*)\s\*\s/$1/;
 		$found = 'sorta' if $_ =~ s/\s*\*\/$//;
+		$first = undef;
 	}
 	elsif ($_ =~ /^\s*\/\*\*\s(.*)/)
 	{
@@ -29,6 +30,7 @@ while (<FILEIN>)
 		if ($_ =~ s/\s*\*\/$//) { $found = 'sorta'; } else { $found = $.; }
 		my $dash;  for (my $x = 0; $x < length($_); $x++) { $dash .= '-'; }
 		$_ = "$_\n$dash\n";
+		$first = 1;
 	}
 	elsif ($_ =~ /\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s*\*\/$/)
 	{

From 33f373fbd53f53c69dd21dfd40773d8b43f00463 Mon Sep 17 00:00:00 2001
From: Russ Rowan <neuromancer@HAL-9000.(none)>
Date: Sun, 8 Jan 2006 13:05:11 -0500
Subject: [PATCH 46/64] Get rid of extraneous spaces left in the output.  I
 have no idea if asciidoc is sensitive enough to fail on these.

---
 doc/tools/code2doc | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index ff22bb7fc..e572441d9 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -17,7 +17,7 @@ while (<FILEIN>)
 {
 	if ($found)
 	{
-		if ($_ =~ /^\s+\*\s$/) { next if $first; $_ =~ s/\*// if not $first; }
+		if ($_ =~ /^\s+\*\s$/) { next if $first; $_ =~ s/\s\*// if not $first; }
 		if ($_ =~ /^\s\*+\/$/ or $_ !~ /^\s/) { $found = undef; next; }
 		$_ =~ s/^(\s*)\s\*\s/$1/;
 		$found = 'sorta' if $_ =~ s/\s*\*\/$//;
@@ -32,9 +32,9 @@ while (<FILEIN>)
 		$_ = "$_\n$dash\n";
 		$first = 1;
 	}
-	elsif ($_ =~ /\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s*\*\/$/)
+	elsif ($_ =~ /\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/)
 	{
-		$_ =~ s/\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s*\*\/$/id[$1]::\n  $2/;
+		$_ =~ s/\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/id[$1]::\n  $2/;
 		$found = 'sorta';
 	}
 	next if not $found;  $found = undef if $found eq 'sorta';

From 8fe7eb03577e4c3d0baf7291a0896fee06910220 Mon Sep 17 00:00:00 2001
From: Laurent MONIN <zas@norz.org>
Date: Sun, 8 Jan 2006 19:06:46 +0100
Subject: [PATCH 47/64] check_uri_file(): use strcspn().

---
 src/protocol/uri.c | 12 +-----------
 1 file changed, 1 insertion(+), 11 deletions(-)

diff --git a/src/protocol/uri.c b/src/protocol/uri.c
index bbd54c501..88020dcc0 100644
--- a/src/protocol/uri.c
+++ b/src/protocol/uri.c
@@ -150,18 +150,8 @@ check_uri_file(unsigned char *name)
 {
 	/* Check POST_CHAR etc ... */
 	static const unsigned char chars[] = POST_CHAR_S "#?";
-	int i;
-	int namelen = strlen(name);
 
-	for (i = 0; i < sizeof(chars) - 1; i++) {
-		unsigned char *pos = memchr(name, chars[i], namelen);
-
-		if (!pos) continue;
-
-		return pos - name;
-	}
-
-	return namelen;
+	return strcspn(name, chars);
 }
 
 /* Encodes URIs without encoding stuff like fragments and query separators. */

From 3529b35da09c8c33fb078742c320e1213abc3aa1 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sun, 8 Jan 2006 20:02:31 +0100
Subject: [PATCH 48/64] Sneak in an extra newline after section headers

---
 doc/tools/code2doc | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index e572441d9..7c404d00a 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -29,7 +29,7 @@ while (<FILEIN>)
 		$_ = $1;
 		if ($_ =~ s/\s*\*\/$//) { $found = 'sorta'; } else { $found = $.; }
 		my $dash;  for (my $x = 0; $x < length($_); $x++) { $dash .= '-'; }
-		$_ = "$_\n$dash\n";
+		$_ = "$_\n$dash\n\n";
 		$first = 1;
 	}
 	elsif ($_ =~ /\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/)

From 97b0635d1ee5f9cee08e5ab9446d031142be6e10 Mon Sep 17 00:00:00 2001
From: Russ Rowan <neuromancer@HAL-9000.(none)>
Date: Sun, 8 Jan 2006 15:37:03 -0500
Subject: [PATCH 49/64] Don't underline leading comment lines that contain ::.

---
 doc/tools/code2doc | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index 7c404d00a..ae5b5cc37 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -28,8 +28,12 @@ while (<FILEIN>)
 		print FILEOUT "\n\n" if $start;
 		$_ = $1;
 		if ($_ =~ s/\s*\*\/$//) { $found = 'sorta'; } else { $found = $.; }
-		my $dash;  for (my $x = 0; $x < length($_); $x++) { $dash .= '-'; }
-		$_ = "$_\n$dash\n\n";
+		if ($_ =~ /::/) { $_ = "$_\n\n"; }
+		else
+		{
+			my $dash; for (my $x = 0; $x < length($_); $x++) { $dash .= '-'; }
+			$_ = "$_\n$dash\n\n";
+		}
 		$first = 1;
 	}
 	elsif ($_ =~ /\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/)

From 877ed348beaadf4362d5e41ba9f6eac60f95c9b2 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sun, 8 Jan 2006 21:46:21 +0100
Subject: [PATCH 50/64] Simplify the doc Makefile tremendously

---
 doc/Makefile | 41 ++++++++++++-----------------------------
 1 file changed, 12 insertions(+), 29 deletions(-)

diff --git a/doc/Makefile b/doc/Makefile
index 08641d124..30e96f2ec 100644
--- a/doc/Makefile
+++ b/doc/Makefile
@@ -56,50 +56,33 @@ IMPORT_FEATURES_CONF	= $(top_srcdir)/doc/tools/import-features.conf
 MAKE_ELINKS_MANPAGE	= $(top_srcdir)/doc/tools/make-elinks-manpage
 MAKE_ELINKSKEYS_MANPAGE	= $(top_srcdir)/doc/tools/make-elinkskeys-manpage
 
-ifeq ($(CONFIG_ASCIIDOC),yes)
-HTML_DOCS_WITH_ASCIIDOC = \
+HTML_DOCS-$(CONFIG_ASCIIDOC) += \
  $(HTML_DIR)/elinks.1.html \
  $(HTML_DIR)/elinkskeys.5.html \
  $(HTML_DIR)/hacking.html \
  $(HTML_DIR)/manual.html
-endif
-
 
 # Only jw is used for generating PDF.
-ifeq ($(CONFIG_XMLTO),yes)
-HTML_DOCS_WITH_XMLTO = \
+HTML_DOCS-$(CONFIG_XMLTO) += \
  $(HTML_DIR)/manual.html-chunked
 
-MAN_DOCS_WITH_XMLTO = \
+MAN_DOCS-$(CONFIG_XMLTO) += \
  $(MAN_DIR)/man1/elinks.1.in \
  $(MAN_DIR)/man5/elinkskeys.5
-endif
 
 # Only jw is used for generating PDF.
-ifeq ($(CONFIG_JW),yes)
-PDF_DOCS_WITH_JW = \
+PDF_DOCS-$(CONFIG_JW) += \
  $(PDF_DIR)/manual.pdf
-endif
 
-ifeq ($(CONFIG_POD2HTML),yes)
-HTML_DOCS_WITH_POD2HTML = \
+HTML_DOCS-$(CONFIG_POD2HTML) += \
  $(HTML_DIR)/perl.html \
  $(HTML_DIR)/perl-hooks.html
-endif
 
-MAN_DOCS_WITH_SHELL = $(MAN_DIR)/man5/elinks.conf.5
+MAN_DOCS += $(MAN_DIR)/man5/elinks.conf.5
 
-MAN_DOCS = \
- $(MAN_DOCS_WITH_SHELL)
- $(MAN_DOCS_WITH_ASCIIDOC)
-
-HTML_DOCS = \
- $(HTML_DOCS_WITH_ASCIIDOC) \
- $(HTML_DOCS_WITH_POD2HTML) \
- $(HTML_DOCS_WITH_JW)
-
-PDF_DOCS = \
- $(PDF_DOCS_WITH_JW)
+MAN_DOCS  += $(MAN_DOCS-yes)
+HTML_DOCS += $(HTML_DOCS-yes)
+PDF_DOCS  += $(PDF_DOCS-yes)
 
 html-asciidoc-yes: doc-dirs $(HTML_DOCS_WITH_ASCIIDOC)
 html-asciidoc-no:
@@ -116,9 +99,9 @@ man-xmlto-no:
 pdf-jw-yes: doc-dirs $(PDF_DOCS_WITH_JW)
 pdf-jw-no:
 
-man-docs: man-xmlto-$(CONFIG_XMLTO) $(MAN_DOCS_WITH_SHELL)
-html-docs: html-asciidoc-$(CONFIG_ASCIIDOC) html-xmlto-$(CONFIG_XMLTO) html-pod2html-$(CONFIG_POD2HTML)
-pdf-docs: pdf-jw-$(CONFIG_JW)
+man-docs: $(MAN_DOCS)
+html-docs: $(HTML_DOCS)
+pdf-docs: $(PDF_DOCS)
 
 all-docs: man-docs html-docs pdf-docs
 

From 4fe54a9d680c2eaf4045de05c6d3f84dc2b266d3 Mon Sep 17 00:00:00 2001
From: Russ Rowan <neuromancer@HAL-9000.(none)>
Date: Sun, 8 Jan 2006 15:54:00 -0500
Subject: [PATCH 51/64] Added another blank line between each block and one
 between each id:[*]:: line for readability.

---
 doc/tools/code2doc | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index ae5b5cc37..da27680ea 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -8,7 +8,7 @@ print "Usage: $0 [FILE]\n\tParses [FILE], outputing the result to [FILE].out.\n"
 
 my ($input)  = @ARGV;
 my ($output) = $input . '.out';
-my ($found, $start, $first);
+my ($found, $start, $first, $gotone);
 print "CopyleftŠ 2006, Russ Rowan (See `COPYING')\n" and exit if $input eq '-v';
 open FILEIN,  "<$input"  or print "File `$input' was not found.\n"     and exit;
 open FILEOUT, ">$output" or print "Cannot open $output for writing.\n" and exit;
@@ -25,7 +25,7 @@ while (<FILEIN>)
 	}
 	elsif ($_ =~ /^\s*\/\*\*\s(.*)/)
 	{
-		print FILEOUT "\n\n" if $start;
+		print FILEOUT "\n\n\n" if $start;
 		$_ = $1;
 		if ($_ =~ s/\s*\*\/$//) { $found = 'sorta'; } else { $found = $.; }
 		if ($_ =~ /::/) { $_ = "$_\n\n"; }
@@ -38,9 +38,11 @@ while (<FILEIN>)
 	}
 	elsif ($_ =~ /\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/)
 	{
+		print FILEOUT "\n" if $gotone;
 		$_ =~ s/\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/id[$1]::\n  $2/;
-		$found = 'sorta';
+		$found = 'sorta'; $gotone = 1;
 	}
+	else { $gotone = undef; }
 	next if not $found;  $found = undef if $found eq 'sorta';
 	print FILEOUT $_ and $start = 1;
 }

From 905aa95a6e73d9a2021de776106443d7a44f7905 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sun, 8 Jan 2006 21:57:55 +0100
Subject: [PATCH 52/64] Simplify some more

---
 doc/Makefile | 48 ++++++++++++++++--------------------------------
 1 file changed, 16 insertions(+), 32 deletions(-)

diff --git a/doc/Makefile b/doc/Makefile
index 30e96f2ec..c8d8502f7 100644
--- a/doc/Makefile
+++ b/doc/Makefile
@@ -57,51 +57,35 @@ MAKE_ELINKS_MANPAGE	= $(top_srcdir)/doc/tools/make-elinks-manpage
 MAKE_ELINKSKEYS_MANPAGE	= $(top_srcdir)/doc/tools/make-elinkskeys-manpage
 
 HTML_DOCS-$(CONFIG_ASCIIDOC) += \
- $(HTML_DIR)/elinks.1.html \
- $(HTML_DIR)/elinkskeys.5.html \
- $(HTML_DIR)/hacking.html \
- $(HTML_DIR)/manual.html
+ elinks.1.html \
+ elinkskeys.5.html \
+ hacking.html \
+ manual.html
 
-# Only jw is used for generating PDF.
 HTML_DOCS-$(CONFIG_XMLTO) += \
- $(HTML_DIR)/manual.html-chunked
+ manual.html-chunked
+
+HTML_DOCS-$(CONFIG_POD2HTML) += \
+ perl.html \
+ perl-hooks.html
 
 MAN_DOCS-$(CONFIG_XMLTO) += \
- $(MAN_DIR)/man1/elinks.1.in \
- $(MAN_DIR)/man5/elinkskeys.5
+ man1/elinks.1.in \
+ man5/elinkskeys.5
 
 # Only jw is used for generating PDF.
 PDF_DOCS-$(CONFIG_JW) += \
- $(PDF_DIR)/manual.pdf
+ manual.pdf
 
-HTML_DOCS-$(CONFIG_POD2HTML) += \
- $(HTML_DIR)/perl.html \
- $(HTML_DIR)/perl-hooks.html
-
-MAN_DOCS += $(MAN_DIR)/man5/elinks.conf.5
+MAN_DOCS += man5/elinks.conf.5
 
 MAN_DOCS  += $(MAN_DOCS-yes)
 HTML_DOCS += $(HTML_DOCS-yes)
 PDF_DOCS  += $(PDF_DOCS-yes)
 
-html-asciidoc-yes: doc-dirs $(HTML_DOCS_WITH_ASCIIDOC)
-html-asciidoc-no:
-
-html-pod2html-yes: doc-dirs $(HTML_DOCS_WITH_POD2HTML)
-html-pod2html-no:
-
-html-xmlto-yes: doc-dirs $(HTML_DOCS_WITH_XMLTO)
-html-xmlto-no:
-
-man-xmlto-yes: doc-dirs $(MAN_DOCS_WITH_XMLTO)
-man-xmlto-no:
-
-pdf-jw-yes: doc-dirs $(PDF_DOCS_WITH_JW)
-pdf-jw-no:
-
-man-docs: $(MAN_DOCS)
-html-docs: $(HTML_DOCS)
-pdf-docs: $(PDF_DOCS)
+man-docs:  $(addprefix $(MAN_DIR)/,$(MAN_DOCS))
+html-docs: $(addprefix $(HTML_DIR)/,$(HTML_DOCS))
+pdf-docs:  $(addprefix $(PDF_DIR)/,$(PDF_DOCS))
 
 all-docs: man-docs html-docs pdf-docs
 

From e15c9d46111e0006ca261a65e138e089f9b525b5 Mon Sep 17 00:00:00 2001
From: Russ Rowan <neuromancer@HAL-9000.(none)>
Date: Sun, 8 Jan 2006 16:20:46 -0500
Subject: [PATCH 53/64] Minor code optimization.

---
 doc/tools/code2doc | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index da27680ea..793df00a4 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -20,13 +20,12 @@ while (<FILEIN>)
 		if ($_ =~ /^\s+\*\s$/) { next if $first; $_ =~ s/\s\*// if not $first; }
 		if ($_ =~ /^\s\*+\/$/ or $_ !~ /^\s/) { $found = undef; next; }
 		$_ =~ s/^(\s*)\s\*\s/$1/;
-		$found = 'sorta' if $_ =~ s/\s*\*\/$//;
-		$first = undef;
+		$found = 'sorta' if $_ =~ s/\s*\*\/$//; $first = undef;
 	}
 	elsif ($_ =~ /^\s*\/\*\*\s(.*)/)
 	{
+		$_ = $1; $first = 1;
 		print FILEOUT "\n\n\n" if $start;
-		$_ = $1;
 		if ($_ =~ s/\s*\*\/$//) { $found = 'sorta'; } else { $found = $.; }
 		if ($_ =~ /::/) { $_ = "$_\n\n"; }
 		else
@@ -34,15 +33,14 @@ while (<FILEIN>)
 			my $dash; for (my $x = 0; $x < length($_); $x++) { $dash .= '-'; }
 			$_ = "$_\n$dash\n\n";
 		}
-		$first = 1;
 	}
 	elsif ($_ =~ /\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/)
 	{
 		print FILEOUT "\n" if $gotone;
 		$_ =~ s/\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/id[$1]::\n  $2/;
-		$found = 'sorta'; $gotone = 1;
+		$found = 'sorta'; $gotone = $.;
 	}
-	else { $gotone = undef; }
+	$gotone = undef if $gotone and $gotone < $.;
 	next if not $found;  $found = undef if $found eq 'sorta';
 	print FILEOUT $_ and $start = 1;
 }

From 15cd8c8d2cf4822ef64fe6cf19dd1a14833a0734 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sun, 8 Jan 2006 22:33:26 +0100
Subject: [PATCH 54/64] Simplify dependency on doc-dirs by making MAKECMDGOALS
 depend on it

---
 doc/Makefile | 34 ++++++++++++++++++----------------
 1 file changed, 18 insertions(+), 16 deletions(-)

diff --git a/doc/Makefile b/doc/Makefile
index c8d8502f7..6ce38a6ca 100644
--- a/doc/Makefile
+++ b/doc/Makefile
@@ -92,6 +92,8 @@ all-docs: man-docs html-docs pdf-docs
 ### Build Rules
 #
 
+$(MAKECMDGOALS): doc-dirs
+
 doc-dirs:
 	$(INSTALL) -d $(DOC_DIRS)
 
@@ -102,36 +104,36 @@ clean-local:
 
 # Autogenerated asciidoc files.
 
-$(TXT_DIR)/import-features.conf.txt: $(FEATURES) doc-dirs $(IMPORT_FEATURES_CONF)
+$(TXT_DIR)/import-features.conf.txt: $(FEATURES) $(IMPORT_FEATURES_CONF)
 	$(IMPORT_FEATURES_CONF) > $@
 
-$(TXT_DIR)/elinks.1.%.txt: $(MAKE_ELINKS_MANPAGE) doc-dirs $(ELINKS)
+$(TXT_DIR)/elinks.1.%.txt: $(MAKE_ELINKS_MANPAGE) $(ELINKS)
 	$(MAKE_ELINKS_MANPAGE) $@ $(ELINKS) $(HELP2DOC) > $@
 
-$(TXT_DIR)/elinkskeys.5.%.txt: $(MAKE_ELINKSKEYS_MANPAGE) doc-dirs $(KBDBIND)
+$(TXT_DIR)/elinkskeys.5.%.txt: $(MAKE_ELINKSKEYS_MANPAGE) $(KBDBIND)
 	$(MAKE_ELINKSKEYS_MANPAGE) $@ $(KBDBIND) > $@
 
 
 # Man Pages
 
-$(XML_DIR)/%.man.xml: $(TXT_DIR)/%.man.txt doc-dirs
+$(XML_DIR)/%.man.xml: $(TXT_DIR)/%.man.txt
 	$(ASCIIDOC) -b docbook -d manpage -o $@ $<
 
-$(MAN_DIR)/man1/elinks.1.in: $(XML_DIR)/elinks.1.man.xml doc-dirs
+$(MAN_DIR)/man1/elinks.1.in: $(XML_DIR)/elinks.1.man.xml
 	$(XMLTO) -o $(MAN_DIR)/man1 man $<
 	mv $(MAN_DIR)/man1/elinks.1 $@
 
-$(MAN_DIR)/man5/elinkskeys.5: $(XML_DIR)/elinkskeys.5.man.xml doc-dirs
+$(MAN_DIR)/man5/elinkskeys.5: $(XML_DIR)/elinkskeys.5.man.xml
 	$(XMLTO) -o $(MAN_DIR)/man5 man $<
 	sed -e 's/\\fI\\fR'\''/\\fI\\'\''\\fR/' < $@ > $@.tmp
 	mv $@.tmp $@
 
-$(MAN_DIR)/man5/elinks.conf.5: doc-dirs $(ELINKS)
+$(MAN_DIR)/man5/elinks.conf.5: $(ELINKS)
 	$(HELP2DOC) --elinks=$(ELINKS) --elinksconf > $@
 
 # XHTML/CSS Man Pages
 
-$(HTML_DIR)/%.html: $(TXT_DIR)/%.html.txt doc-dirs
+$(HTML_DIR)/%.html: $(TXT_DIR)/%.html.txt
 	$(ASCIIDOC) -b xhtml11 -d manpage -o $@ $<
 
 # The Manual
@@ -141,28 +143,28 @@ MANUAL_EXTRA_FILES = \
  $(TXT_DIR)/elinks.1.html.txt \
  $(TXT_DIR)/elinkskeys.5.html.txt
 
-$(HTML_DIR)/manual.html: $(MANUAL_FILES) doc-dirs $(MANUAL_EXTRA_FILES)
+$(HTML_DIR)/manual.html: $(MANUAL_FILES) $(MANUAL_EXTRA_FILES)
 	$(ASCIIDOC) -b xhtml11 -d book -o $@ -n $<
 
-$(HTML_DIR)/hacking.html: $(top_srcdir)/doc/hacking.txt doc-dirs
+$(HTML_DIR)/hacking.html: $(top_srcdir)/doc/hacking.txt
 	$(ASCIIDOC) -b xhtml11 -d book -o $@ -n $<
 
-$(HTML_DIR)/dev-intro.html: $(top_srcdir)/doc/dev-intro.txt doc-dirs
+$(HTML_DIR)/dev-intro.html: $(top_srcdir)/doc/dev-intro.txt
 	$(ASCIIDOC) -b xhtml11 -d book -o $@ -n $<
 
-$(XML_DIR)/manual.xml: $(MANUAL_FILES) doc-dirs $(MANUAL_EXTRA_FILES)
+$(XML_DIR)/manual.xml: $(MANUAL_FILES) $(MANUAL_EXTRA_FILES)
 	$(ASCIIDOC) -b docbook -d book -o $@ $<
 
-$(HTML_DIR)/manual.html-chunked: $(XML_DIR)/manual.xml doc-dirs
+$(HTML_DIR)/manual.html-chunked: $(XML_DIR)/manual.xml
 	$(XMLTO) -o $@ html $<
 
-$(PDF_DIR)/manual.pdf: $(XML_DIR)/manual.xml doc-dirs
+$(PDF_DIR)/manual.pdf: $(XML_DIR)/manual.xml
 	$(JW) -o $(PDF_DIR) -b pdf $<
 
-$(HTML_DIR)/perl.html: $(top_srcdir)/doc/perl.pod doc-dirs
+$(HTML_DIR)/perl.html: $(top_srcdir)/doc/perl.pod
 	$(POD2HTML) --outfile=$@ < $<
 
-$(HTML_DIR)/perl-hooks.html: $(top_srcdir)/contrib/perl/hooks.pl doc-dirs
+$(HTML_DIR)/perl-hooks.html: $(top_srcdir)/contrib/perl/hooks.pl
 	$(POD2HTML) --outfile=$@ < $<
 
 include $(top_srcdir)/Makefile.lib

From b2cbbf11397b8ff0965869709a2d5d9142bd6133 Mon Sep 17 00:00:00 2001
From: Russ Rowan <neuromancer@HAL-9000.(none)>
Date: Sun, 8 Jan 2006 16:46:46 -0500
Subject: [PATCH 55/64] Don't left align enums.

---
 doc/tools/code2doc | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index 793df00a4..ad90559e7 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -37,7 +37,7 @@ while (<FILEIN>)
 	elsif ($_ =~ /\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/)
 	{
 		print FILEOUT "\n" if $gotone;
-		$_ =~ s/\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/id[$1]::\n  $2/;
+		$_ =~ s/(\s*)([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/$1id[$2]::\n$1  $3/;
 		$found = 'sorta'; $gotone = $.;
 	}
 	$gotone = undef if $gotone and $gotone < $.;

From 0f8f5b06356a4e64f2b79c58101ede6a2daef910 Mon Sep 17 00:00:00 2001
From: Russ Rowan <neuromancer@HAL-9000.(none)>
Date: Sun, 8 Jan 2006 16:57:10 -0500
Subject: [PATCH 56/64] Undo last change and left align enums.

---
 doc/tools/code2doc | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index ad90559e7..793df00a4 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -37,7 +37,7 @@ while (<FILEIN>)
 	elsif ($_ =~ /\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/)
 	{
 		print FILEOUT "\n" if $gotone;
-		$_ =~ s/(\s*)([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/$1id[$2]::\n$1  $3/;
+		$_ =~ s/\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/id[$1]::\n  $2/;
 		$found = 'sorta'; $gotone = $.;
 	}
 	$gotone = undef if $gotone and $gotone < $.;

From b70133760f83b85ff6021bbbe91c263b81ba0d38 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sun, 8 Jan 2006 23:15:47 +0100
Subject: [PATCH 57/64] Make *-docs depend on doc-dirs; beware if you build
 individual files

---
 doc/Makefile | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/doc/Makefile b/doc/Makefile
index 6ce38a6ca..18a383e09 100644
--- a/doc/Makefile
+++ b/doc/Makefile
@@ -83,17 +83,15 @@ MAN_DOCS  += $(MAN_DOCS-yes)
 HTML_DOCS += $(HTML_DOCS-yes)
 PDF_DOCS  += $(PDF_DOCS-yes)
 
-man-docs:  $(addprefix $(MAN_DIR)/,$(MAN_DOCS))
-html-docs: $(addprefix $(HTML_DIR)/,$(HTML_DOCS))
-pdf-docs:  $(addprefix $(PDF_DIR)/,$(PDF_DOCS))
+man-docs:  doc-dirs $(addprefix $(MAN_DIR)/,$(MAN_DOCS))
+html-docs: doc-dirs $(addprefix $(HTML_DIR)/,$(HTML_DOCS))
+pdf-docs:  doc-dirs $(addprefix $(PDF_DIR)/,$(PDF_DOCS))
 
 all-docs: man-docs html-docs pdf-docs
 
 ### Build Rules
 #
 
-$(MAKECMDGOALS): doc-dirs
-
 doc-dirs:
 	$(INSTALL) -d $(DOC_DIRS)
 

From 5b818b20bab3b5e4b47d205ee41d9d04f612908d Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sun, 8 Jan 2006 23:36:07 +0100
Subject: [PATCH 58/64] Use the new asciidoc code markup to document the DOM
 sgml parser

---
 src/dom/sgml/parser.h | 118 ++++++++++++++++++++++++++++++++++--------
 1 file changed, 95 insertions(+), 23 deletions(-)

diff --git a/src/dom/sgml/parser.h b/src/dom/sgml/parser.h
index b81c6a933..dd3542632 100644
--- a/src/dom/sgml/parser.h
+++ b/src/dom/sgml/parser.h
@@ -11,74 +11,146 @@ struct sgml_parser;
 struct string;
 struct uri;
 
+/** enum:[sgml_parser_type]: SGML parser type
+ *
+ * There are two kinds of parser types: One that optimises one-time access to
+ * the DOM tree and one that creates a persistent DOM tree. */
 enum sgml_parser_type {
-	/* The first one is a DOM tree builder. */
-	SGML_PARSER_TREE,
-	/* The second one will simply push nodes on the stack, not building a
+	/** id:[SGML_PARSER_STREAM]::
+	 * The first one will simply push nodes on the stack, not building a
 	 * DOM tree. This interface is similar to that of SAX (Simple API for
 	 * XML) where events are fired when nodes are entered and exited. It is
 	 * useful when you are not actually interested in the DOM tree, but can
 	 * do all processing in a stream-like manner, such as when highlighting
 	 * HTML code. */
 	SGML_PARSER_STREAM,
+	/** id:[SGML_PARSER_TREE]::
+	 * The second one is a DOM tree builder, that builds a persistent DOM
+	 * tree. When using this type, it is possible to do even more
+	 * (pre)processing than for parser streams. For example you can sort
+	 * element child nodes, or purge various node such as text nodes that
+	 * only contain space characters.  */
+	SGML_PARSER_TREE,
 };
 
+/** enum:[sgml_parser_flag]: SGML parser flags
+ *
+ * These flags control how the parser behaves.
+ */
 enum sgml_parser_flag {
-	SGML_PARSER_COUNT_LINES	= 1,
-	SGML_PARSER_COMPLETE = 2,
-	SGML_PARSER_INCREMENTAL = 4,
-	SGML_PARSER_DETECT_ERRORS = 8,
+	SGML_PARSER_COUNT_LINES	= 1,	/*:: Make line numbers available. */
+	SGML_PARSER_COMPLETE = 2,	/*:: Used internally when incremental. */
+	SGML_PARSER_INCREMENTAL = 4,	/*:: Parse chunks of input. */
+	SGML_PARSER_DETECT_ERRORS = 8,	/*:: Report errors. */
 };
 
+/** struct:[sgml_parser_state]: SGML parser state
+ *
+ * The SGML parser has only little state.
+ */
 struct sgml_parser_state {
-	/* Info about the properties of the node contained by state.
+	/** id:[sgml_parser_state.info]::
+	 * Info about the properties of the node contained by state.
 	 * This is only meaningful to element and attribute nodes. For
 	 * unknown nodes it points to the common 'unknown node' info. */
 	struct sgml_node_info *info;
-	/* This is used by the DOM source renderer for highlighting the
+	/** id:[sgml_parser_state.end_token]::
+	 * This is used by the DOM source renderer for highlighting the
 	 * end-tag of an element. */
 	struct dom_scanner_token end_token;
 };
 
+/** enum:[sgml_parser_code]: (Error) codes for the SGML parser
+ *
+ * These enum values are used for return codes.
+ */
 enum sgml_parser_code {
-	SGML_PARSER_CODE_OK,		/* The parsing was successful */
-	SGML_PARSER_CODE_INCOMPLETE,	/* The parsing could not be completed */
-	SGML_PARSER_CODE_MEM_ALLOC,	/* Failed to allocate memory */
+	SGML_PARSER_CODE_OK,		/*:: The parsing was successful */
+	SGML_PARSER_CODE_INCOMPLETE,	/*:: The parsing could not be completed */
+	SGML_PARSER_CODE_MEM_ALLOC,	/*:: Failed to allocate memory */
 
-	/* FIXME: For when we will add support for requiring stricter parsing
+	/** id:[SGML_PARSER_CODE_ERROR]::
+	 * FIXME: For when we will add support for requiring stricter parsing
 	 * or even a validator. */
 	SGML_PARSER_CODE_ERROR,
 };
 
-/* If the return code is not SGML_PARSER_CODE_OK the parsing will be ended and
- * that code will be returned. */
+/** callback:[sgml_error_T]: SGML error callback
+ *
+ * Called by the SGML parser when a parsing error has occurred.
+ *
+ * If the return code is not ref:[SGML_PARSER_CODE_OK] the parsing will be
+ * ended and that code will be returned. */
 typedef enum sgml_parser_code
 (*sgml_error_T)(struct sgml_parser *, struct dom_string *, unsigned int);
 
+
+/** struct:[sgml_parser]: The SGML parser
+ *
+ * This struct hold info used while parsing SGML data.
+ *
+ * NOTE: The only variable the user should set is ref:[error_func]. */
 struct sgml_parser {
-	enum sgml_parser_type type;	/* Stream or tree */
-	enum sgml_parser_flag flags;	/* Flags that control the behaviour */
+	enum sgml_parser_type type;	/*:: Stream or tree */
+	enum sgml_parser_flag flags;	/*:: Flags that control the behaviour */
 
-	struct sgml_info *info;		/* Backend dependent info */
+	struct sgml_info *info;		/*:: Backend dependent info */
 
-	struct dom_string uri;		/* The URI of the DOM document */
-	struct dom_node *root;		/* The document root node */
+	struct dom_string uri;		/*:: The URI of the DOM document */
+	struct dom_node *root;		/*:: The document root node */
 
-	sgml_error_T error_func;
+	sgml_error_T error_func;	/*:: Called for detected errors */
 
-	struct dom_stack stack;		/* A stack for tracking parsed nodes */
-	struct dom_stack parsing;	/* Used for tracking parsing states */
+	struct dom_stack stack;		/*:: A stack for tracking parsed nodes */
+	struct dom_stack parsing;	/*:: Used for tracking parsing states */
 };
 
+
+/** func:[init_sgml_parser]: Initialise an SGML parser
+ *
+ * Initialise an SGML parser with the given properties.
+ * 
+ * type::	Stream or tree; one-time or persistant.
+ * doctype::	The document type, this affects what sub type nodes are given.
+ * uri::	The URI of the document root.
+ * flags::	Flags controlling the behaviour of the parser.
+ *
+ * Returns the created parser or NULL.
+ */
 struct sgml_parser *
 init_sgml_parser(enum sgml_parser_type type, enum sgml_document_type doctype,
 		 struct dom_string *uri, enum sgml_parser_flag flags);
 
+/** func:[done_sgml_parser]: Release an SGML parser
+ *
+ * Deallocates all resources, _expect_ the root node.
+ *
+ * parser::	The parser being released.
+ */
 void done_sgml_parser(struct sgml_parser *parser);
 
+/** func:[parse_sgml]: Parse a chunk of SGML source
+ *
+ * Parses the given `buffer`. For incremental rendering the last buffer can be
+ * signals through the `complete` parameter.
+ *
+ * parser::	A parser created with ref:[init_sgml_parser].
+ * buffer::	A string containing the chunk to parse.
+ * complete::	Whether this is the last chunk to parse.
+ *
+ * The returned code is ref:[SGML_PARSER_CODE_OK] if the buffer was
+ * successfully parserd, else a code hinting at the error.
+ */
 enum sgml_parser_code
 parse_sgml(struct sgml_parser *parser, struct dom_string *buffer, int complete);
 
+/** func:[get_sgml_parser_line_number]: Get the line position in the source
+ *
+ * Returns what line number the parser is currently at or zero if there has
+ * been no parsing yet.
+ *
+ * NOTE: Line numbers are recoderded in the scanner tokens.
+ */
 unsigned int get_sgml_parser_line_number(struct sgml_parser *parser);
 
 #endif

From 0eef02b8f3508e600baaf74a4bfe28ecc0c49b66 Mon Sep 17 00:00:00 2001
From: Russ Rowan <neuromancer@HAL-9000.(none)>
Date: Sun, 8 Jan 2006 17:42:04 -0500
Subject: [PATCH 59/64] Handle: SGML_PARSER_CODE_OK = 1,         /**: The
 parsing was successful */

---
 doc/tools/code2doc | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index 793df00a4..59d4ed099 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -20,12 +20,12 @@ while (<FILEIN>)
 		if ($_ =~ /^\s+\*\s$/) { next if $first; $_ =~ s/\s\*// if not $first; }
 		if ($_ =~ /^\s\*+\/$/ or $_ !~ /^\s/) { $found = undef; next; }
 		$_ =~ s/^(\s*)\s\*\s/$1/;
-		$found = 'sorta' if $_ =~ s/\s*\*\/$//; $first = undef;
+		$found = 'sorta' if $_ =~ s/\s*\*\/$/\n/; $first = undef;
 	}
 	elsif ($_ =~ /^\s*\/\*\*\s(.*)/)
 	{
 		$_ = $1; $first = 1;
-		print FILEOUT "\n\n\n" if $start;
+		print FILEOUT "\n\n" if $start;
 		if ($_ =~ s/\s*\*\/$//) { $found = 'sorta'; } else { $found = $.; }
 		if ($_ =~ /::/) { $_ = "$_\n\n"; }
 		else
@@ -34,13 +34,13 @@ while (<FILEIN>)
 			$_ = "$_\n$dash\n\n";
 		}
 	}
-	elsif ($_ =~ /\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/)
+	elsif ($_ =~ /\s*([A-Z0-9_]+)(\s+=\s+[0-9]+)?,\s*\/\*::\s*(.*)\s+\*\/$/)
 	{
 		print FILEOUT "\n" if $gotone;
-		$_ =~ s/\s*([A-Z0-9_]+),\s*\/\*::\s*(.*)\s+\*\/$/id[$1]::\n  $2/;
+		$_ =~ s/\s*([A-Z0-9_]+)(\s+=\s+[0-9]+|),\s*\/\*::\s*(.*)\s+\*\/$/id[$1]::\n  $3/;
 		$found = 'sorta'; $gotone = $.;
 	}
-	$gotone = undef if $gotone and $gotone < $.;
+	print FILEOUT "\n" and $gotone = undef if $gotone and $gotone < $.;
 	next if not $found;  $found = undef if $found eq 'sorta';
 	print FILEOUT $_ and $start = 1;
 }

From 61efbc4c9a6b0879f43cea3cdf1538f9f7c15119 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sun, 8 Jan 2006 23:44:59 +0100
Subject: [PATCH 60/64] First attempt to build html/api docs

It only does src/dom/sgml/parser.h for now.
---
 doc/Makefile | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/doc/Makefile b/doc/Makefile
index 18a383e09..121961c23 100644
--- a/doc/Makefile
+++ b/doc/Makefile
@@ -14,6 +14,7 @@ TXT_DIR	 = $(top_srcdir)/doc/txt
 
 DOC_DIRS = \
  $(HTML_DIR) \
+ $(HTML_DIR)/api \
  $(MAN_DIR)/man1 \
  $(MAN_DIR)/man5 \
  $(TXT_DIR) \
@@ -51,12 +52,14 @@ FEATURES = $(top_srcdir)/features.conf
 ### Scripts
 #
 
+CODE2DOC		= $(top_srcdir)/doc/tools/code2doc
 HELP2DOC		= $(top_srcdir)/doc/tools/help2doc
 IMPORT_FEATURES_CONF	= $(top_srcdir)/doc/tools/import-features.conf
 MAKE_ELINKS_MANPAGE	= $(top_srcdir)/doc/tools/make-elinks-manpage
 MAKE_ELINKSKEYS_MANPAGE	= $(top_srcdir)/doc/tools/make-elinkskeys-manpage
 
 HTML_DOCS-$(CONFIG_ASCIIDOC) += \
+ api/dom-sgml-parser.html \
  elinks.1.html \
  elinkskeys.5.html \
  hacking.html \
@@ -165,4 +168,10 @@ $(HTML_DIR)/perl.html: $(top_srcdir)/doc/perl.pod
 $(HTML_DIR)/perl-hooks.html: $(top_srcdir)/contrib/perl/hooks.pl
 	$(POD2HTML) --outfile=$@ < $<
 
+## API Docs
+#
+
+$(HTML_DIR)/api/dom-sgml-parser.html: $(top_srcdir)/src/dom/sgml/parser.h
+	$(CODE2DOC) $< | $(ASCIIDOC) -f code2doc.conf -b xhtml11 -d book -o $@ -n -
+
 include $(top_srcdir)/Makefile.lib

From 7acb3059a84c32de6420106b272b973d2a16fc4b Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Sun, 8 Jan 2006 23:57:14 +0100
Subject: [PATCH 61/64] Make code2doc print to stdout

---
 doc/tools/code2doc | 15 +++++----------
 1 file changed, 5 insertions(+), 10 deletions(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index 59d4ed099..6fbf56259 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -3,16 +3,13 @@ use strict;
 use warnings;
 use diagnostics;
 
-print "Usage: $0 [FILE]\n\tParses [FILE], outputing the result to [FILE].out.\n"
+print "Usage: $0 [FILE]\n\tParses [FILE], outputing the result to stdout.\n"
 	and exit if not @ARGV;
 
 my ($input)  = @ARGV;
-my ($output) = $input . '.out';
 my ($found, $start, $first, $gotone);
 print "CopyleftŠ 2006, Russ Rowan (See `COPYING')\n" and exit if $input eq '-v';
 open FILEIN,  "<$input"  or print "File `$input' was not found.\n"     and exit;
-open FILEOUT, ">$output" or print "Cannot open $output for writing.\n" and exit;
-print "Processing $input...\n";
 while (<FILEIN>)
 {
 	if ($found)
@@ -25,7 +22,7 @@ while (<FILEIN>)
 	elsif ($_ =~ /^\s*\/\*\*\s(.*)/)
 	{
 		$_ = $1; $first = 1;
-		print FILEOUT "\n\n" if $start;
+		print STDOUT "\n\n" if $start;
 		if ($_ =~ s/\s*\*\/$//) { $found = 'sorta'; } else { $found = $.; }
 		if ($_ =~ /::/) { $_ = "$_\n\n"; }
 		else
@@ -36,14 +33,12 @@ while (<FILEIN>)
 	}
 	elsif ($_ =~ /\s*([A-Z0-9_]+)(\s+=\s+[0-9]+)?,\s*\/\*::\s*(.*)\s+\*\/$/)
 	{
-		print FILEOUT "\n" if $gotone;
+		print STDOUT "\n" if $gotone;
 		$_ =~ s/\s*([A-Z0-9_]+)(\s+=\s+[0-9]+|),\s*\/\*::\s*(.*)\s+\*\/$/id[$1]::\n  $3/;
 		$found = 'sorta'; $gotone = $.;
 	}
-	print FILEOUT "\n" and $gotone = undef if $gotone and $gotone < $.;
+	print STDOUT "\n" and $gotone = undef if $gotone and $gotone < $.;
 	next if not $found;  $found = undef if $found eq 'sorta';
-	print FILEOUT $_ and $start = 1;
+	print STDOUT $_ and $start = 1;
 }
-close FILEOUT;
 close FILEIN;
-print "$output completed.\n";

From 4aa57ffc59851b328a36e14612d78c9ebe057f40 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Mon, 9 Jan 2006 00:00:10 +0100
Subject: [PATCH 62/64] Add asciidoc configuration file for output of code2doc

---
 doc/code2doc.conf | 52 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 52 insertions(+)
 create mode 100644 doc/code2doc.conf

diff --git a/doc/code2doc.conf b/doc/code2doc.conf
new file mode 100644
index 000000000..039f5a022
--- /dev/null
+++ b/doc/code2doc.conf
@@ -0,0 +1,52 @@
+[specialwords]
+emphasizedwords=\bAsciiDoc\b
+monospacedwords=\basciidoc\(1\)
+
+[id-inlinemacro]
+<a id="{0}" href="#{0}">{0}</a>
+
+[enum-inlinemacro]
+<a id="{0}" href="#{0}">enum {0}</a>
+
+[func-inlinemacro]
+<a id="{0}" href="#{0}">{0}()</a>
+
+[struct-inlinemacro]
+<a id="{0}" href="#{0}">struct {0}</a>
+
+[callback-inlinemacro]
+<a id="{0}" href="#{0}">callback {0}</a>
+
+[ref-inlinemacro]
+<a href="{target}#{0}">{0}</a>
+
+[replacements]
+(^|[^-])--($|[^-])=\1--\2
+
+[tags]
+ilisttext=|
+olisttext=|
+vlisttext=|
+qlisttext=|
+colisttext=|
+
+[tags]
+title1=<h1>|</h1>
+title2=<h2>|</h2>
+title3=<h3>|</h3>
+
+[literalparagraph]
+<table border="1" class="code"><tr><td><pre>
+|
+</pre></td></tr></table>
+
+[listingblock]
+<p><b>{title}</b></p>
+<table border="1" class="code"><tr><td><pre>
+|
+</pre></td></tr></table>
+
+[noteblock]
+<div><p><b>{title}</b></p>
+|
+</div>

From a7fb4a71e392abfa17685ec8f069ebc997a61ad7 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Mon, 9 Jan 2006 00:01:09 +0100
Subject: [PATCH 63/64] Insert missing ':' after 'id', use tab for indentation

---
 doc/tools/code2doc | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index 6fbf56259..3adb53d7f 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -34,7 +34,7 @@ while (<FILEIN>)
 	elsif ($_ =~ /\s*([A-Z0-9_]+)(\s+=\s+[0-9]+)?,\s*\/\*::\s*(.*)\s+\*\/$/)
 	{
 		print STDOUT "\n" if $gotone;
-		$_ =~ s/\s*([A-Z0-9_]+)(\s+=\s+[0-9]+|),\s*\/\*::\s*(.*)\s+\*\/$/id[$1]::\n  $3/;
+		$_ =~ s/\s*([A-Z0-9_]+)(\s+=\s+[0-9]+|),\s*\/\*::\s*(.*)\s+\*\/$/id:[$1]::\n\t$3/;
 		$found = 'sorta'; $gotone = $.;
 	}
 	print STDOUT "\n" and $gotone = undef if $gotone and $gotone < $.;

From 2d7356d29a85e8493acb91d1f2fbc5194c4f4677 Mon Sep 17 00:00:00 2001
From: Jonas Fonseca <fonseca@diku.dk>
Date: Mon, 9 Jan 2006 00:53:41 +0100
Subject: [PATCH 64/64] Improve support for struct documentation by saving the
 struct name

This makes it easy to create

id:[$struct$member]::
	One line description
---
 doc/tools/code2doc | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/doc/tools/code2doc b/doc/tools/code2doc
index 3adb53d7f..d6b03202b 100755
--- a/doc/tools/code2doc
+++ b/doc/tools/code2doc
@@ -7,9 +7,10 @@ print "Usage: $0 [FILE]\n\tParses [FILE], outputing the result to stdout.\n"
 	and exit if not @ARGV;
 
 my ($input)  = @ARGV;
-my ($found, $start, $first, $gotone);
+my ($found, $start, $first, $gotone, $idpath);
 print "CopyleftŠ 2006, Russ Rowan (See `COPYING')\n" and exit if $input eq '-v';
 open FILEIN,  "<$input"  or print "File `$input' was not found.\n"     and exit;
+$idpath = '';
 while (<FILEIN>)
 {
 	if ($found)
@@ -24,6 +25,7 @@ while (<FILEIN>)
 		$_ = $1; $first = 1;
 		print STDOUT "\n\n" if $start;
 		if ($_ =~ s/\s*\*\/$//) { $found = 'sorta'; } else { $found = $.; }
+		if ($_ =~ /struct:[[]([^\]]+)[\]]/) { $idpath = "$1."; } else { $idpath = ''; }
 		if ($_ =~ /::/) { $_ = "$_\n\n"; }
 		else
 		{
@@ -31,10 +33,10 @@ while (<FILEIN>)
 			$_ = "$_\n$dash\n\n";
 		}
 	}
-	elsif ($_ =~ /\s*([A-Z0-9_]+)(\s+=\s+[0-9]+)?,\s*\/\*::\s*(.*)\s+\*\/$/)
+	elsif ($_ =~ /^(\s|[^\s=]+)*[\s*]([A-Za-z0-9_]+)(\s+=\s+[^,;]+)?[,;]\s*\/\*::\s*(.*)\s+\*\/$/)
 	{
 		print STDOUT "\n" if $gotone;
-		$_ =~ s/\s*([A-Z0-9_]+)(\s+=\s+[0-9]+|),\s*\/\*::\s*(.*)\s+\*\/$/id:[$1]::\n\t$3/;
+		$_ = "\nid:[$idpath$2]::\n\t$4\n";
 		$found = 'sorta'; $gotone = $.;
 	}
 	print STDOUT "\n" and $gotone = undef if $gotone and $gotone < $.;