I have little idea of what best practice is in the JavaScript world, but
I wonder if we could clean up the core JavaScript code a little, perhaps
even introduce some coding standards, etc.
I've been perusing "Prototype" - a JavaScript library that supports much
AJAX goodness, and really like w
I've been using the following local extension
Selenium.prototype.doFocus = function(locator) {
var element = this.page().findElement(locator);
triggerEvent(element, 'focus', false);
};
to programmatically fire a focus event on an element whose "onFocus"
instigates a server round-trip. I'
When I introduced the PatternMatcher stuff recently, I used the same
kind of "reflection" trick we use for locators, commands, etc;
PatternMatcher tried to map a prefix like "regexp" to a function-name
like "RegexpMatchStrategy".
I realised tonight that it's simpler to just build the registry
Patrick Lightbody wrote:
Has anyone given thought to a JavaScript/iframe combo (or FireFox
plugin) that captures the clicks/events/etc by a user and records them
as Selenium commands? It could start to get a little hairy (see example
below), but would be very useful.
Ian Bicking wrote a Test r
Darrell DeBoer wrote:
How about something like:
Selenium.commands.myCommand = ...
... etc ...
I think you're right - this mechanism is a bit clearer for cases where
we don't need subclasses to override behaviour. I'm not so sure how
we'd do it for locators where we have different implem
David Kemp wrote:
Are you also suggesting that we copy their code for
Class.create() and Object.extend() ?
Class.create(), anyway. And I like the more structured class-definition
style, e.g.
MyClass.prototype = {
method1: function(arg1, arg2) { ... },
method2: function() { ... }
}
Hi all,
Our JIRA server is currently unavailable... in case anyone has been
trying to reach it. We hope to have it back online tomorrow.
--
cheers, MikeWhttp://www.dogbiscuit.org/mdub/
___
Selenium-devel mailing list
Sel
On Tue, July 5, 2005 19:15, Mike Williams said:
> Our JIRA server is currently unavailable... in case anyone has been
> trying to reach it. We hope to have it back online tomorrow.
... and ... now it's back.
--
cheers, Mike
___
Sel
Darrell DeBoer wrote:
> I think you're right - this mechanism is a bit clearer for cases
> where we don't need subclasses to override behaviour. I'm not so
> sure how we'd do it for locators where we have different
> implementations for different BrowserBot subclasses - thoughts?
A very good
Darrell DeBoer wrote:
I'm not sure what you mean. ...
Maybe I'm missing something, though?
I didn't really explain it very well.
Let's use the locators functions as an example. I'm figuring that they
still get invoked in an OO-way, with "this" referring to a pageBot
instance. Currently
Ed Chapel wrote:
I have locally modified the SeleneseRunner to have a button to turn
the log on and off.
Great idea. Thanks, Ed.
--
cheers, MikeWhttp://www.dogbiscuit.org/mdub/
___
Selenium-devel mailing list
Selenium
Ochsner, Andrew wrote:
A lot of people (me included) have noticed that IE will go into an
infinite loop especially after a |type| command. After digging through
this for hours, I found the problem, and the solution.
The issue (for me) is around IE caching requests. I do an |open|
command fol
aslak hellesoy wrote:
While looking at the Greets of the otherwise brilliant
http://www.ripcord.co.nz/behaviour/ I got curious about the
getElementBySelector function by Simon Williamson. Google for it.
http://simon.incutio.com/js/getElementsBySelector.js
Nifty. So, it would allow somethi
In recent days I re-styled the logging console such that it floats above
the AUT frame in a absolutely-positioned DIV. It works quite nicely,
as you can see the log messages scrolling as the test runs.
Unfortunately, though, it doesn't work on Konqueror, where a bug seems
to cause the loggin
neko bsd wrote:
> I want to get selenium of FreeBSD ports. So, I make it.
Great!
> I will put it on /usr/ports/security/selenium. Or,
> /usr/ports/www/selenium. Which do you like better, www and security ?
I don't know anything about how FreeBSD structures it's packages, but
"www" sounds more a
David Kemp wrote:
I have just committed code that automatically generates asserts,
assertNots, verify, and verifyNots from getters.
Nice one, David!
Besides the obvious removal of duplication, I think this is a good step
to allowing "driven" Selenium to be more useful, since accessors like
On Tue, August 16, 2005 23:51, Alexander Dvoretskiy said:
> I had not found in selenium functionality for testing prompts
> (window.prompt) and added it by myself.
Wonderful. Thanks for the contribution, Alex!
If you have a moment, could you create an enhancement issue at
http://jira.public.
> If you have a moment, could you create an enhancement issue at
>
> http://jira.public.thoughtworks.org/browse/SEL
>
> and attach your patch there?
Sure. I have done this just now. But I didn't find how to attach
patch there
On the LHS of the issue page, under "Operations", you should see an
Harman, Elaine wrote:
Looks like there's a bug in PageBot.locateElementByName, when
retrieving a form array element. e.g.: Option #1 Option #2
This is not exactly a bug - more a limitation of the design. At the
moment, element locators are expected to return only a single element.
This work
David Kemp wrote:
we could leave the default behaviour as is, but support both "trim"
and "untrimmed" for those wanting to be explicit about it.
assertValue | foo | untrimmed:tolower:regexp:^fo*$
I'd been thinking that stacked prefixes would be similar to nested
functions. From that point o
Yes, it seems sensible to make the locator-prefix matching
case-insensitive. Patch applied: thanks.
--
cheers, MikeWhttp://www.dogbiscuit.org/mdub/
___
Selenium-devel mailing list
Selenium-devel@lists.public.thoughtworks.or
Harman, Elaine wrote:
Apologies in advance for the long, dull novel.
No problem. Thanks for taking the time.
At present, everything that uses an element-locator expects the return
value to be, well, an element. I don't think we wanna have to check in
each case whether it is in fact an ele
On Thu, August 25, 2005 9:11, Jamo Smith said:
> HOWDY!
>
> I wrote my own assertElementOrder a few months back and today I finally
> updated to the newest version 0.5 of selenium (from 0.3).
>
> So I moved my custom assertion into the user-extensions.js as instructed.
> However it no longer works
On Thu, August 25, 2005 18:03, Alexander Dvoretskiy said:
> Can anyone suppose why next code fails (selenium-browserbot.js lines
> 168-171):
Fails in what context? Which browser? Is there an error message? Could
"windowToModify" be null, or not a window object?
--
cheers, Mike
[EMAIL PROTECTED] wrote:
I don't see any information about the plan for the next release of
Selenium.
There isn't one - a plan, that is. But thanks for the question, because
it's probably about time we did a release - there's nothing left on the
TODO-list for 0.6, right now.
Darrell DeBo
David Kemp wrote:
I have just committed an extension to the command handler framework so
that assert, assertNot, verify, and verifyNot are all generated
automatically from no-arg getters. Previously it only worked with
single-arg getters.
As a result of this, I have now rewritten assertTitl
Mike Williams wrote:
Can I hear from the other committers: all in favour of a release?
"Do I contradict myself?
Very well then I contradict myself ..."
David, I just re-read your message, and in particular, the "Still to go"
section. Would it be best to hold off the r
Is anyone else out there running Selenium in Konqueror?
I'm seeing some instability problems, ie. occasional crashes of the
browser after running Selenium. It's nothing I can pin down, but often
if I run the full suite of tests, Konq dies some minutes later.
Has anyone seen this kind of beha
On Wed, August 31, 2005 8:48, Eric Starr said:
> I have been looking over the issues list considering the possibility
> of lending a hand.
Excellent!
> By looking at the list, I begin to think that the
> list is not up to date.
That's entirely possible.
> For one, there really is no Roadmap wi
David Kemp wrote:
Please, don't wait for me to finish the command autogeneration stuff.
What has been done so far all seems to work fine, and the rest can be
included in later releases.
Okay, great.
I think we should give people a chance to test what's on the SVN trunk.
I'm particularly k
Tres Seaver wrote:
Is there anything wrong with the patch I inlined into the description
Looks okay ... but not being a Python guy, I really wouldn't know. I'm
happy to apply it for you though.
(I couldn't figure out the jira UI to attach a file at the time)?
No; sorry about that. Unti
Unfortunately, recent testing has confirmed that Selenium has stability
issues on Konqueror and Safari. They're bad enough that I think it's
worth postponing the next release until we sort them out.
For those of you wishing to use the "bleeding edge" version of Selenium
in the meantime, it's
I've arranged for the Selenium Confluence pages at
http://confluence.public.thoughtworks.org/display/SEL/Home
to be editable by any (registered, authenticated) user. I hope that
this will provide a place for tips and tricks, contributed
user-extensions, and the like.
--
cheers, MikeW
I believe I've fixed the problems with Konqueror that caused me to
postpone the 0.6 release.
I've installed the latest version of Selenium on my personal web-site
for "remote"[1] testing.
http://dogbiscuit.org/mdub/software/selenium/current/TestRunner.html
If you have a chance, please tak
Aahz wrote:
Well, in OS X 10.4.2, it gets to the end of test 17, then it bombs
(Safari dies).
Thanks Aahz. That's not good news.
"Test 17"? Are we talking TestXPathLocators??
--
cheers, MikeWhttp://www.dogbiscuit.org/mdub/
David Kemp wrote:
Strange that TestWaitInPopupWindow still hangs in firefox 1.0.6 when I
point it at your site.
I've certainly seen that problem in Firefox on Windows, though not on
Linux. I can't reproduce it now.
I wonder if you've got an old version of Selenium cached? Try a
Shift-Ref
On Mon, September 19, 2005 6:18, Michael Dunstan said:
> On 9/19/05, Michael Dunstan <[EMAIL PROTECTED]> wrote:
>> It seems to be the next test "TestGoBack" that crashes Safari if it is
>> either ran or walked. Though it works fine if stepped.
>
>> Also the visibility tests, "TestVisibility" and
On Mon, September 19, 2005 13:20, Aahz said:
> ... the Safari TestSuite works fine (35 tests run, 282 commands passed).
Excellent! Many thanks.
--
cheers, Mike
___
Selenium-devel mailing list
Selenium-devel@lists.public.thoughtworks.org
http://list
Robert S. Sfeir wrote:
Mike all tests pass.
Excellent; thanks Robert.
I did hit a snag in 2 tests but then realized that my popup blocker
was on, and that made the 2 test popup window and wait in popup
fail. I wonder if there could be a workaround for this. The popup
window will gene
Alex Dvoretskiy wrote:
Is it possible to force selenium use separate window for AUT?
Not at the moment, but it's certainly a frequently-requested feature.
How difficult to add such a functionality?
Not sure: no-one's tried to implement it yet :-)
--
cheers, MikeW
So, the Konqueror/Safari issues have been solved.
There's still a weird problem in Firefox (SEL-152), but it's somewhat
obscure, and only seems to kick in when running Selenium against a
remote server. So, I'd prefer not to let it delay the release.
I propose to cut a release this weekend.
Selenium-0.6.0 has been released, and is available for download from
http://selenium.thoughtworks.com/download.html
Notable features of this release include:
pattern-matching: Support for regular-expression and exact
matching. Allow patterns to be used to match alert and confirmation
On Tue, September 27, 2005 4:34, Luke Closs said:
> I actually got it working with the help of user-extensions.js. I
> created an inputKeys command that calls the textarea onkeypress
> callback for every key. If the callback returns true, it'll also add
> it to the textbox.
Nice work, Luke.
-
neko bsd wrote:
> I had to want to log function. It is supported Release 0.6.0.
>
> But, of which test log I am is not understand at all test.
>
> Executing: |open | http://hoge.com/index.html | |
>
> I want to add Test Suite name and Test Case Name.
You want extra context-info in the log? Fair e
On Thu, September 29, 2005 0:34, Scott Fitchet said:
> One thing I have to been able to figure out yet is how to "check" or
> "uncheck" a box instead of "clicking".
Hi Scott. You're absolutely right; this is a glaring hole in the API at
the moment, and shouldn't be too hard to fix up. I've logg
On Thu, September 29, 2005 6:25, Warren Pollans said:
> I converted the the TestRunner tests that were included in the
> yesterday's svn version of selenium to run as ruby-driven tests.
Thanks Warren - this is great.
I must admit I'm completely ignorant of the "driven" stuff, and
unfortunately t
On Thu, September 29, 2005 11:46, Jason R Huggins said:
>> pending anyone more familiar with that code getting involved.
> That would be me... Sure. I'll take care of applying the patch.
Fantastic - thanks Jason. Welcome back!
--
cheers, Mike
___
S
Jason R Huggins wrote:
This is what I have:
61fa8ecab81cebd05bd3502f58bfd8cf
Mike Williams is the one who uploaded the original zip file to the
gforge server. We'll probably need to confirm with him that he gets the
same checksum value on the file he originally uploaded. Mike?
I c
Well, now that 0.6 is out, let's try to work out: what next?
There are already a few things on "the plan" for 0.7 - mostly things
I've added because I (personally) think are important, or would provide
a quick-win. See:
http://jira.public.thoughtworks.org/browse/SEL?report=com.atlassian.jira
We're yet to find a better term to describe the "TestRunner" mode of
operation, ie. the one that's not "driven" by an external script.
(Yeah, that old can of worms.)
It would be nice if it were another adjective, for symmetry with "Driven".
Do any of these stick?
- Simple mode
- Dumb mode
- Au
Stephen Haberman wrote:
So we have:
1) Document Literal Selenium
and
2) RPC Selenium
I like it.
[Not sure what happened to my previous reply].
I like it too. A lot. I think "Document" vs "RPC" is the best
distinction I've heard so far.
"RPC" communicates the concept that the "
On Mon, October 10, 2005 23:51, Aahz said:
> Another option might be "Server" mode. Regardless of who calls whom,
> the browser is definitely acting as a server. Also, the driver
> initiates the initial connection, right? So the browser is doing a
> callback on the driver. We could call it "Ca
On Fri, October 14, 2005 7:20, Grig Gheorghiu said:
> My "Tool Look: A Look at Selenium" article was published in the Oct.
> 2005 issue of Better Software.
Congratulations, Grig. Nice article, and thanks for your comments on the
Selenium community.
--
cheers, Mike
___
Greg Fodor wrote:
Please don't nuke the .NET Driver, we're using it extensively ...
Well, that answers the "is anyone using it" question :-) Thanks Greg.
I'm very happy to see the Java and .NET functionality stay, if it's
being used. Feel free to submit patches if you see the .NET driver
actually tend to remain fairly static, while user-visible
labels tend to change occasionally (upon the whims of the customer).
--
cheers,
Mike Williams
___
Selenium-devel mailing list
Selenium-devel@lists.public.thoughtworks.org
http://lists.public.thoughtworks.org/mailman/listinfo/selenium-devel
On 23/10/2005, "Keenan" <[EMAIL PROTECTED]> wrote:
>I'm noticing that when 'click' or 'clickAndWaiting' a url in the sub
>window [that will load a url into the subwindow] it seems to be
>double submitting it to the server. Is anyone else noticing this?
>
>When putting a break point in my jsp c
JamesWang wrote:
Today I spent two hours in the example "TestWaitInPopupWindow" of
selenium 0.6 and I confirmed it was wrong and should not exist.
The test script of TestWaitPopupWindow is wrong because command
"clickAndWait" doesn't work for popup.
Am I right?
Yes, you're right:
JamesWang wrote:
Today I spent two hours in the example "TestWaitInPopupWindow" of
selenium 0.6 and I confirmed it was wrong and should not exist.
The test script of TestWaitPopupWindow is wrong because command
"clickAndWait" doesn't work for popup.
Am I right?
Yes, you're right:
Quoting JC Hearn <[EMAIL PROTECTED]>:
We are trying to test an application which has a frameset which contains
some javascript (in the actual frameset file). That last fact (contains
javascript) seems to freak out selenium. What happens is the frames become
ridiculously sized (millions of pixe
[EMAIL PROTECTED] said:
It's cool to post frameset questions/problems/solutions even though Selenium
doesn't officially support them though, right? We have a heavily framed
site, so I for one am interested in frameset posts... plus I have some
experience adapting Selenium in that direction, and
Selenium's own self-test suite is in fact four separate suites: one each
for Firefox, IE, Konqueror and Safari. I find this duplication mildly
irritating, and certainly it's not easy to see (at a glance) what
Selenium functionality is unsupported on each platform.
I suspect that this would al
Luke Closs wrote:
What about a more general 'skip' feature. Test cases or individual
commands could be skipped based on some test. I see this in perl test
suites all the time, something like:
unless package foo is installed {
skip all tests
}
In the interests of keeping things simple
Robin Becker wrote:
I was able to use storeValue to obtain the value of a hidden filed
called period, but failed with store and javascript.
... this failed
store|javascript{document.getElementById('period').value}|PERIOD
I think that when your JavaScript expression is eval'ed, "document"
r
I recently added a couple of commands for dealing with toggle-buttons,
ie. checkboxes and radio-buttons.
check( locator )
uncheck( locator )
assertChecked( locator )
assertNotChecked( locator )
As you can see, they currently just accept a single element-locator
argument.
Now,
David Kemp wrote:
I like the idea of extending the locator.
Great.
Should you be able to capture the state of a single check box with:
| storeChecked | options cheese | myVariable |
Can you then use that variable later to set the checked status of
another check box:
| check | options onio
Aahz wrote:
> | check | options cheese | |
The main problem I see with this is that "options cheese" is a valid
name. Do we currently require that to be in quotes?
Hmmm. No, we neither support nor require quotes, right now.
I understand that it might be *technically* valid to have an elem
Luke Closs wrote:
On Fri, Nov 04, 2005 at 02:02:08PM -0800, Thomas Milne wrote:
Someone on my team managed to track this issue down. The patch to fix
our problem is ...
Should I:
- file a proper selenium bug and add the patch there
relax as someone else patches selenium with the above fi
Quoting JamesWang <[EMAIL PROTECTED]>:
Anybody can help me confirm this problem?
Should I propose it as a bug in JIRA?
Hi James.
Yes, sounds like a bug to me. By all means report it in JIRA. Thanks.
--
cheers, Mike
This mes
Dan Fabulich wrote:
Browsing through the lists, it doesn't look like anyone has gotten
Java Driven mode to work in some time. If so, I'd like to help fix it.
Fantastic. Welcome aboard!
For my part, I haven't really tried to get the Java driver working, and
don't know much about the driven
Bret Pettichord wrote:
We [...] have a problem with tests hanging on clickAndWait or
typeAndWait.
The problem is that our pages have javascript validation. These
actions only result in a post if the javascript validates. We have
tests that hang because the data fails the validation. The
c
Patrick Lightbody wrote:
Any thoughts on how we can handle step #6? The mailing lists for both
users and dev are ready to go:
Users: http://www.openqa.org/selenium/userForums.action
Developers: http://www.openqa.org/selenium/devForums.action
Hello, anybody still here? The wiki-page seem
Bret Pettichord wrote:
How to set myself up to view the logs?
Click the "Show Log" button to pop up the logging window.
I thought i would share the related code from Watir ...
It never assumes (like the Selenium code) that a new page is being
loaded. Rather it simply tests three states: br
Patrick Lightbody wrote:
I agree as well. Can someone who maintains the lists send me the
subscribers list for both dev and users? I can take care of the rest.
Hey Patrick.
Actually, any subscriber can view subscribers, via the mailman info page
for each list. Unfortunately though, excludes
Everything works fine on Firefox. But I still wonder if the blur should
always be fired. Maybe a check could be made to see if the select box still
has focus? I couldn't find a way to write this check myself.
Sounds like a good idea, but I'm not sure how to implement it, either.
Does anyone
74 matches
Mail list logo
