connot set handlers on XMLHttpRequest instance in FF1.5

[Spasmoid]

I have the following code running in FireFox 1.5 (a remote XUL
application):

var req = new XMLHttpRequest();
// execution will block until request is
completed
// third param set to false makes request
synchronous
req.open(strMethod, uriTarget, false);
if(strMethod == "POST") {

req.setRequestHeader('Content-Type','application/x-www-form-urlencoded;
charset=UTF-8');
}
req.onreadystatechange = function () {
if (req.readyState == 4) {
if(req.status == 200) {
strResult = req.responseText;
xmlResult = req.responseXML;
}
else {
strLastError = uriTarget + " returned http error code " +
req.status + "\n" + req.statusText;
if(intVerbosity > 1) throw new Error(strLastError);
}
}
}
req.onerror = function() {
strLastError = uriTarget + " fetch failed while " +
arrReadyState[req.readyState];
if(intVerbosity > 1) throw new Error(strLastError);
}
req.send(strRequest);

By the time it gets to req.send() (at the bottom) all the callback
functions like req.onerror still have values of null. I've verified
this using the Venkman debugger with req as a watch variable.

the code is clearly assigning functions to those callback function
objects. Any idea why they are all turning up null?

No errors are showing up in the JS console.

 

[Thomas 'PointedEars' Lahn]

I have the following code running in FireFox 1.5 (a remote XUL
application):

var req = new XMLHttpRequest();
// execution will block until request is
completed
// third param set to false makes request
synchronous

Code should be posted (and written in the first place) so that it is not
subject to potential word-wrap, that is, it should not exceed 80 characters
per line.

req.open(strMethod, uriTarget, false);
[...]
req.onreadystatechange = function () {
[...]
}
req.onerror = function() {
[...]
}
req.send(strRequest);

By the time it gets to req.send() (at the bottom) all the callback
functions like req.onerror still have values of null.

I've verified this using the Venkman debugger with req as a watch
variable.

the code is clearly assigning functions to those callback function
objects. Any idea why they are all turning up null?

`onreadystatechange' is for asynchronous requests only. I do not
know why `onerror' is `null' even if you assigned a value to it;
that may have to do with the synchronous request as well.

<URL:http://xulplanet.com/references/objref/XMLHttpRequest.html>


HTH

PointedEars

 

[Spasmoid]

Ahh yeah. Forgot about that. Thanks for reminding me.
Dum mistake

I will try to remember to make the code post friendly next time.

 

[Randy Webb]

Thomas 'PointedEars' Lahn said the following on 12/22/2005 8:18 PM:

Spasmoid wrote:




Code should be posted (and written in the first place) so that it is not
subject to potential word-wrap, that is, it should not exceed 80 characters
per line.

To state that code should be written at less than 80 characters per line
is ludicrous. There is no limitation in the language that even makes
that remotely related.

If your text editor wraps code at 80 characters, you don't need to
change your coding style, you need to change text editors.

 

[Lasse Reichstein Nielsen]

To state that code should be written at less than 80 characters per
line is ludicrous. There is no limitation in the language that even
makes that remotely related.

No, it is a limit in some, widely used, newsreaders, that they
automatically and unavoidably breaks lines longer than 80 characters
both when sending and receiving. I.e., The limit is not on code in
general, but on code to be posted to a newsgroup.

Shorter lines are recommended because it is easier to make people
write shorter lines than it is to make them use a newsreader that
doesn't munge longer lines - although in the long run, that would
be best for them.

If your text editor wraps code at 80 characters, you don't need to
change your coding style, you need to change text editors.

And when your newsreader breaks longer lines, you need to write
shorter lines when asking for help with your code, because the code
you post is not going to work otherwise. Sure, anybody competent
enough to help with the problem can probably also fix the improper
line breaks, but it adds extra work that will surely make some
readers, myself included, reconsider whether I don't have something
better to spend my time on (yeah right, but in theory .

/L

 

[Dr John Stockton]

JRS: In article <[email protected]>, dated Thu, 22 Dec
2005 22:32:44 local, seen in Randy Webb

Thomas 'PointedEars' Lahn said the following on 12/22/2005 8:18 PM:

To state that code should be written at less than 80 characters per line
is ludicrous. There is no limitation in the language that even makes
that remotely related.

If your text editor wraps code at 80 characters, you don't need to
change your coding style, you need to change text editors.

Code that is written to be read in Usenet News should be written,
whenever practicable (as it always is in javascript) with no more than
72 characters per line, since that is the accepted Usenet norm. Not 80;
not even 76, as called for by the fascist bully-boy, but 72.

Code that has not been written for News, but needs to be posted to News,
must not be machine-wrapped by the posting agent; it may be manually &
sensibly wrapped (in which case it should be re-tested before posting),
or it may be posted with longer lines.

A reader agent that cannot display longer lines faithfully is defective,
and should not be pandered to.

A text editor should wrap lines at the setting that its user chooses,
over a wide range of widths including 80.

Material to be read as text should be wrapped at about 72, even if code
is longer.

A copy of <URL:http://www.merlyn.demon.co.uk/js-quick.htm> can wrap
paragraphs to selected width - range (IIRC) 1 to enormous - though it
may not work if line-endings are not CRLF.

 

[Randy Webb]

Lasse Reichstein Nielsen said the following on 12/24/2005 8:55 AM:

No, it is a limit in some, widely used, newsreaders, that they
automatically and unavoidably breaks lines longer than 80 characters
both when sending and receiving. I.e., The limit is not on code in
general, but on code to be posted to a newsgroup.

The implication that I got from the statement that you snipped was that
one should code at less than 80 characters. Now, that is well true for
posting code, but it is ludicrous to state that about writing code in
general.

Shorter lines are recommended because it is easier to make people
write shorter lines than it is to make them use a newsreader that
doesn't munge longer lines - although in the long run, that would
be best for them.

But that still has no bearing on how I should/would write code that was
never to be posted to Usenet.

And when your newsreader breaks longer lines, you need to write
shorter lines when asking for help with your code, because the code
you post is not going to work otherwise. Sure, anybody competent
enough to help with the problem can probably also fix the improper
line breaks, but it adds extra work that will surely make some
readers, myself included, reconsider whether I don't have something
better to spend my time on (yeah right, but in theory .

I totally agree. But that was not my point about what he said. Re-read
what he posted and my reply. My reply is about coding in general, you
are referring to code posted to Usenet.

 

[Randy Webb]

Dr John Stockton said the following on 12/24/2005 11:57 AM:

JRS: In article <[email protected]>, dated Thu, 22 Dec
2005 22:32:44 local, seen in Randy Webb




Code that is written to be read in Usenet News should be written,
whenever practicable (as it always is in javascript) with no more than
72 characters per line, since that is the accepted Usenet norm. Not 80;
not even 76, as called for by the fascist bully-boy, but 72.

Agreed - with respect to Usenet.

Code that has not been written for News, but needs to be posted to News,
must not be machine-wrapped by the posting agent; it may be manually &
sensibly wrapped (in which case it should be re-tested before posting),
or it may be posted with longer lines.

Agreed - with respect to Usenet.

A reader agent that cannot display longer lines faithfully is defective,
and should not be pandered to.

A text editor should wrap lines at the setting that its user chooses,
over a wide range of widths including 80.
Material to be read as text should be wrapped at about 72, even if code
is longer.

None of which I was pointing out. The problem is his implication that
one should *always* write code at less than ~72 characters. And that is,
well, ludicrous

It is not until code is introduce to Usenet that it's line length's
become relevant.

 

[Lasse Reichstein Nielsen]

I totally agree. But that was not my point about what he said. Re-read
what he posted and my reply. My reply is about coding in general, you
are referring to code posted to Usenet.

While I can see that you can read Mr Lahn's original answer to refer
to all code, I could also understand it as only referring to the code
that is to be posted. Picking one reading of the posting and calling
it ludicrous, instead of just pointing out that the wording is less
than precise and making it precise, makes more noise than is necessary.

It also feels somewhat disingenious, perhaps because it seems that
several people are going out of their way to find flaws in each
others' postings, just because the poster is who he is.

I perfectly understand the feeling of wanting to rub someone's nose in
every little error they might make, but personal vendettas are not
endearing, and they do lower the perceived credibility of whoever
pursues them.

/L 'never argue with idiots, they merely bring you down to their
level, then beat you with experience'