December 2008 Archives
I've been battling (on and off) with an extJs combobox not being wide enough for the past few days and finally google has come up trumps for me. The hint about setting "hideMode" to "offsets" seems to solve the problem.
It does seem to cause a new rendering oddity though, the combobox suffering the issue is contained inside of an anchor inside of a form layout inside of the fifth of a set of tab panels. Multiple comboboxes on the first tab (which incidentally is default visible) don't seem to exhibit the problem. The tab the combo is on contains the combo, and a grid with two associated buttons:
<tab>
When initially rendered, the label for the combo doesn't appear, nor do the buttons for the grid, they in fact do't render until the combo's drop-down is triggered. There's also a noticable lag when the tab is selected between the drop-down button part of the combobox appearing (show straight away) and the main text-box part showing (shows after a second). Grrrr!
It does seem to cause a new rendering oddity though, the combobox suffering the issue is contained inside of an anchor inside of a form layout inside of the fifth of a set of tab panels. Multiple comboboxes on the first tab (which incidentally is default visible) don't seem to exhibit the problem. The tab the combo is on contains the combo, and a grid with two associated buttons:
<tab>
<form>
<anchor>
<combo />
</anchor>
<anchor>
<label />
<grid>
<columnModel />
<buttons>
<button1 />
<button2 />
<buttons>
<grid>
</anchor>
</form>
</tab>
When initially rendered, the label for the combo doesn't appear, nor do the buttons for the grid, they in fact do't render until the combo's drop-down is triggered. There's also a noticable lag when the tab is selected between the drop-down button part of the combobox appearing (show straight away) and the main text-box part showing (shows after a second). Grrrr!
Previously: http://www.robertwray.co.uk/blog/2008/12/extjs-getclientrects-error.html
After encountering the getClientRects() issue again I downgraded to IE7 to eliminate a beta browser as the causal factor. It still happened. Grr!
Having found a thread about it on the extJs forums, it looks like it's caused by a bit of craptitude in the extJs version that's included in Coolite v0.6. I want 0.7!! (At least there's a work-around/solution in the linked forum posting for those not using Coolite).
After encountering the getClientRects() issue again I downgraded to IE7 to eliminate a beta browser as the causal factor. It still happened. Grr!
Having found a thread about it on the extJs forums, it looks like it's caused by a bit of craptitude in the extJs version that's included in Coolite v0.6. I want 0.7!! (At least there's a work-around/solution in the linked forum posting for those not using Coolite).
Inspired by: http://www.codinghorror.com/blog/archives/000869.html
Never has a truer word been spoken. Trying to shoehorn desktop application UI paradigms into a web based application never ends well. Just look at 'A qWikiOffice Desktop' for a shining example of all that is wrong with doing this. When loaded on a Mac, it's going to look horribly, horribly out of place thanks to using pseudo-Windows theming. When loaded on an XP box (or even a Vista box running a 'classic' theme) it's going to look horribly out of place. More importantly, the question is what does your application gain from looking like this?
The answer to the question above will almost certainly always be 'it looks good' or 'it's got lots of features'. A web application can look good without apeing that of a desktop application. It can look better.
Whilst I'm mentioning this, I have to say that my most loathed example of the UI Uncanny Valley has to be itunes. It uses the Mac look and feel on Windows and thus doesn't fit in at all. Bleurgh.
Never has a truer word been spoken. Trying to shoehorn desktop application UI paradigms into a web based application never ends well. Just look at 'A qWikiOffice Desktop' for a shining example of all that is wrong with doing this. When loaded on a Mac, it's going to look horribly, horribly out of place thanks to using pseudo-Windows theming. When loaded on an XP box (or even a Vista box running a 'classic' theme) it's going to look horribly out of place. More importantly, the question is what does your application gain from looking like this?
The answer to the question above will almost certainly always be 'it looks good' or 'it's got lots of features'. A web application can look good without apeing that of a desktop application. It can look better.
Whilst I'm mentioning this, I have to say that my most loathed example of the UI Uncanny Valley has to be itunes. It uses the Mac look and feel on Windows and thus doesn't fit in at all. Bleurgh.
If you encounter the error message "Failed to access IIS metabase" when making a request to a newly created ASP.net virtual directory under IIS, or even one that was created a whle ago on a machine that has had Visual Studio added/removed/upgraded/repaired, the solution is generally to go to WINDOWS_DIRECTORY\Microsoft.Net\Framework\FRAMEWORK_VERSION in a command prompt and run "aspnet_regiis -i"
Discovered at: http://cgaskell.wordpress.com/2006/11/22/failed-to-access-iis-metabase/
I was looking through my web stats and saw that my most-read posts were:
- "Insufficient System Resources Exist to Complete the API."
- Lamb Shanks with Cannellini beans and Balsamic Vinegar
- Using Classic ASP on Vista/Longhorn/IIS7
(Actually in my stats for the URL: http://robertwray.co.uk/blog/classic-asp/2007/05/ ) - Sausage Casserole
- ASP.net: Viewstate
I've been encountering a strange error with extJs (when using Coolite) that I think may be IE8 specific, but I haven't been able to prove it yet.
When I use the <ext:Grid> and set the ClicksToEdit to "1", rather than the default of 2, if I click through the editable fields in the grid too quickly I receive an "unspecified error" on a call to getClientRects. Very odd. Also, the editable textboxes for the bottom-most (of the two rows in the grid) row won't render/become editable unless the grid has been sorted (at least once). Even more odd.
When I use the <ext:Grid> and set the ClicksToEdit to "1", rather than the default of 2, if I click through the editable fields in the grid too quickly I receive an "unspecified error" on a call to getClientRects. Very odd. Also, the editable textboxes for the bottom-most (of the two rows in the grid) row won't render/become editable unless the grid has been sorted (at least once). Even more odd.
I needed a VType for validating a time using the extJs framework in the format HH:MM (24 hour clock, no ‘AM’ or ‘PM’) and couldn’t find one that quite fit the bill, so a tweaked version is:
Ext.form.VTypes[“timeVal”] = /^([0-9]|1[0-9]|2[0-3]):([0-5][0-9])$/i;Related Linkage:
Ext.form.VTypes[“timeMask”] = /[\d\s:amp]/i;
Ext.form.VTypes[“timeText”] = ‘Not a valid time. Must be in the format hh:mm”.’;
Ext.form.VTypes[“time”] = function(v)
{
return Ext.form.VTypes[“timeVal”].test(v);
}
- A decent collection of VTypes (one of which forms the basis of the one above)
- Advanced Data Validation using ExtJs (a good overview of VTypes)
- Displaying Form Submit Errors by Saki (Mixing it up with VTypes and server-side validation)
Whilst hunting for a way to do this I found this sample/example at agrinei.com. [via: Joteke's aspadvice.com blog]
For the avoidance of doubt (and hopefully so it'll make anyone else who wonders about this have an easier time searching), the insertRecord method on the Coolite implementation of the extJs grid is added by Coolite.
Be prepared to see more like this :-S
Be prepared to see more like this :-S
If you look through comments on the IE Blog, you'll usually find at least one comment against each recent entry from someone saying something along the lines of "IE should pass Acid3". Given that (and I quote Wikipedia here) " It includes several elements from W3C CSS3 working drafts that have not made it to candidate recommendations yet", why on earth would any browser want to "pass" it?
Yes - the intentions of the Acid3 test are laudable, but c'mon people, passing a test that requires implementation of standards that are working drafts is a bit mental. (Before anyone comments - my opinion applies equally to *any* browser)
Yes - the intentions of the Acid3 test are laudable, but c'mon people, passing a test that requires implementation of standards that are working drafts is a bit mental. (Before anyone comments - my opinion applies equally to *any* browser)
As of today, the fifth result in a google search for Robert Wray (restricted to UK sites) is a page on hotfrog.co.uk. Quite how they've come to the conclusion that I offer to sell things ranging from Coffee Grinders to Laptops to Red Wine, I have no idea. Oh wait, I do. I've mentioned all those things, plus the others mentioned in various different blog postings. Automated page scrapingis all well and good, just not so much when it's done so poorly.
I can heartily recommend the Coolite framework for this, rather than battling with Javascript syntax. It's currently at v0.6 (c'mon, hurry up and release v0.7!) and thus is missing some functionality that I'd really quite like to use but it is pretty damn good. The big win for me is the fact that you get to describe your extJs "stuff" using markup that's validated by Visual Studio for you. Big. Win.
Follow coolite on twitter. (That's how I'm watching for v0.7!)
Follow coolite on twitter. (That's how I'm watching for v0.7!)
