[docs] xmldocs - racke modified 2 files

[docs at icdevgroup.org]

User:      racke
Date:      2005-09-21 12:07:25 GMT
Modified:  .        Makefile
Modified:  bin      refs-autogen
Log:
new commandline options --validate and --skip-invalid for refs-autogen
script, requires XML::Twig
--validate validates XML chunks which helps to relate XML errors to symbols
--skip-invalid ignores invalid XML chunks for unattended builds

new Makefile variables REFS_AUTOGEN and REFS_AUTOGEN_FLAGS

Revision  Changes    Path
1.72      +8 -4      xmldocs/Makefile


rev 1.72, prev_rev 1.71
Index: Makefile
===================================================================
RCS file: /anon_cvs/repository/xmldocs/Makefile,v
retrieving revision 1.71
retrieving revision 1.72
diff -u -r1.71 -r1.72
--- Makefile	3 Sep 2005 17:09:47 -0000	1.71
+++ Makefile	21 Sep 2005 12:07:25 -0000	1.72
@@ -28,6 +28,10 @@
 PSR         = xsltproc
 PSR_FLAGS   = --xinclude
 
+# Scripts and options
+REFS_AUTOGEN = bin/refs-autogen
+REFS_AUTOGEN_FLAGS ?=
+
 VPATH       = guides refs howtos glossary
 .SILENT:
 .PHONY: all complete
@@ -239,7 +243,7 @@
 # Silly, rewrite this, I forgot about $*. Or $* wouldn't help? I'm not 
 # willing to think about it right now.
 refxmls: BOTH = --both
-refxmls: bin/refs-autogen $(foreach stype,$(SYMBOL_TYPES),refs/$(stype).xml) howtos/howtos.xml glossary/glossary.xml
+refxmls: $(REFS_AUTOGEN) $(foreach stype,$(SYMBOL_TYPES),refs/$(stype).xml) howtos/howtos.xml glossary/glossary.xml
 	:
 $T/%.list: BNAME = $(subst $T/,,$@)
 refs/%.xml: BNAME = $(subst refs/,,$@)
@@ -247,10 +251,10 @@
 refs/%.xml: FNAME = $(subst .xml,,$(BNAME))
 # A little 'overwork' here: we regenerate all .xml files even if just
 # one file changes.
-$T/%.list refs/%.xml: $(foreach icver,$(IC_VERSIONS),cache/$(icver)/.cache.bin) $(shell find refs/ -regex '.+[^(\.xml)]$$') bin/refs-autogen
+$T/%.list refs/%.xml: $(foreach icver,$(IC_VERSIONS),cache/$(icver)/.cache.bin) $(shell find refs/ -regex '.+[^(\.xml)]$$') $(REFS_AUTOGEN)
 	# PEH, -g is useless since tags migrate between tag groups
-	#bin/refs-autogen -g $(FNAME) -o $@ $(BOTH) $(IC_VERSIONS)
-	bin/refs-autogen -o $@ $(BOTH) $(TARGET_RELEASE) $(IC_VERSIONS)
+	#$(REFS_AUTOGEN) $(REFS_AUTOGEN_FLAGS) -g $(FNAME) -o $@ $(BOTH) $(IC_VERSIONS)
+	$(REFS_AUTOGEN) $(REFS_AUTOGEN_FLAGS) -o $@ $(BOTH) $(TARGET_RELEASE) $(IC_VERSIONS)
 
 
 #############################################################



1.96      +60 -1     xmldocs/bin/refs-autogen


rev 1.96, prev_rev 1.95
Index: refs-autogen
===================================================================
RCS file: /anon_cvs/repository/xmldocs/bin/refs-autogen,v
retrieving revision 1.95
retrieving revision 1.96
diff -u -r1.95 -r1.96
--- refs-autogen	7 Sep 2005 16:29:16 -0000	1.95
+++ refs-autogen	21 Sep 2005 12:07:25 -0000	1.96
@@ -43,6 +43,8 @@
 my $output_spec; # 'list' produces tag list, 'xml' produces real xml source
 my $output_both; # Unconditionally override $output_spec
 my $no_autorefs; # Generate autorefs.ent collection of entities by default
+my $validate = 0; # Whether to validate XML chunks
+my $skip_invalid = 0; # Whether to skip invalid XML chunks
 my $autopath = "docbook/autorefs.ent";
 my %dups; # List of symbols names that are not unique
 my $last_path; # Last version we want docs generated for (say, 5.2.0).
@@ -60,6 +62,8 @@
 	"both|b!"    => \$output_both,
 	"noentities|noents!" => \$no_autorefs,
 	"last-path|last|lp=s" => \$last_path,
+	"validate|c!" => \$validate,
+	"skip-invalid|s" => \$skip_invalid,
 )) { die "Error parsing options\n" }
 
 # Determine which stuff to output
@@ -70,6 +74,10 @@
 	die "Unknown output combination '$output_spec'\n";
 }
 
+if ($validate) {
+	require XML::Twig;
+}
+
 $specific_only and $specific_only =~ s/s$//; # if user entered name in plural
 @ARGV or die "Usage: $0 version[s]\n";
 
@@ -781,7 +789,25 @@
 
 	## Add items
 	for my $key (sort keys %{ $symbols{$group} }) {
-		$refpage .= $symbols{$group}{$key}
+		## Validate item
+		my ($buf, $msg);
+
+		$buf = '<!DOCTYPE reference PUBLIC "-//OASIS//DTD DocBook-Interchange XML V4.2//EN"
+	"../docbook/docbookxi.dtd">';
+		$buf .= $symbols{$group}{$key};
+
+		if ($validate) {
+			unless (validate_xml($buf, \$msg)) {
+				if ($skip_invalid) {
+					warn "ERR: symbol ${group}::$key: $msg\n";
+					next;
+				} else {
+					die "ERR: symbol ${group}::$key: $msg\n";
+				}
+			}
+		}
+		
+		$refpage .= $symbols{$group}{$key};
 	}
 
 	## Close up
@@ -1182,6 +1208,39 @@
 	} else {
 		return "@orig"
 	}
+}
+
+sub validate_xml {
+	my ($xml, $msgref) = @_;
+	my ($buf);
+
+	my $twig = new XML::Twig;
+
+	if ($twig->safe_parse($xml)) {
+		# sane XML
+		return 1;
+	}
+
+	# try to analyze error message
+	my $error = $@;
+		
+	if ($error =~ /^\s+(.*?) at line (\d+), column (\d+), byte (\d+)/s) { # at(.*?)line (\d+)$/) {
+		my ($cause, $line, $col, $byte) = ($1, $2, $3, $4);
+		
+		$cause = $1;
+		if ($cause eq 'undefined entity') {
+			$buf = substr($xml, $4);
+			if ($buf =~ /^&([\w_-]+);/) {
+				$$msgref = "undefined entity $1";
+				return;
+			}
+		}
+
+		$$msgref = "$cause " . substr($xml, $4, 60);
+	} else {
+		$$msgref = $error;
+	}
+	return;
 }
 
 sub O { print "@_\n"; print STDOUT "@_\n" if $verbose }