[patch][cla] NumberTextBox does not properly support leading zeros in editOptions.pattern

[Terence Kent]

When a NumberTextBox? contains an editOptions.pattern which expects leading 0s, and the constraints.min is greater than 0, the _isValidSubset() method doesn't work properly.

For example, with editOptions.pattern = "000", and constraints = { min: 1, max: 500}, the user will get see error indicators while typing the first two keys when entering "004".

I've put together a jsfiddle showing both the issue and a patch for the issue here: ​http://jsfiddle.net/bYG37/4/

I'm currently using the aspect included to get around the problem.

[Terence Kent]

I've updated the patch we're using to include significant digit checking and fix a few bugs that we ran into. Unfortunately, I can't edit my issue description, so sorry for the spam.

I'll keep the most updated example here: ​http://jsfiddle.net/terencekent/SktJ8/

[dylan]

Hi @tkent, thanks for your explanation and proposed patch!

Do you have time to submit your patch as a pull request? And do you have a CLA on file? ​https://github.com/dojo/dojo/blob/master/CONTRIBUTING.md goes through some of the steps to make contributing as efficient as possible.

[Terence Kent]

Hey @Bill

I'd be happy submit my patch as a pull request. I'll do that tomorrow if I have time, or over the weekend if not.

I don't have a CLA on file, but I'll follow the process outlined here to get one: ​http://dojofoundation.org/about/cla

Thanks, Terence

[bill]

Cool, thanks. The other that needs to be done (hopefully by you) is to add an automated test case for this. We have a bunch of validation tests in dijit/tests/form/robot/ValidationTextBox.html, but apparently not for this case. There's also validationMessages.html. I forget why the tests are spread across two files.

[Terence Kent]

Bill,

I've sent a pull request with my patch and some unit tests for your review.

After thinking about the problem for a while, I changed the implementation of my solution. The new solution is functionally equivalent, but it's very different from the jsfiddle patch I originally included.

In case you have feedback on the patch, I've joined the contributors mailing list (though have been approved by the moderator yet)

[bill]

See ​https://github.com/dojo/dijit/pull/46.

Thanks @tkent. We usually add feedback to the PR (pull request), and I've done that.

[Bill Keese <bill@…>]

In 9094e1f8fd1202f28241f471cb16c13b7ca29798/dijit:

Error: Processor CommitTicketReference failed

Unsupported version control system "git": Can't find an appropriate component, maybe the corresponding plugin was not enabled?