need help debugging opensource script on Mac IE

[Patrick Fitzgerald]

I have a GNU licensed Javascript slideshow script:
http://slideshow.barelyfitz.com/

Recently some users have complained about problems in Mac IE; unfortunately
I don't have a Mac system that I can test. It runs cleanly on Windows IE
and Mozilla.

Is anyone willing to take a look at this?
Your fellow Mac users will thank you. :)
--
Patrick Fitzgerald - pa*@barelyfitz. com
BarelyFitz Designs - http://www.barelyfitz.com/

[Nobody]

"Patrick Fitzgerald" <pa*@barelyfitz .com> wrote in message
news:Xn******** **************@ 216.77.188.51.. .
| I have a GNU licensed Javascript slideshow script:
| http://slideshow.barelyfitz.com/
|
| Recently some users have complained about problems in Mac IE;
unfortunately
| I don't have a Mac system that I can test. It runs cleanly on Windows IE
| and Mozilla.

Have you tried Opera and Netscape as well?

I see one problem at a glance. The innerHTML reference is used without
checking if it is supported. This is bad news for a lot of agents. My own
slideshow works on Mac IE, but as I recall, it required some careful object
detection to make it work (Mac IE is one of the strangest animals out
there.)

Usual suspects are things dealing with filters (BTW, your slideshow will
lose transitions for non-IE browsers if it uses filters), plugins, audio,
etc. Only the first one applies here based on a quick glance at the script.

You can make your slideshow work with transitions in Gecko based browsers
with the moz-opacity style (search Google for examples.) Your dynamic
document must hide this style from everything BUT Gecko. How to do that is
left as an exercise.

|
| Is anyone willing to take a look at this?
| Your fellow Mac users will thank you. :)
| --
| Patrick Fitzgerald - pa*@barelyfitz. com
| BarelyFitz Designs - http://www.barelyfitz.com/

[Richard Cornford]

"Nobody" <no**@nope.ne t> wrote in message
news:bK******** **********@fe3. columbus.rr.com ...
<snip>

with the moz-opacity style (search Google for examples.)
Your dynamic document must hide this style from
everything BUT Gecko. How to do that is left as an
exercise.

Why? The CSS specification clearly states that unrecognised properties
are to be ignored by conforming implementations . It might be necessary
to hide the CSS from dubious implementations like Netscape 4 but there
are plenty of CSS based methods for achieving that.

Richard.

[Nobody]

IE is highly dubious as well. You absolutely cannot hide everything you
need to hide with CSS tricks involving comments and the like. Server-side
script is needed.

"Richard Cornford" <Ri*****@litote s.demon.co.uk> wrote in message
news:bj******** ***********@new s.demon.co.uk.. .
| "Nobody" <no**@nope.ne t> wrote in message
| news:bK******** **********@fe3. columbus.rr.com ...
| <snip>
| >with the moz-opacity style (search Google for examples.)
| >Your dynamic document must hide this style from
| >everything BUT Gecko. How to do that is left as an
| >exercise.
|
| Why? The CSS specification clearly states that unrecognised properties
| are to be ignored by conforming implementations . It might be necessary
| to hide the CSS from dubious implementations like Netscape 4 but there
| are plenty of CSS based methods for achieving that.
|
| Richard.
|
|

[Richard Cornford]

"Nobody" <no**@nope.ne t> wrote in message
news:sc******** ***********@fe1 .columbus.rr.co m...

IE is highly dubious as well. You absolutely cannot hide
everything you need to hide with CSS tricks involving comments
and the like. Server-side script is needed.

A server-side script stands less chance of determining the quality of
client side CSS support than it has of working out what software is
operating at the client end.

Richard.

[Patrick Fitzgerald]

"Nobody" <no**@nope.ne t> wrote in
news:bK******** **********@fe3. columbus.rr.com :

| http://slideshow.barelyfitz.com/

Have you tried Opera and Netscape as well?
Yes, no problems in Opera or Netscape 4.8
I see one problem at a glance. The innerHTML reference is used
without checking if it is supported. This is bad news for a lot of
agents.
Thanks for catching that. I think at one point I checked like this:

if (r.innerHTML)

but that caused a problem when innerHTML was set to "". I guess I'll have to
check like this:

if (typeof r.innerHTML != 'undefined')

I don't know if this will solve my Mac problem though.
My own slideshow works on Mac IE, but as I recall, it
required some careful object detection to make it work (Mac IE is one
of the strangest animals out there.)
So it would seem.

Usual suspects are things dealing with filters (BTW, your slideshow
will lose transitions for non-IE browsers if it uses filters),
plugins, audio, etc. Only the first one applies here based on a quick
glance at the script.

Yeah, I just added the IE filters on request; I'll eventually add a cross-
browser solution.

Thanks again.
--
Patrick Fitzgerald - pa*@barelyfitz. com
BarelyFitz Designs - http://www.barelyfitz.com/

[Nobody]

I pass it the DOM-indicated UA string. After that it is all based on
empirical evidence (testing on numerous browser/platforms.) Clearly CSS
support is different across different platforms and browser versions and the
only way to keep track of such things is with a database. It is an
imperfect world with lots of imperfect browsers and this is an imperfect
method, but effective enough to be quite practical. If half of the web
sites out there competently supported NS, IE, Moz and Opera (and okay throw
in Konquerer) across Windows, Mac and Unix, things would be 1000% better
than they are today (where most sites don't seem to support anything well.)
If half of them could learn to suppress tables when confronted with cell
phones (well, mine anyway) then we would really be getting somewhere. I
have a tool that can make all of them at least this good (ASP based, watch
for it soon!) So call me crazy!

"Richard Cornford" <Ri*****@litote s.demon.co.uk> wrote in message
news:bj******** ***********@new s.demon.co.uk.. .
| "Nobody" <no**@nope.ne t> wrote in message
| news:sc******** ***********@fe1 .columbus.rr.co m...
| >IE is highly dubious as well. You absolutely cannot hide
| >everything you need to hide with CSS tricks involving comments
| >and the like. Server-side script is needed.
|
| A server-side script stands less chance of determining the quality of
| client side CSS support than it has of working out what software is
| operating at the client end.
|
| Richard.
|
|