ReCaptcha challenge API rendering method may cause "operation aborted" error in IE6-7 based browsers

[Arech]

I didn't find a ReCaptcha's bug tracker and have to write here - I
guess someone from ReCaptcha team could comment it?

Please, navigate to Zend Framework bug tracker http://framework.zend.com/issues/browse/ZF-6891
for bug description.

Thank you.

[Charles Sweeney]

Got a URL for the form in question (I read your bug report)?

--
Charles Sweeney
http://FormToEmail.com
PHP mail script with reCAPTCHA

[Arech]

I've already uploaded an AJAX-based version of ReCaptcha and the bug
doesn't trigger anymore. Sorry, I can't leave it unfixed.

On 3 июн, 17:28, Charles Sweeney <cont...@identipic.com> wrote:
> Arech wrote:
> > I didn't find a ReCaptcha's bug tracker and have to write here - I
> > guess someone from ReCaptcha team could comment it?
>

> > Please, navigate to Zend Framework bug trackerhttp://framework.zend.com/issues/browse/ZF-6891

> > for bug description.
>
> > Thank you.
>
> Got a URL for the form in question (I read your bug report)?
>
> --

> Charles Sweeneyhttp://FormToEmail.com

[Charles Sweeney]

Arech wrote:
> I've already uploaded an AJAX-based version of ReCaptcha and the bug
> doesn't trigger anymore. Sorry, I can't leave it unfixed.

Make a copy of your original form so people can check to replicate
your bug/error. Can you do that? Seems pretty pointless otherwise.

--
Charles Sweeney

[Arech]

Sorry for a long reply.
You are right, Charles, thanks. I had to think about it.

Please, look into http://framework.zend.com/issues/secure/attachment/11988/Zend_Recaptcha_IE_broken2.php
It is almost an exact copy of registration form, I mentioned in my
report. This code reproduces the error every time I run it.
(Requires Zend Framework installed to ../libs/Zend and Dojo installed
to /js/dojo folder of htdocs. Dojo can be downloaded from
http://dojotoolkit.org/downloads, and ZF from http://framework.zend.com/download/latest.
APP_ROOT constant in the test should point to a parent folder of
htdocs).

Clear IE6 cache (it may be not necessary for the first try, but IS
necessary for any subsequent tries) and navigate browser to this test
page. Page will start to load and in some moment it will display an
"operation aborted" error. To that point a "register" button usually
doesn't get instantiated.
If the test works fine, try to run it on a old-slow PC, like Athlon /
Pentium 1500Mhz.

What do you think?

On 5 июн, 00:14, Charles Sweeney <cont...@identipic.com> wrote:
> Arech wrote:
> > I've already uploaded an AJAX-based version of ReCaptcha and the bug
> > doesn't trigger anymore. Sorry, I can't leave it unfixed.
>
> Make a copy of your original form so people can check to replicate
> your bug/error.  Can you do that?  Seems pretty pointless otherwise.
>
> --

> Charles Sweeneyhttp://FormToEmail.com

[Charles Sweeney]

Arech wrote:
> Sorry for a long reply.
> You are right, Charles, thanks. I had to think about it.
>
> Please, look into http://framework.zend.com/issues/secure/attachment/11988/Zend_Recaptcha_IE_broken2.php
> It is almost an exact copy of registration form, I mentioned in my
> report. This code reproduces the error every time I run it.
> (Requires Zend Framework installed to ../libs/Zend and Dojo installed
> to /js/dojo folder of htdocs. Dojo can be downloaded from
> http://dojotoolkit.org/downloads, and ZF from http://framework.zend.com/download/latest.
> APP_ROOT constant in the test should point to a parent folder of
> htdocs).
>
> Clear IE6 cache (it may be not necessary for the first try, but IS
> necessary for any subsequent tries) and navigate browser to this test
> page. Page will start to load and in some moment it will display an
> "operation aborted" error. To that point a "register" button usually
> doesn't get instantiated.
> If the test works fine, try to run it on a old-slow PC, like Athlon /
> Pentium 1500Mhz.
>
> What do you think?

I tried your link but got prompted to download a file. Can't you
simply provide a link to the form in question????

I doubt very much if this bug, if that's what it is, will ever be a
problem for real visitors since one has to jump through many hoops to
even get a chance of replicating it!!

Best wishes.

--
Charles Sweeney

[i-imagine]

I hope this can help - reCAPTCHA has been good for me so far.

The exact same symptoms and problem occurred for me as follows.

- reCAPTCHA integrated into Coppermine Gallery installation - php
server side validation (links below)
- script added to a comment form submit
/////////////////////////////////////
<script>
var RecaptchaOptions = {
theme : 'white'
};
</script>
<script type="text/javascript"
src="http://api.recaptcha.net/challenge?k=<YOUR PUBLIC RECAPTCHA
KEY HERE>">
</script>

<noscript><iframe src="http://api.recaptcha.net/noscript?k=<YOUR
PUBLIC RECAPTCHA KEY HERE>"
height="300" width="500" frameborder="0"></iframe><br>
<textarea name="recaptcha_challenge_field" rows="3" cols="40">
</textarea>
<input type="hidden" name="recaptcha_response_field"
value="manual_challenge"></noscript>
////////////////////////////////////

All was fine and working well until I added a Coppermine plugin -
"Lightbox Slideshow"
This plugin is based on the popular Lightbox open source software.

Afterwards IE7 and only IE (cross checked only against FF and Opera)
would do the
exact same thing when it opened a page with the comment form and
reCAPTCHA.

IE seemed to get 'lost" - "operation aborted" was also one error
message that I saw.
A blank page was common, after a delay lasting more than a few
seconds.

- with a clear cache everything was fine until proceeding to a new
page with a reCAPTCHA
- with anything in the cache (from the same site and therefore the
Lightbox) it would do the above.

I too had to resolve it with the api version which seems to be running
flawlessly. This all happened
a couple of months ago and was thought to have happened because of my
lack of knowledge about
coding, etc. For certain, the IE "crash" with the above code and the
lightbox js was very repeatable.

Links:
reCAPTCHA api in comment/user reg. etc.
- www.i-imagine.net and
www.gallery.josephcarver.com/natural/

Details of how Coppermine Gallery Comments were modded with reCAPTCHA
(api solution there also)
- http://forum.coppermine-gallery.net/index.php/topic,58757.0.html


Again, I hope this helps.
JR Carver

On Jun 7, 8:50 am, Charles Sweeney <cont...@identipic.com> wrote:
> Arech wrote:
> > Sorry for a long reply.
> > You are right, Charles, thanks. I had to think about it.
>

> > Please, look intohttp://framework.zend.com/issues/secure/attachment/11988/Zend_Recaptc...

> > It is almost an exact copy of registration form, I mentioned in my
> > report. This code reproduces the error every time I run it.
> > (Requires Zend Framework installed to ../libs/Zend and Dojo installed
> > to /js/dojo folder of htdocs. Dojo can be downloaded from

> >http://dojotoolkit.org/downloads, and ZF fromhttp://framework.zend.com/download/latest.

> > APP_ROOT constant in the test should point to a parent folder of
> > htdocs).
>
> > Clear IE6 cache (it may be not necessary for the first try, but IS
> > necessary for any subsequent tries) and navigate browser to this test
> > page. Page will start to load and in some moment it will display an
> > "operation aborted" error. To that point a "register" button usually
> > doesn't get instantiated.
> > If the test works fine, try to run it on a old-slow PC, like Athlon /
> > Pentium 1500Mhz.
>
> > What do you think?
>
> I tried your link but got prompted to download a file.  Can't you
> simply provide a link to the form in question????
>
> I doubt very much if this bug, if that's what it is, will ever be a
> problem for real visitors since one has to jump through many hoops to
> even get a chance of replicating it!!
>
> Best wishes.
>
> --

> Charles Sweeneyhttp://FormToEmail.com

[Arech]

On 7 июн, 17:50, Charles Sweeney <cont...@identipic.com> wrote:

> I tried your link but got prompted to download a file.  Can't you
> simply provide a link to the form in question????

No, I can't do that.

That is amasing! Who do you think is interested in making ReCaptcha to
be a stable anti-spam solution? Me? Not exactly, I'm interested in my
own projects and can choose any other captcha solution to fit my
needs. I've solved the problem and spent a lot of my own time
describing it. I gave all the necessary info to reproduce it and gave
a link to Microsoft Knowledge Base article with complete description
why the ReCaptcha's Challenge API solution is invalid for IE6-7.
And what am I hearing? "Oooh, can't you upload the test case to a
hosting, I'm too ??? to make it myself". That is really strange for
me, sorry, if I'm too emotional right now.
I did a really good job for anyone, who really want to solve the
problem. Just take the file any try it yourself, if you are
interested.

> I doubt very much if this bug, if that's what it is, will ever be a
> problem for real visitors since one has to jump through many hoops to
> even get a chance of replicating it!!

Have you tried http://www.google.ru/search?q=recaptcha+%22operation+aborted%22
? There are tons of messages about this bug!


http://code.google.com/p/recaptcha/issues/detail?id=38 - The bug has
been reported 4 months ago.

[Charles Sweeney]

> www.gallery.josephcarver.com/natural/

The above form page has 281 errors and 29 warnings.

If any problem is browser-specific, it is invariably due to errors on
the form page.

--
Charles Sweeney

[i-imagine]

How right you are! Why don't you check the some of the "official"
reCAPTCHA pages for HTML errors,
why I just validated one of YOUR pages and found errors there too!


Does the "average" visitor really care about validation or is it what
is shown on their
browser? What choice would a webmaster make - a very nice ready-to-run
script that provides a pleasant user experience (Lightbox) or a
security add-on that
benefits the webmaster more than the users?

Perhaps I was not clear enough in the first message. The site is NOT
using the script above.
That script was discarded months ago. "This all happened a couple of
months ago" is what I wrote.
It is using the AJAX method instead of the in-line js. Did you even
look at the source of the page?
Did you even consider for a second that multiple reports of trouble
might be an actual issue?

No, I am not a by any stretch a software engineer or a developer, so
maybe you decided not to
actually pay any attention to the issue, preferring to spend your time
proving how wrong I am.
Good for you.

However, I have enough experience with product development to know
when someone would prefer to hear their own voice of absolute
righteousness rather than consider that their customers might be right
when calling attention to a problem. Yes, we are all customers of your
service. "Free' or not, paid or not - "it's your baby" - so get used
to it.

You must assume that all browsers operate the same. In the real world
they don't. You must also assume that there is no such thing as a
javascript conflict - try a google search for the term, there are at
least a couple of results to be found.

I was going to offer a test page, the direction of your inquiry and
your response means that would be a wasted effort since you already
know the answers.

In case you are actually interested in fixing the problem a test bed
server could easily be set up to provide the following conditions.
-Coppermine Photo Gallery
-reCAPTCHA added to comments form as noted above - all OK
-Lightbox plugin added to Coppermine - IE 7 not OK - other browsers OK

It was as repeatable as a light switch.

[Paul Herring]

Perhaps before insulting/having a go at volunteers who attempt to
freely give their time and advice on this list, you should perhaps
consider that if you want a timely, official, response from the people
who actually are in charge of of reCAPTCHA, you should perhaps pay for
a contract with them?

To re-iterate, no-one who has thus-far answered on this thread works
for reCAPTCHA - stop treating them as such.

And stop being so damn rude.

--
PJH

http://shabbleland.myminicity.com/com
http://www.chavgangs.com/register.php?referer=9375

[reCAPTCHA Support]

Hi everybody,

We're aware of this issue. So far, nobody has been able to isolate the reason that IE gives the operation aborted error. It seems that the best solution for sites which encounter this is to use the AJAX API.

- Ben

[Charles Sweeney]

i-imagine wrote:
> How right you are! Why don't you check the some of the "official"
> reCAPTCHA pages for HTML errors,
> why I just validated one of YOUR pages and found errors there too!

Sure about that?

> Does the "average" visitor really care about validation

No, but they would care if the site was broken.

> Perhaps I was not clear enough in the first message. The site is NOT
> using the script above.
> That script was discarded months ago. "This all happened a couple of
> months ago" is what I wrote.

Perfectly clear, but as it was the older form that was giving the
problem, that's the one I checked for errors. Maybe it's just a
coincidence that the form with 300 errors was the one giving the
problem.

> Did you even consider for a second that multiple reports of trouble
> might be an actual issue?

Indeed, but invariably with browser-specific problems, it's down to
validation errors on the form page (in my experience from my
customers).

> No, I am not a by any stretch a software engineer or a developer, so
> maybe you decided not to
> actually pay any attention to the issue, preferring to spend your time
> proving how wrong I am.
> Good for you.

I'm no expert either, but I do get a lot of feedback and see a lot of
problems and can sometimes see common causes.

> However, I have enough experience with product development to know
> when someone would prefer to hear their own voice of absolute
> righteousness rather than consider that their customers might be right
> when calling attention to a problem. Yes, we are all customers of your
> service. "Free' or not, paid or not - "it's your baby" - so get used
> to it.

It's not my service.

It wasn't so much my voice of absolute righteousness but more my voice
of pointing out that the form with the problem had 300 errors on it.

> You must assume that all browsers operate the same. In the real world
> they don't. You must also assume that there is no such thing as a
> javascript conflict - try a google search for the term, there are at
> least a couple of results to be found.

I would say that browsers are getting more fussy if anything, and
certainly if you declare a doctype, they will work to it, hence the
reason for problems where validation errors exist.

Sorry if I came across the wrong way.

[i-imagine]

Apologies for the harsh tone of the previous message. In no way is/was
it intended to be personal, more of a cry of frustration from someone
who was also providing free information to support your cause.

I still strongly support and will promote reCAPTCHA as a robust method
to stop spam. As previously mentioned reCAPTCHA has been great for me
so far.

Best Regards
JR Carver