Page with embedded scaffold refreshes improperly following CRUD if Javascript disabled

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Scenario: AS scaffold embedded on a page with other HTML elements
2. User has disabled Javascript in browser
3. During and after a CRUD action using standard AS controls, the rest of the 
page is absent.

What is the expected output? What do you see instead?
Expected: properly formed page with everything on it
Actual: page contains AS table plus page header, footer, navbar, etc., but no 
other HTML content.

NOTE: cannot reproduce this bug in demo.activescaffold.com

What version (or revision) of the product are you using?
1.0.1

If this bug causes an exception, please paste at least the first 20 lines 
below. N/A

Original issue reported on code.google.com by goo...@jamesgagne.com on 31 May 2007 at 4:33

[GoogleCodeExporter]

ack. yeah, i know what's happening here. with javascript turned off, there's no 
ajax,
which means that activescaffold busts out of any embedding when you try and 
click
*anything*. no clue how to resolve this.

Original comment by cainl...@gmail.com on 31 May 2007 at 6:08

• Changed state: Accepted

[GoogleCodeExporter]

Perhaps use <noscript> to  show a warning that javascript should be enabled for
proper use ?
http://www.w3schools.com/tags/tag_noscript.asp

Original comment by tom.bi...@gmail.com on 26 Jun 2007 at 9:14

[GoogleCodeExporter]

Yeah, we'd probably be best off just not showing the embedded scaffold at all. I
can't think what value an embedded scaffold could have with javascript turned 
off.

Original comment by cainl...@gmail.com on 26 Jun 2007 at 10:57

• Added labels: Milestone-1.1

[GoogleCodeExporter]

We should probably send the user to a page with only the embedded scaffold on 
it. The
short-term fix might be to just disabled all nested scaffold links in the 
absence of JS.

Original comment by rrwh...@gmail.com on 8 Jul 2007 at 5:55

[GoogleCodeExporter]

Isn’t there a standard HTML call that will tell you the URL of the referring 
page? It MAY require either Javascript or 
PHP, but if I’m not mistaken, the PHP only needs to be on the server side. 
Then when you’re done, zip back 
whence you came.

A less satisfying solution is to let AS users know of this behavior, and why 
demo.activescaffold.com works: stay 
on the main controller page the whole time. 

Original comment by goo...@jamesgagne.com on 8 Jul 2007 at 7:39

[GoogleCodeExporter]

Rich: Disabling nested scaffold links doesn't cover it, since people're using
embedding in other situations. And currently the behavior _is_ to go to a page 
with
only the embedded scaffold, which is part of the problem (the other part is 
getting
back).

James: The problem with HTTP_REFERRER is knowing when to care about it.

Tom: Yeah, a <noscript> warning might be the solution. Sigh.

Original comment by cainl...@gmail.com on 26 Sep 2007 at 4:19

[GoogleCodeExporter]

added a <noscript> warning
r644

Original comment by cainl...@gmail.com on 26 Sep 2007 at 4:36

[GoogleCodeExporter]

Original comment by cainl...@gmail.com on 26 Sep 2007 at 4:37

• Changed state: Fixed

[GoogleCodeExporter]

In Safari for Mac this is causing the warning to always display - even when 
Javascript is enabled. I know "Who 
uses a Mac?".

Original comment by edwin.m...@gmail.com on 26 Sep 2007 at 8:03

[GoogleCodeExporter]

Yeah, what is this "mac" you speak of?

Seriously though? <noscript> tags displaying anyways? That's really odd. What 
if you
remove teh active-scaffold class from the tag?

Original comment by cainl...@gmail.com on 26 Sep 2007 at 8:06

[GoogleCodeExporter]

k, tried that and it is still there.

Original comment by edwin.m...@gmail.com on 26 Sep 2007 at 9:19

[GoogleCodeExporter]

findings: safari will always display the <noscript> if it's not paired with a
<script> tag. but prototype always removes (and evaluates) <script> tags from 
ajax
calls. therefore, my patch fails.

while safari's behavior is annoying, i can hardly fault <noscript> tags for not
working during an ajax call. :) unfortunately i'm not sure where to generate the
<noscript> tags without affecting ajax calls as well.

Original comment by cainl...@gmail.com on 26 Sep 2007 at 9:53

• Changed state: Accepted

[GoogleCodeExporter]

Original comment by cainl...@gmail.com on 3 Dec 2007 at 11:24

• Removed labels: Milestone-1.1