dynapi-dev

[Dynapi-Dev] ns6 w/h and browser advocacy

[Cameron H. <ca...@bi...>]

Hi,

I remember a post from Robert saying he'd got offsetWidth and offsetHeight
working in NS6 using a setTimeout. Was this specific to one of the widgets
Robert, or did it fix W/H for everything in NS6?

If you need a setTimeout for the thing to work properly, it sounds like a
bug in the browser. The great thing about Mozilla is if you post a bug,
they'll fix it for you ;-) It can take a while, it depends on how important
the developers feel that peice of functionality is. But I think it would be
worth writing a test case and a bug report for the W/H bug. I can submit one
if someone (Robert I guess) tells me how to reproduce the bug, that is how
it works versus how you think it should work.

Something else which might be if interest to DOM enthusiasts, I was
wondering why offsetHeight and offsetWidth and similar measurements aren't
in any of the DOM's. I posted a query about this to ww...@w3..., and the
response was it's been put on the back burner as there hasn't been that much
interest in this functionality. Apparently it will probably end up in DOM3.
I thought it would have been pretty core functionality...

Re: [Dynapi-Dev] ns6 w/h and browser advocacy

[Joachim L. <lu...@ho...>]

Time for me to bud in... I did quite a lot of testing for this bug some 
time ago.

Mozilla.org had issued the bug at that time as #56810 :
         http://bugzilla.mozilla.org/show_bug.cgi?id=56810

I never got the work-around to, well, work :-(

But I got the right result by calling the function as the result of a timeout:
         setTimeout("alert("+o+".elm.offsetWidth)", 0);
In other words: timeout in 0 ms - just after the layout engine has done its 
deed

Now I checked bugzilla again, and there is another one:
         http://bugzilla.mozilla.org/show_bug.cgi?id=65548

Guess it never got fixed...

The problem as I see it is this: Because the bug is in Netscape 6 (I 
assume, the installer crashes for me, but I use Mozilla 0.6 which is 
supposed to be from the same codebase) the fix must work for the release 
version.

I was hoping there was a way to programmatically get the result, such as
         this.elm.ownerDocument.defaultView.getComputedStyle(this.elm,null).getPropertyValue("width");
but whatever I tried I couldn't get it to work.

The only reliable way was the setTimeout() trick. I think the easiest for 
you to make it work in DynAPI is to skip the precreation code for that 
browser, at least when the width and height aren't known.

(The code here is from memory - it seems to be right though.)

/Lunna


At 2001-01-30 11:24, you wrote:
>Hi,
>
>I remember a post from Robert saying he'd got offsetWidth and offsetHeight
>working in NS6 using a setTimeout. Was this specific to one of the widgets
>Robert, or did it fix W/H for everything in NS6?
>
>If you need a setTimeout for the thing to work properly, it sounds like a
>bug in the browser. The great thing about Mozilla is if you post a bug,
>they'll fix it for you ;-) It can take a while, it depends on how important
>the developers feel that peice of functionality is. But I think it would be
>worth writing a test case and a bug report for the W/H bug. I can submit one
>if someone (Robert I guess) tells me how to reproduce the bug, that is how
>it works versus how you think it should work.
>
>Something else which might be if interest to DOM enthusiasts, I was
>wondering why offsetHeight and offsetWidth and similar measurements aren't
>in any of the DOM's. I posted a query about this to ww...@w3..., and the
>response was it's been put on the back burner as there hasn't been that much
>interest in this functionality. Apparently it will probably end up in DOM3.
>I thought it would have been pretty core functionality...

Re: [Dynapi-Dev] ns6 w/h and browser advocacy

[Michael Br. <mb...@st...>]

> From: Joachim Lundgren <lu...@ho...>
> Reply-To: dyn...@li...
> Date: Tue, 30 Jan 2001 15:15:59 +0100
> To: dyn...@li...
> Subject: Re: [Dynapi-Dev] ns6 w/h and browser advocacy

[...]

> But I got the right result by calling the function as the result of a timeout:
> setTimeout("alert("+o+".elm.offsetWidth)", 0);


this is also true for IE5 on the Mac, but you have to wait a little longer.
I did some tests some time ago and there wasn't a safe time, in which the
values would get updated. So I implemented a function (no DynAPI) that
stores the old values and then checks every few ms if the values have been
updated and if so, it calls whatever you like.
This was the only solution I could come up with, I haven't ever found a clue
on how it could be done differently. ...

the drawback is of course that you can't use return values to directly do
more stuff with it, there has to be another function that does the things
you want to do after you know the size of the content.
if this mechanism would be implemented into the DynAPI, it would change the
way one has to deal with getContentWidth/getContentHeight (maybe using a
"onThe_moment_our_contentsize_values_are_finally_correct"-event or something
like that)

i really need this feature in IE5/Mac too, so if nobody comes up with a
better solution, I will definitely implement it that way.

--
Michael Buerge

BTW: i have the impression that there arent any Mac-developers here. the
examples in this and last weeks snapshots mostly didn't work on the Mac and
there wasn't one post regarding that...
...and what about the bug i posted (elm problem)? can somebody reproduce it?
can somebody help me figuring it out?

RE: [Dynapi-Dev] ns6 w/h and browser advocacy

[Matthew A. S. <ms...@go...>]

I have to support the macintosh as well, though because I get very little
responses on this list I usually have to limit myself to the things that I
can get to work.  Not being able to get the content size has been a very
irritating problem for me.  Usually I end having to set sizes manually which
can cause things to look a little wierd on some of the browsers.

M.

-----Original Message-----
From: dyn...@li...
[mailto:dyn...@li...]On Behalf Of Michael
Burge
Sent: Tuesday, January 30, 2001 3:38 PM
To: dyn...@li...
Subject: Re: [Dynapi-Dev] ns6 w/h and browser advocacy


> From: Joachim Lundgren <lu...@ho...>
> Reply-To: dyn...@li...
> Date: Tue, 30 Jan 2001 15:15:59 +0100
> To: dyn...@li...
> Subject: Re: [Dynapi-Dev] ns6 w/h and browser advocacy

[...]

> But I got the right result by calling the function as the result of a
timeout:
> setTimeout("alert("+o+".elm.offsetWidth)", 0);


this is also true for IE5 on the Mac, but you have to wait a little longer.
I did some tests some time ago and there wasn't a safe time, in which the
values would get updated. So I implemented a function (no DynAPI) that
stores the old values and then checks every few ms if the values have been
updated and if so, it calls whatever you like.
This was the only solution I could come up with, I haven't ever found a clue
on how it could be done differently. ...

the drawback is of course that you can't use return values to directly do
more stuff with it, there has to be another function that does the things
you want to do after you know the size of the content.
if this mechanism would be implemented into the DynAPI, it would change the
way one has to deal with getContentWidth/getContentHeight (maybe using a
"onThe_moment_our_contentsize_values_are_finally_correct"-event or something
like that)

i really need this feature in IE5/Mac too, so if nobody comes up with a
better solution, I will definitely implement it that way.

--
Michael Buerge

BTW: i have the impression that there arent any Mac-developers here. the
examples in this and last weeks snapshots mostly didn't work on the Mac and
there wasn't one post regarding that...
...and what about the bug i posted (elm problem)? can somebody reproduce it?
can somebody help me figuring it out?



_______________________________________________
Dynapi-Dev mailing list
Dyn...@li...
http://lists.sourceforge.net/lists/listinfo/dynapi-dev

Re: [Dynapi-Dev] ns6 w/h and browser advocacy

[Dann <da...@to...>]

Hi,

Has anybody ever done some testing or wondered about the safety nature of
concurrent execution of code in JS ? It struck me that this timeout gimmick has
been discovered over and over again, to apparantly fix a recurring issue. It was
already pointed out by Cameron, that if you could fix this kind of behaviour by
using a Timeout, you probably stumbled on a browser bug... usually, in my years
of fooling around with JS, a Timeout could even prevent a browser from crashing
entirely.

What exactly happens if you would hook load event handlers to to two images, who
would fire simultaneously and call the same function, to change some common
global variables ?

CU,
 Dann