Archive for August, 2007

Litmus – Web Page and Email Testing Platform

Today I heard about the impending launch of the new incarnation of the SiteVista web page and email testing tool ‘Litmus‘. A quick email later and I got myself a shiny beta invite and took it for a spin (thanks Paul).

There are some screen shots up on the SiteVista blog which give a fair impression of the kind of thing we are talking about here. Some of the features that might not be so obvious from the screen shots are such things as test versioning, public access/collaboration control, the ability to save browser combinations, multi-user collaboration, test histories and I’m sure a whole load of other stuff that I am yet to discover in my limited noodling…

Having tried a couple of the other tools of this ilk (Browsercam, Browsershots, Browsrcamp) I would say that Litmus will definitely ‘raise the bar’ in terms of features, UI and performance.

So… keep your eye on the SiteVista blog and expect some action any day now…

Disclosure: I have worked with the gentlemen of Salted but I am a man of morals and wouldn’t post something here that I didn’t mean.

E-TextEditor V1.0 Released

E-TextEditor, the Windows text-editor inspired by and based upon the bundles for TextMate has come out of beta.

If you haven’t tried it you should give it a whirl. It is definitely the closest us, aesthetically challenged, yesterday’s news Windows users are going to get to the lofty heights of TextMate.

UPDATE: I notice the Powershell tag is pulling a fair few people into this post from Technorati but only a small proportion of visitors are actually clicking through and checking ‘E’ out. If you are searching for a Powershell editor dear reader give it a look. I think this could be the Powershell editor that people have been waiting for. Whilst there isn’t, to my knowledge at least, a Powershell bundle, there will certainly be a demand for one as this editor takes off. As the bundles are coming currently from the Mac world there obviously has been nobody interested in writing Powershell stuff on the Mac. With ‘E’ this will change and the ‘bundle’ set up is very extensible/flexible. Give it a whirl…

Beware of GMail not enforcing an SSL connection…

I’ve just been reading this: TG Daily – Point and click Gmail hacking at Black Hat (thanks for passing that on Al)

Basically, if you aren’t logging into GMail using SSL then someone can grab your cookie and replay it. Well, that is pretty obvious if you are familiar with how all that kind of thing works and of course, I only log into GMail using SSL in fact, Google enforces this. Oh… actually, now I look it doesn’t enforce it at all…

If you enter the url mail.google.com it automatically redirects you to https://mail.google.com and all is well. But if you log in and then close your browser, re-open it and enter it again it takes you straight to http://mail.google.com presumably passing the unencrypted cookie along the way.

Opening up your e-mail is not good, especially when you consider the alarming wealth of sites that still send password reminders as plain text which are all sitting in your mail archive along with those that send your full credit card number when you get an order confirmation.

GMail works perfectly well if you add the all important little ‘s’ into any of its URLs so why don’t they just enforce it and save us the bother?

Fix Jerky Scriptaculous (Script.aculo.us) Animation Effects

I was adding some BlindDown, BlindUp animation effects to something I was working on today with Scriptaculous and had a bit of an issue with jerky animation. I’ve had this before and ended up ditching them but never really investigated why sometimes they are smooth as nobody’s business and others not.

As it happened, today, I had it working nicely, made a change, and the issue arose, thus leading me to the cause.

It turns out that if you add a BlindDown or BlindUp animation to a div that has padding or a margin set in its CSS as the effect begins to execute the div jumps to the padding/margin size and then the BlindUp/BlindDown (or probably others but I haven’t investigated) works on the remaining ‘area’ of the element.

So this is jerky:

<script language="javascript" type="text/javascript">
    Effect.BlindDown("divJerky");
</script>
<div style="height:50px; padding:10px; display:none" id="divJerky">
    
</div>

The solution is this:

<script language="javascript" type="text/javascript">
    Effect.BlindDown("divJerky");
</script>
<div style="display:none" id="divJerky">
    <div style="height:50px; padding:10px">
    </div>
</div>

If you add the padding to an ‘inner’ div all is well in the world again and everything works smoothly…

I know this isn’t exactly astrophysics but it is Friday afternoon and I’ve only just realised what the issue was so I thought I’d share…

Have a good weekend :)

Clicking more than once may charge your credit card multiple times…

I bought some tickets from the Royal Liverpool Philharmonic Hall (that sounds like I’m all very cultured but, sorry to disappoint, they were for a pirate fancy dress party). At the checkout I was confronted with this:

buy_click_once

Jeysus… Come on. It’s the 21st century for heaven’s sake. Surely this problem has been dealt with before?

Well, yes it has. The common solution is to disable the ‘Buy’ button using JavaScript in its “onclick” handler (or change its text or some such thing). This is probably what I would have done but it has the problem that it falls over if JavaScript isn’t enabled and will incur the wrath of Uncle Dave… You could argue that those without JavaScript enabled probably number less than 10% and then only a tiny fraction of those are going to be trigger happy so it solves the problem more or less. To complement that you could provide a big red warning as above and hide it with JavaScript as the page loads. This would mean it is displayed to those without JavaScript and would be hidden from those with it. This probably fits in with the Unobtrusive JavaScript mantra.

A more meaty non JavaScript solution is to put some sort of form id into a hidden field in the form on the server side when it is first rendered and then store it when the order is first received. Then only process orders that have not already been received. This works but the problem that arises is that after the second, erroneous, click you have to deal with where the user is redirected to so you need to be able to retrieve the order using the hidden value to allow the next page, confirmation or whatever, to be properly served (I can’t use served in this context anymore without thinking of South Park). This requires a bit of thought as the order may not have been processed when the second click is ‘received’ so you may not yet have an order to work against.

So, has anyone else got any bright ideas or should those with an itchy trigger finger be left to suffer the duplicate order fate of their own making?



Follow

Get every new post delivered to your Inbox.

Join 37 other followers