[DynAPI-notify] [ dynapi-Bugs-1255125 ] Html input bug.

[SourceForge.net <no...@so...>]

Bugs item #1255125, was opened at 2005-08-09 13:27
Message generated for change (Settings changed) made by warp9pnt9
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=105757&aid=1255125&group_id=5757

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
>Group: Verified
>Status: Closed
Resolution: None
>Priority: 2
Submitted By: qq365 (qq365)
>Assigned to: L W (warp9pnt9)
Summary: Html input bug.

Initial Comment:

1.in firefox 1.0.4, open "javascript console";

2.open your example 
file : "dynapi/examples/dynapi.gui.htmltextbox.html"

3.Move your mouse over the input control, you will get 
lots errors' info as following:
---------------------------------------------------------------
Error： [Exception... "'?^' 
HTMLDivElement.parentNode ?CP
3' when calling method: 
[nsIDOMEventListener::handleEvent]"  
nsresult: "0x8057001e 
(NS_ERROR_XPC_JS_THREW_STRING)"  
location: "<unknown>"  data: no]
---------------------------------------------------------------

----------------------------------------------------------------------

Comment By: L W (warp9pnt9)
Date: 2005-08-09 21:15

Message:
Logged In: YES 
user_id=706287

I believe this is an old old old browser bug.  But your text
string is mangled or incomplete.  The complete is as follows.

Error: [Exception... "'Permission denied to get property
HTMLDivElement.parentNode' when calling method:
[nsIDOMEventListener::handleEvent]"  nsresult: "0x8057001e
(NS_ERROR_XPC_JS_THREW_STRING)"  location: "<unknown>" 
data: no]

I first noticed it with Mozilla 1.4 about 2 years ago.  My
comments are burried in the dynapi-dev mailing list. Text
input and Textarea fields trigger this, maybe others.  The
error message did not appear with Mozilla 1.3.

This is also a known bug at Mozilla which nobody has ever
adressed.  People keep reporting it and it keeps getting
marked as a duplicate and people have long discussions about
why they're right or the other person is wrong but nobody
has resolved this.

It does not appear to prevent anything from working
correctly, as far as I can tell, so I've just been ignoring
it.  Go to Mozilla's Bugzilla database and ping the bug
report(s), if you can find the original in the quagmire of
unmarked duplicates.

The bug has nothing at all to do with DynAPI.  You can quite
simply reproduce the bug with just some standard JavaScript
event handler.  I don't know of any JavaScript workaround to
make this bug silent.  It is agreeably annoying when trying
to debug JavaScript when your mouse will pass over text
input and textarea fields.

The only solution is to fix the browser.  It may not be a
coding problem but a design problem deep seated in someone
else's stubborn policy.  Or merely just a lack of people to
get to such a low priority bug, when faced with other more
critical security fixes.

The only way to fix it is to get the source code and learn
how to build it.  The only way to do that is with a full
build environment (which was not all that well documented at
the time), and to build the browser code and fiddle with it.
 Just a very thin Firefox client requires a good 15-30GB
(yes, GigaBytes) of disc space for all of the libraries and
binaries that get built as separate objects with all their
debugging symbols and the libraries composed of these
modules, and the binaries with these libraries statically
linked.  And then you will likely need extra tools to have
an installer build environment, and then build an installer
from config files.

It's definitely a non-trivial task to get that far even. 
Then you've got to learn the source layout, the internal
APIs and whatnot, probably little or no documentation or
support.  You've got to have a debugging environment.  Then
even if you manage to fix this, you have to make sure you
fix it correctly, generate a patch, unpack a clean source
tree, apply the patches, see if it builds, and then do this
in Linux to make sure it doesn't break anything.

And then probably the hardest task is to lobby for the code
to be applied.  You'll have to convince someone at Mozilla
that it will address the problem without breaking anything,
and present evidence of your rigorous testing.  You'll sound
better if you can convince other developers to apply your
patch and try it, and view your very simple JavaScript/HTML
Form test code.  Probably by attaching a patch to the bug
report, which will again go unnoticed for another few years
and may never be applied in the end.

But if it's going to have any chance at all, different
people have to fix it themselves, test and verify that
patches apply cleanly, and make noise about it.  I already
made my noise, and I couldn't figure out how to build it
without any proper documentation or support, and other
things became more important.  If I continue to make noise
they just ignore the whole thing entirely as "too noisy" or
something.  :p

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=105757&aid=1255125&group_id=5757