How to express progress of longer run of JS?

[Martin Mrazek]

Hi,
I check data validity in html form by JS. Something like

for (i=0; i<document.for m[0].elements.lengt h; i++) {
chechkValidity( i);
}

Unfortunately, the form[0] has about two thousands elements (it is
statistical questioning form for companies) and execution of this one
cycle takes about twenty seconds.

The problem is, that during these 20 seconds the browser "freezes":
The button activating the cycle disappears and the mouse pointer
doesn't change into "watch" (sandglass) showing that something is
being carried out.

Although after the 20s everything is ok, it is not the nicest behavior
of the form. I don't assume somebody is so impatient to restart during
those 20 seconds. I just want to minimize the users' swearing on my
address and tha's why my question:

??? How to show the progress of that cursed for-cycle?

Thanks
Martin

[Andrew Thompson]

On Tue, 18 May 2004 19:06:03 +0100, Richard Cornford wrote:

Andrew Thompson wrote:
<snip>

(shrugs) Seems to work in both IE6 and
Moz 1.3 on XP. Did I miss something (still)?

We try not to think in terms of "both browsers" these days.

Aaah well. The only ones I had immediate
access to. I'd have to hunt down and install
NN 4.7, ..or Lynx off the filesystem. Just
checking now I also see 'pheonix' (shrugs
vaguely).

I know I have a CD with an older version
of Opera tucked away somewhere...

(wanders off, muttering..) ;-)

[Ian Richardson]

Martin Mrazek wrote:

Hi,
I check data validity in html form by JS. Something like

for (i=0; i<document.for m[0].elements.lengt h; i++) {
chechkValidity( i);
}

I'd suggest:

1. Create an array, e.g. myformstatus[]

1a. Set a new variable to setInterval("is Complete(myform )",20);

2. have a for loop disabling all the form elements, assuming that you
don't want user interaction until the process is complete, just user
feedback. Populate each element of myformstatus[] with a default value,
e.g. null.

3. use your loop with something like myform[i] =
setTimeout(chec hkValidity(i),1 0); so your function is called and the
browser is only hanging for the period of one execution of
chechkValidity( ) once.

[Note your setTimeout could be
("chechkValidit y(i);chechkVali dity(i+1);chech kValidity(i+2); ",10"); so
that these are executed in blocks rather than just one at a time,
depending on how intensive each call is...]

4. In chechkValidity( ), set myformstatus[i] to some other default value,
e.g. "done"

5. Create a function isComplete() which goes through myformstatus[] to
check that all elements have a value of "done". If so, clear the
setInterval() and then have a for loop reenabling all form elements.

Then in your step 3 or 4 you can do whatever feedback to the user you can.

Note this may cause the overall process to take _longer_, but if you do
it in one for loop as you've described you'll hit three problems:

1. It may cause the browser to freeze so long that the browser offers
the option for the user to abandon it, thus potentially causing all
manner of state-related problems. I've found that recent builds of
Mozilla seem to have shortened this time considerably, but there's no
easy way I've found of pre-determining how long this actually is.

2. You won't be able to get feedback to the user, as the browser needs
time to breathe between bursts of Javascript...

3. A browser freeze may be considered by the user as a browser crash!

Just my thoughts - I've not tested this out but it seems to make sense,
at least to me!

Thanks,

Ian

[Jim Ley]

On Tue, 18 May 2004 20:14:34 +0100, Ian Richardson
<za*****@chaos. org.uk> wrote:

Martin Mrazek wrote:

Hi,

3. use your loop with something like myform[i] =
setTimeout(che chkValidity(i), 10); so your function is called and the
browser is only hanging for the period of one execution of
chechkValidity () once.

So on win98 this will add a minimum of 106 seconds? (53ms minimum
setTimeout resolution *2000 items, or is my maths out?)

(not that I don't disagree with the need to use setTimeout, but
remember it's a lot of overhead.)

Jim.
--
comp.lang.javas cript FAQ - http://jibbering.com/faq/

[Dr John Stockton]

JRS: In article <40************ ****@news.indiv idual.net>, seen in
news:comp.lang. javascript, Jim Ley <ji*@jibbering. com> posted at Wed, 19
May 2004 22:53:18 :

So on win98 this will add a minimum of 106 seconds? (53ms minimum
setTimeout resolution *2000 items, or is my maths out?)

More exactly, (2000*) 54.9255 milliseconds. 86400000 / 0x1800B0.

Anyone know of javascript systems where the time quantum is NOT either
55 ms (possibly rounded to 50/60) or 10 ms? RSVP.

--
© John Stockton, Surrey, UK. ?@merlyn.demon. co.uk Turnpike v4.00 MIME. ©
Web <URL:http://www.merlyn.demo n.co.uk/> - w. FAQish topics, links, acronyms
PAS EXE etc : <URL:http://www.merlyn.demo n.co.uk/programs/> - see 00index.htm
Dates - miscdate.htm moredate.htm js-dates.htm pas-time.htm critdate.htm etc.