The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.
R1_1_6-19980514:

## +++. Leakage: make new method to replace print() that does cleanup
##    too. -- USE THIS: render() 
## 0. fix >< in commments (i think this was done in 1.001005)
## x. make default constructor, move processing into tl_init()... in
##    widgets, you shouldn't deal with redefining new() as a
##    consequence, only tl_init().
##    8. i have a feeling someone is going to tell me that my
##    constructors suck and need to be rewritten (doh! I spoke too
##    soon; someone already has!)
## x. improve componentcontainer class to be a mroe usefule generic
##    container...also export symbol "container" to take advantage of
##    it.
## 6. add clone() to the (full) API (based on Data::Dumper)
## x. remove internal compatibility _isa--this breaks compatibility
##    with older perls... sorry.  speeds up everyone else.
## x. change exception handling to use die/warn.


R1_1_5-19980416:  

## 1. fixed another sawanampersand problem pointed out by AK.
## 2. slipped the forms inside the table (and cell) so they don't
##    cause an extra space in the containing table (if any).
## 3. added support for breaks after list elements
## 4. fixed problem with >< in FormComponentMulti items, plus can now
##    embed real components (not just text) in FormComponentMulti.
## 9. Make a generic container for grouping things together (no
##    formatting)... call it "box" or somesuch.2/8/1998 11:43pm
##    (modified componentcontainer to have a tl_print)
## 11. added "getName()" to Form class... this will get the
##     appropriate "name"... if there is no name, or you set the
##     "force_numeric" flag, then it returns a number for the name.
##     This number, of course, happens to correspond to the number of
##     the component in the javascript document model.  this is
##     important for supporting IE3, which doesn't support symbolic
##     naming of such.
## +. moved form back *outside* of the cell, otherwise it incurs a
##    mess of extra padding. (but should be *inside* table... gotta
##    love it)
## +. added setParams()
## +. added HTML::TableLayout::VERSION variable.
## +. tinkered with padding of text (tl_do_not_pad()) 4/14/1998 
## +. added delete() method to base class.  NOT YET IMPLEMENTED. 4/14/1998 


R1_1_4-121297: Patches from andreas koenig to reduce number of
warnings w/-w flag; applied a.k.'s recommendation to make the tags
always meet as such: "><" to avoid abberations in how some browsers
(notably netscape) format tables becauses of spaces between the table
tags.  Plus now the HTML output looks decidely funky.

R1_1_3-120897: Changed version number to be Perl compliant; Changed
license from GPL to Perl-like (Artistic) as per suggestion by andreas
koenig; Updated comments;Fixed some form handling bugs

R1_1_2-110697: This will be the first fully public release of this
module as I am quite happy with the API, and thus it should remain
stable for at least a little while =).  I removed all use of
tl_strav(), which, incidentally, was correctly named, as it was a
travesty.  Also removed another loser: tl_setupComponents().

D1_1_1-102697: Changed timing model to add "setup" phase as a distinct
phase, numerous other changes and bugfixes, but still wants some more
cleanup.  Added some new components, changed name of select to choice.

D1_1_0-101197: This IS the first public release.  Improvements
include (a) much more organized OOness and object heirarchy (b)
redid "timing" so now there is "late" and "early".  Stuff happens
late that needs to.  This is good (it means you can stick things
in other things without thinking about timing at all). (c) added
many more comments to code, improved numerous variable and method
names.  (d) changed "cell header" mechanism to use embedded tables.
cell header can be left,right,top or bottom.  (e) every component
gets a pointer to its container, its window, and its form; it can
now find these efficiently and reliably (no dangers of deep nesting
confusing us).

D1_0_0-092997: This snapshot is after initial profiling (&
improvements), and after fixing the spelling of the word component.
This should be considered the "stable" snapshot before reorganizations
and improvements that have been brewing over the past week or so.

(Just as a note, "D" prefix for a CVS tag means it's earmarked for
development only... "R" is for releases.)