jsmenu-codedev — Discussions regarding jsmenu code & development of it

[jsmenu-codedev] Changelog 2002-10-13

[Morten W. <war...@on...>]

have tried to get item tracking faster (by making the item force it
instead of using MTMTrackExpand()), as well as working on how submenus
and tracking is done.  not done yet.

2002-10-13:
 - Worked on bug #455582, tracked item lost on sub action.  Changes to
   how items do submenu opening & closing etc is changed.  All items
   that point to URLs have an itemClick() method defined and that
   takes care of tracking & submenu opening (submenu closing is still
   done through MTMSubAction()).
 - Changed tracking code so it's able to track multiple items
 - Added direct setting of tracked item, forced update through
   itemClick() and ignore-variable so that the tracking code call from
   the document an item points doesn't refresh the menu.  Should speed
   up things by quite an amount.
 - Changed how submenus are auto-closed, MTMCloseSubs() has changed
   and is a recursive function go upwards, and is now called from
   MTMSubAction() and not from MTMDisplayMenu().
 - Added childNodes-reference (points to submenu.items) so that the
   code can say item.childNodes[i] instead of item.submenu.items[i].
 - Code has bugs regarding handling of items.  Tracking is turned on
   at times where it shouldn't, creating unwanted effects.
 - Scrolling doesn't work (it's turned off for a rewrite)


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-08-25

[Morten W. <war...@on...>]

I've taken a couple of steps backwards to try to get development going
again.  over the months that have passed since v2.3.2 got released
I've tried to do a large amount of change in the code to incorporate
some important fixes (tracking of multiple items, direct expansion and
highlight of items, and more usage of DHTML capabilities) but ended up
walking into a dead-end with those.  therefore I've gotten the
v2.3.2-macfriendly code into CVS now, and will do smaller patches
against that.  I've kept the code I've written up to now and will
merge some of those changes when I go along so it won't be necessary
to do all the work again.

so today's Changelog says:

2002-08-25:
 - Merged IE stylesheet patch to mtmcode.js and updated CREDITS.
 - Merged patch #536699 to mtmcode.js, but it ended up being a simple
   change of location.hostname to location.host, since that works.
   Updated CREDITS, of course.
 - Fixed bug #527465 by changing the code that sets MTMUA.preHREF.
   Now it'll also work with a query string containing "/" (i.e
   "code.html?foo=bar/baz").
 - Removed all unwanted references to MTMUA.preHREF, changing them to
   use MTMUA.resolveURL().  This makes image, stylesheet and script URLs
   correctly resolve relative to the path of code.html.

2002-08-25:
 - Checked in v2.3.2-macfriendly so that CVS and release are in sync
 - Updated COPYING, CREDITS & ChangeLog with the files from the
   v2.3.2-macfriendly release.
 - Tagged all files with rel-2-3-2-macfriendly so they're marked.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-02-26

[Morten W. <war...@on...>]

I checked in a small patch 20 minutes ago attempting to fix the
IE5/Mac problems:

2002-02-26:
 - Altered the if()-statement in the IE detection code deciding whether
   the browser is DOM-capable or not so that it rejects Mac-browsers.
   In other words they'll fall back to the IE4/Windows-approach.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] ToDo-list for v2.3.3

[Morten W. <war...@on...>]

I've been having a look at the list of open bugs & feature requests
and have found a few I'd like to look at for v2.3.3.  The problems
with IE5/Mac is given first priority, after that comes the following
things, not in order of priority:

1: style-element's id-attribute doesn't validate.
   this isn't a very difficult thing to fix, grab the HTML 4.01
   Transitional DTD, add the id-attribute, write some instructions,
   publish on the web, include it in the documentation
2: Pre/postload all images.
   (this of course includes any items in the icon list)
3: tracked item lost on sub action.
4: test the iframe-support patch.

#3 is a slightly larger issue.  I'm thinking along the lines of
rewriting the tracking functionality somewhat, since it's not optimal
at the moment.  my ideas are:

1: all items should have an itemClick() method defined, and they
   should use an onclick-event to call that method, setting the
   tracked item, refreshing the menu
   (the first call to MTMDisplayMenu() after this is ignored, so the
    menu isn't refreshed twice)
2: the code behind itemClick() will store the previously tracked item
   so that it can use DOM-methods directly to set/unset the class
   properties, thereby removing the need for a redisplay of the menu.
3: itemClick() for submenus will not redisplay the menu either, just
   add the necessary items to the document if needed, and set their
   style-properties to "display:none;" when they're removed.

this change means that the tracking code will only be used when the
user hits his/her back-button or follows a link in document.  instead
of working on a first-hit basis it should store any items found in an
array, enabling hits on more than one item.  secondly it won't need to
open & close submenus while it's doing this since that can be done by
walking _up_ the tree once a hit is found using the .parentNode
property of the item.  this will of course also mean changes to the
code for auto-closing of submenus is needed.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-02-24

[Morten W. <war...@on...>]

with the release of v2.3.2 there were a few changes to the
release-archive part of the CVS repository.  the following files were
added or changed:

tree-menu/example-site/install-guide/options/time-out.shtml
tree-menu/example-site/upgrading/version-2.2.6.html
tree-menu/example-site/upgrading/version-2.3.0.html
tree-menu/example-site/upgrading/version-2.3.1.html
tree-menu/example-site/code.html
tree-menu/example-site/javascript-frameset.html
tree-menu/example-site/mtmcode.js
tree-menu/example-site/mtmtrack.js
tree-menu/CREDITS
tree-menu/ChangeLog
tree-menu/README

the only real change was MTMTimeOut which due to patch #510937 now can
be set to 0 and upwards, not 1 like before.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-02-22

[Morten W. <war...@on...>]

Two quick bugfixes to get v2.3.2 ready:

2002-02-22:
 - Refixed bug #517717 so MTMcreateStyleSheet() doesn't barf when the
   Mozilla quirk style element isn't present.
 - Fixed a part of the output code for non-DOM-capable browsers which
   made the first parts of the document not being output correctly.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-02-19

[Morten W. <war...@on...>]

Fixed one bug, declared bug #424050 a "won't fix", and had a closer
look at bug #424794 in Mozilla 0.9.8 noticing that they're doing
things better now.  Things are progressing nicely, in my opinion:

2002-02-19:
 - Rewrote the browser sniffing code in function MTMBrowser() so it
   sniffs for Opera first.  This simplified things a bit since it means
   Opera pretending to be IE/Netscape is still sniffed as Opera.
 - Fixed bug #517717 through disabling existing style sheets in DOM-
   capable browsers.  Function MTMdisableStyleSheets() created to do
   this since it'll be used for both inline and linked style sheets
   (before they're written to the document).


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-02-18

[Morten W. <war...@on...>]

The bug fixes continues.  Looked at two of the bugs introduced in
v2.3.1 and fixed both of them.  Yesterday the web site got a few
updates too (particularly to the list of sites in use), but that's not
mentioned here :)

2002-02-18:
 - Removed usage of "height: 100%;" in inline style sheet, thereby
   fixing bug #510929.  Tried reproducing the non-tiling background
   image, but couldn't, am unsure where it came from but now things
   work like they should.
 - Removed the NN6-scrollTo(0,0) and changed scrollBy() to scrollTo()
   in the scroll code, which temporarily fixes bug #512447.  Scrolling
   is done correctly, but it now also scrolls even though the link is
   in view.  Created feature request #519613 for this issue.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-02-15

[Morten W. <war...@on...>]

So far just a small patch inserted, changing an if()-statement in
MTMStartMenu() which in turn means MTMTimeOut now can be set to 0 for
no delay (previously it had to be 1).  I imported the release archive
into the CVS repository, so all documentation and everything for that
is also available from CVS.

2002-02-15:
 - Changed if()-statement in MTMStartMenu() in accordance with patch  
   #510937.
 - Imported release archive into CVS repository as "release-archive".


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-02-11

[Morten W. <war...@on...>]

one of the more useful features is now supported.  I've tried to test
the function to make sure it works correctly, as far as I can tell it
does.  tracking of "../"-URLs rocks :)

2002-02-11:
 - Finished creating MTMresolveURL() which now means that the tree correctly
   tracks relative URLs using "../" (it's not fool-proof though, "foo/../bar"
   isn't resolved to "bar", and a "../" too many will result in URLs like
   "http://main.html").
   Notes:
     - All protocols/schemes found in absoluteArray[] will be regarded as
       non-local URLs, and will be used directly.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-02-08

[Morten W. <war...@on...>]

2002-02-08:
 - Initial work on function MTMresolveURL (becoming method
   resolveURL() of object MTMBrowser()), giving the rough outline of
   the function as well as those parts of MTMCheckURL() that are
   useful.
 - Bumped version number to 2.3.2-alpha.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Development of 2.3.2

[Morten W. <war...@on...>]

It's time to start working on v2.3.2.  The development of this version
will be interesting since Michael Richards is joining as a
co-developer, a first for this project (but hopefully not the last).

I'll be starting off by looking at feature request #490933, writing a
URL resolver for the menu so it'll start working with relative URLs
using "../".  since preHREF is a property of the MTMBrowser() object
I'm thinking of making MTMBrowser.resolveURL(), which will take a
given URL as its argument and resolve it.  if it's an absolute URL
it'll just return the given URL, if it's relative it'll be resolved
relative to MTMBrowser.preHREF, which is how the menu works now.

since it'll be a replacement to MTMCheckURL() (which does some of this
functionality now) it'll also take a second argument, and return
true/false depending on whether the given URL is relative or not.

then I'll be looking into bugs 510929 about the constant scrollbars in
Mozilla/Netscape 6, and 512447 which is the buggy scrolling.

after that I'll look into something else :)


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-02-03

[Morten W. <war...@on...>]

I'm posting this even though it's 3 days since the release just to
make sure it's publicly annonced.

2002-02-03:
 - Updated CREDITS & ChangeLog so they're ready for release of v2.3.1


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-01-29

[Morten W. <war...@on...>]

couldn't go to bed without fixing this:

2002-01-29:
 - Fixed bug #509834 by adding a small while() loop to function
   MTMinitializeDOMDocument(), thereby removing any existing
   content in MTMUA.document.body.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-01-28

[Morten W. <war...@on...>]

tackled the IE4 right-click message bug, and failed.  started doing
various odd things that need to be done before release:

2002-01-28:
 - Created default values for all user-configurable options.
   These will be documented when the install guide is rewritten,
   which is before release, of course.
 - Tried to fix bug #509230, but IE4 refused to let me fix it.
   Therefore MTMrightClickMessage remains for IE4 users, with the
   alert() that unfortunately follows it.
 - Converted all usage of spaces to indent lines to tabs.  Those who
   still want 2-space indendation should set their tab-settings
   accordingly


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-01-27

[Morten W. <war...@on...>]

bugfixes and more bugfixes (and the introduction of a bug more):

2002-01-27:
 - Fixed bug #425628 by removing the mouseover-related arguments to
   MTMakeLink() in the case where MTMEmulateWE is false and submenu
   item has a URL.
 - Fixed bug #439482 by adding calls to focus the link after the menu
   is refreshed.  This created bug #509210, and ideally should have
   some more checking before it occurs.  See also feature request
   #503543 about this.
 - Fixed bugs #494822 & #426664 by testing some of the item's properties
   before returning true from MTMSubAction().  in some cases it'll then
   do a forced call to MTMDisplayMenu().
 - Added a teste for the existence of the stored tracked item's frame
   before loading it, just in case we store a frame that isn't part of
   the parent frameset.
 - Fixed bug #439473 by removing the test for MTMFirstRun from the if()-
   statement handling the scroll code.  if either MTMTrackedItem or
   MTMClickedItem is set there _is_ an item we can scroll to in the
   document, so no need to test for MTMFirstRun.  Particularly useful
   now that we store both the expansion state and the tracked item.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-01-25

[Morten W. <war...@on...>]

two small bugfixes today.  wrapped up the NN4.x document.anchors[] bug
that arrived yesterday so that everything regarding the IE6
back-button bug were done.  also fixed the O5-bug which appears when
using a dynamic frameset.  in other words:

2001-01-25:
 - Fixed the NN4.x error regarding MTMUA.document.anchors[] that was
   introduced yesterday.  This means bug #475697 is closed.
 - Fixed bug #459277, Opera 5.12 creating a JS error on startup.  This
   was due to O5 not creating the 'parent' object before all frames were
   loaded.  I fixed this by re-assigning the onload-call to MTMStartMenu()
   to parent.onload in O5 and that seemed to fix the bug.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-01-24

[Morten W. <war...@on...>]

worked on cleaning up the code and getting support for IE4 back.
extensive comments are below.  note that NN4.x throws JS-errors now,
something I'll look into next.

2001-01-24:
 - Removed variable MTMDoc.  New property of MTMUA object, .document,
   is used instead.
 - MTMenuFrame is now only a user configurable variable (it was for a
   moment an object reference to the menu frame.  New property of
   MTMUA object, .menuFrame, is used instead.
 - Changed function MTMgetFrames() since it's now a method of the
   MTMBrowser() object.  References 'this' and calls this.setDocument()
   when found frame name corresponds to MTMenuFrame.
 - Added method setDocument() of object MTMBrowser() pointing to function
   MTMsetDocument)().  This function is new and gets a frame object
   reference as the argument (from function MTMgetFrames()).  Will set
   properties 'menuFrame', 'document', and if necessary 'head'.
 - Added method appendElement() of object MTMBrowser(), referencing
   function MTMappendElement().  This function is also new, and used for
   browsers support DOM methods (IE5+ and Mozilla/Netscape 6).  For more
   information about the method, see its extensive comments in mtmcode.js.
 - Global variable MTMTable is removed, instead property 'menuTable' of
   object MTMBrowser() (which means MTMUA.menuTable) is used.
 - Re-introduced script element in menu document for MTMRightClickMessage.
   This is to give the ability to prevent right-click also for IE4 users
   since linked .js-files is not supported in that browser.  All other
   browsers are to use a linked .js-file.
 - all if()-statements should now be correct, IE4's working again.
 - Netscape 4.x produces some errors regarding MTMUA.document.anchors[],
   will look into that.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-01-23

[Morten W. <war...@on...>]

continued work on fixing bugs left from fixing the IE6 bug.  the
Mozilla stuff worked, but slowed the speed down, which in turn means
further development is needed later to speed it up again.  IE4 gave
plenty of headaches since it constantly crashes when writing a script
element with src set to the document, and none of the other options
seemed to work.  IE5, IE6, NN4.x and Mozilla behave quite nicely.

2001-01-23:
 - Mozilla 0.9.7 doesn't support the scrollbar properties, so they're
   not inserted into the document at all, fixed with an appropriate
   if()-statement.  also, Moz likes having 'height: 100%' for body, if
   not it doesn't repeat the background image like one wants it to.
 - right-click prevention code moved to right-click.js.  file will be included
   at a later stage.
 - added configuration variable MTMLinkedInitFunction which is the name of
   the function found in the linked .js file that initializes setup.  this
   is _required_ for a linked .js file to work.
 - IE4 doesn't seem to want to support a linked .js file, it only hangs.
   attempts to use insertAdjancentHTML() didn't work, neither did attempts
   to give the script and ID and use the .src property to load the .js-file.
 - menu doesn't work in IE4 at this time.  will split the document output
   code ASAP to ease maintenance, debugging and development.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-01-21

[Morten W. <war...@on...>]

no large changes.  most of the time spent today went into researching
dynamic adding of style sheets in Mozilla (not supported, I found out).
apart from that:

2001-01-21:
 - changed some of the more complex if()-statements back to
   "if(!MTMUA.DOMable)" since they didn't work correctly with regards
   to Internet Explorer.  am believing that it would be a lot simpler
   to split the document handling code into one that keeps using
   document.write() and one that doesn't.
 - as per comment added to bug #475697 the root element was previously
   removed from the table on each refresh.  that is no longer done and
   the appropriate if()-statement wrapped around MTMAddCell() for the
   root item.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-01-17

[Morten W. <war...@on...>]

fixed bug #503595, that a sub menu which is set as expanded doesn't
also expand the sub's parent items.  also did the rest of the work
needed to get the IE6 back-button bug finally fixed.  introduced
another two bugs: linked .js-files currently do not work in IE5, IE6
and Mozilla, and also that Mozilla doesn't use the user-defined
colors, fonts, etc.  that part of the code is currently IE-centric
until I have found out how to create a dynamic inline style sheet in
Mozilla.  changelog for today therefore is:

2001-01-17:
 - the IE6 back-button bug is no more, since the initial document now is
   created only through dynamically inserting HTML elements.  the current
   code contains a few challenges though, like how to insert the linked
   script element and the right-click prevention code.  currently broken
   are:
   - linked .js-file
   - user-defined colour, font, etc
     (since they rely on an inline style sheet)
   current code is very rough, cleaning is needed, but it works.
 - removed the test for isExpanded and instead added an if()-check to the
   parentNode-adding for()-loop, so that if any submenu item is expanded
   its parent element (which contains the submenu we iterate over) is
   expanded.  this correctly fixes bug 503595, the previous patch didn't.

2001-01-16:
 - added a test for isExpanded and a quick while() loop to fix bug 503595
   so that an expanded submenu also expands its parent items.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-01-13

[Morten W. <war...@on...>]

continued work on the IE6-related bug as well as other bugfixes and
improvements:

2001-01-13:
 - current codebase is not usable due to the rewrite of the initial document
   creation, which is trying to fix the IE6 backbutton bug.  it can therefore
   not be used for anything besides code-examination.
 - removed function MTMakeA() and MTMakeBackImage() as they were no
   longer used. - added object MTMstyleRuleSet() with method .addRule()
   and changed the option MTMExtraCSS so it uses this ruleset.  this
   is a result of working towards more usage of dHTML capabilities, of
   which one meant rewriting the whole CSS-output code.
 - fixed a bug in the browser-detection code which didn't correctly identify
   the minor versions of MSIE.
 - changed MTMClickedItem to be an object reference instead of a number,
   and changed the affected if()-statements to check MTMClickedItem.number.
   the change helps dynamically loading submenus since one has a reference
   to the expanded/closed submenu throught MTMClickedItem and won't have to
   write a lookup function.
 - bugfixed MTMAlwaysLinkIfWE which didn't correctly create the whole submenu
   item as a link when the option was turned off.

2001-01-11:
 - changed timeout in mtmtrack.js to 50ms to speed up refreshment of
   the menu.  50ms should be a good tradeoff between refresh speed
   and any slowdown of the page display from the menu update.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2002-01-02

[Morten W. <war...@on...>]

started the new year with some bug hunting, here's what went into CVS
last night:

2002-01-02:
 - added relocation script to code.html to fix the bug where IE
   bookmarks code.html.  simple script that tests for a frameset.
 - started using TAB instead of spaces at the beginning of lines for
   indentation, saving file size and easing maintenance.
 - initial work on fixing the IE6 "back button doesn't work" bug.
   IE4 has not got a lot of possibility for manipulation of the
   document head.  therefore a later update will instead separate
   IE4 from the rest and use more advanced dHTML.

2001-12-27:
 - added function MTMakeSVG(), given to me by Sachin Ghaisas and Chris
   Coulthrust.  the function can be used in replacement of MTMakeImage()
   to insert an SVG object into the menu document.  it is currently not
   used in any way, though.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2001-12-19

[Morten W. <war...@on...>]

small feature-add update:

2001-12-19:
 - added variable MTMCodeFrame (default: "code") which contains the
   name of the frame containing code.html.  enables usage of multiple
   menus.  this variable was also added to mtmtrack.js so that one can
   have the right frame called from the tracking code too.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.

[jsmenu-codedev] Changelog 2001-12-10, parts I & II

[Morten W. <war...@on...>]

mostly fixes to make menu building a lot easier, removing the need to
know what item index in the items[] array you need and also the need
for "new MTMenuItem()".  the menu building update is currently
undocumented, for an example of how the new structure looks, see <URL:
http://groups.yahoo.com/group/tree_menu/message/942 >

2001-12-10 (part II):
 - added MTMTrackedCookie as a global variable.  would cause a problem if
   cookies were turned off since there's an if(MTMTrackedCookie), browsers
   would throw an error since the variable doesn't exist.

2001-12-10:
 - added function MTMakeLastSubmenu() and created this as a method named
   "makeLastSubmenu", part of the MTMenuItem() object.
 - added method makeSubmenu, part of the MTMenuItem() object.  no need
   to name the methods the same as the function when they're part of an
   object.  foo.makeSubmenu() is much easier to read and work with than
   foo.MTMakeSubmenu().
 - added a test to MTMAddItem() converting the first argument to a string.
   if it's an object we add it as usual, if not we create a new MTMenuItem()
   object.  this makes menu creation even easier since users can now say
   menu.MTMAddItem("foo"); instead of having to create the new MTMenuItem().
 - added method addItem, part of the MTMenu() object.  again no need to
   name the methods the same as the function, addItem() is easier to read
   than MTMAddItem().  the function name stays the same to be unique though.


Morten!

-- 
"...a liquid which was almost, but not quite, entirely unlike tea."
  My opinions are not necessarily those of my provider,
  not necessarily mine, and probably not necessary.