[docs] xmldocs - docelic modified 2 files

[docs at icdevgroup.org]

User:      docelic
Date:      2004-09-22 09:02:47 GMT
Modified:  .        TODO
Modified:  bin      refs-autogen
Log:
- bin/refs-autogen:
  - compress the "Available in" field (like, 4.6.0-5.2.0) as Jon suggested

- TODO: items

Revision  Changes    Path
1.21      +17 -22    xmldocs/TODO


rev 1.21, prev_rev 1.20
Index: TODO
===================================================================
RCS file: /var/cvs/xmldocs/TODO,v
retrieving revision 1.20
retrieving revision 1.21
diff -u -r1.20 -r1.21
--- TODO	22 Sep 2004 08:02:47 -0000	1.20
+++ TODO	22 Sep 2004 09:02:47 -0000	1.21
@@ -6,29 +6,17 @@
   of plain "C").
 - See why the autogenerated navigation links are so retarded (just home/up,
   no next/prev)
-- In refs-autogen at a few places we keep both the hash/array representation
-  of values, and their string representation. (We always generate string
-  when modifying array/hash). It should be made so that we only modify
-  array/hash during the whole program, then in the last step before
-  generating the template they get stringified.
-- Visually mark obsolete items (those not present in cvs-head)
-- leaf nodes
 - not to forget, fix cases where context goes to negative values
 
 - In iccattut:
-  - mention which paths would be valid for /usr/local tarball
-    install.
   - under important file and directories, mention default username,
-    and /cgi-bin/ vs /cgi-bin/ic/ in the Catalog line. And rename the section
-	  appropriately.
+    And rename the section appropriately.
   - make a "translation map" of /etc/interchange/* to RPM-equivs.
   - item for package names, like interchange-cat-foundation, wenglish, etc..
-  - fix markup in iccattut section "pages/checkout.html"
-	  and "etc/profiles.order" and 2 sections after it and below
+    ( This is ok, docbook 4.4 will have <package> element )
   - at the top where describing IC, mention it's built on Perl and ITL
-	- replace [area search] in phase5 with better (file and directory)
-	- replace <tag> with <code> where putting tag alone in the source would work.
-	- explain syntax accepted in profile files
+  - replace <tag> with <code> where putting tag alone in the source would work.
+  - explain syntax accepted in profile files
 
 DOCUMENTATION SYSTEM:
 - copy the definition for <example> to a
@@ -40,7 +28,7 @@
   (%tag.* structures) and warn if invalid option is found in any tag file.
 
 iccattut:
-- give examples for the tasks in 'do yourself' section
+- give examples for the tasks in 'do yourself' section (in progress)
 - give good practices about filtering, security
 - see problems from old docs/TODO notes on iccattut
 
@@ -58,15 +46,22 @@
   say, 15 lines of context, but it'll come great when you have a copy
   of a 300-lines usertag. Example for this could be taken directly out of
   mwforum demo on mwforum.org
-- Make that thing Jon suggested, to write like "4.8.0-5.3.0" instead of
-  "4.8.0, 5.0.0, 5.2.0, 5.3.0" (will probably look better in the future).
+- In refs-autogen at a few places we keep both the hash/array representation
+  of values, and their string representation. (We always generate string
+  when modifying array/hash). It should be made so that we only modify
+  array/hash during the whole program, then in the last step before
+  generating the template they get stringified.
+- Visually mark obsolete items (those not present in cvs-head)
+- leaf nodes
 
  Long-term:
 - filenames in Source contexts should also be clickable. this is longterm
   because it'll involve perltidy and other stuff I have in mind ...
-- rebuild the make system
-- isolate build steps and make it possible to build specific "pages"
-  without the rest
+
+- rebuild the make system:
+  - isolate build steps and make it possible to build specific "pages"
+    without the rest
+  - better dependencies 
 
 DOCUMENTATION ITSELF:
 - Resolve items from tmp/missing file. (You need to run 'make' in your tree



1.31      +40 -0     xmldocs/bin/refs-autogen


rev 1.31, prev_rev 1.30
Index: refs-autogen
===================================================================
RCS file: /var/cvs/xmldocs/bin/refs-autogen,v
retrieving revision 1.30
retrieving revision 1.31
diff -u -r1.30 -r1.31
--- refs-autogen	22 Sep 2004 08:02:47 -0000	1.30
+++ refs-autogen	22 Sep 2004 09:02:47 -0000	1.31
@@ -58,6 +58,7 @@
 my %templates;
 my $max_ctxs = 10;
 my @set_missing_all;
+my @parsed_versions;
 
 my @page_order = (qw/purpose default structure synopsis description example notes bugs/, "symbol type", "source", "author", "copyright", "see also");
 
@@ -112,6 +113,8 @@
 	# no more).
 	loaddb( "$cachedir/$dumppath" );
 
+	push @parsed_versions, $hash{version};
+
 	# Outer loop: symbol types (pragmas, globvars, ...)
 	# Inner loop: actual symbols
 	while ( my ($gkey,$gval) = each %{ $hash{symbols} } ) {
@@ -300,6 +303,9 @@
 	}
 	$ag{'see also'} = join ", ", @see_items;
 
+	# Compress 4.6.0, 4.8.0, 5.0.0 to 4.6.0-5.0.0
+	$ag{'available in'} = compress_availability($ag{'_available in'});
+
 	# Finally, set default values if they weren't overriden by real information
 	for my $field (@page_order) {
 		if ( ! $ag{$field} ) {
@@ -327,6 +333,40 @@
 
 	# Save
 	$symbols{$k} = $template unless ( $ag{ignore} and $ag{ignore} =~ /[y1d]/i );
+}
+
+sub compress_availability {
+	my @avails = my @orig = @{ (shift) };
+	@avails or die "compress_availability(): Shouldn't happen";
+
+	# XXX this needs work, but at the moment it serves us well
+	# (it can't handle like x-y, z, a-b. It only handles start-end)
+
+	my $start = shift @avails;
+	my $si; # starting index
+	my $cl = 0; # compress level
+	for (my $i=0; $i<@parsed_versions;$i++) {
+		if ( $parsed_versions[$i] eq $start ) {
+			$si = $i;
+		}
+	}
+	my $end;
+	while ( $_ = shift @avails ) {
+		if ( $parsed_versions[++$si] eq $_ ) {
+			$end = $_;
+			$cl++;
+		} else {
+			unshift @avails, $_;
+			last
+		}
+	}
+	local $" = ", ";
+	if ( $cl > 2 ) {
+		unshift @avails, "$start-$end";
+		return "@avails"
+	} else {
+		return "@orig"
+	}
 }