webunitproj-users — Users of web unit.

[Webunitproj-users] Problem with iFrame

[Alex G. <ag...@ma...>]

Hello mates,

I have the problems with iFrames. At the moment when the page loading is finished the iFrames are empty. It starts to be populated by javascript a little bit later. The same problem was described in the earlier posts from 2002-11-15/16 but I couldn't find a solution.

Could you please help me.

Thanks in advance,
Alex

[Webunitproj-users] RE: Submiting form having image as submit button

[Jacqui G. <jgu...@pu...>]

I am also having problems submitting a form with the submit button as an
input image button:

=20

<INPUT onmouseup=3DbuttonClickOff(this); =
onmousedown=3DbuttonClick(this);
id=3DLogin onmouseleave=3DbuttonClickOff(this); style=3D"CURSOR: hand"
onclick=3Ddocument.loginForm.submit(); type=3Dimage alt=3DLogin
src=3D"images/buttons/login_btn.gif" name=3DLogin>

=20

Below is the code that I use to find it and click the button:

            WebLink submit =3D null;

            // set login form fields

            loginForm.getTextField(loginUserField,
WebTag.FIND_EQUALS).setValue(loginUserValue);

            // delay

Thread.sleep(1000);

            loginForm.getTextField(loginPswdField,
WebTag.FIND_EQUALS).setValue(loginPswdValue);

            // delay

            Thread.sleep(1000);

            try {

                submit =3D (WebLink)(loginForm.find("alt",
loginSubmitText, WebTag.FIND_EQUALS, WebLink.class)[0]);

                System.err.println("TEST: Found submit button
["+submit.toString()+"]");

             } catch (Exception e) {

                wc.close();

                fail("TEST: Did not find Login submit button image [link
text =3D "+loginSubmitText+"]");

            }

           =20

            submit.click();

            currentPage =3D wc.getCurrentState();

=20

The following error is thrown when I run my test:

=20

Error invoking method:close er:java.rmi.RemoteException: Error invoking
method:java.io.EOFException

      at java.io.DataInputStream.readInt(DataInputStream.java:397)

      at com.zeborg.iorpc.Util.readObject(Util.java:98)

      at
com.zeborg.iorpc.RemoteProcessClient.invokeMethod(RemoteProcessClient.ja
va:94)

      at com.zeborg.iorpc.Stub.inokeMethod(Stub.java:41)

      at
com.zeborg.webunit.iorpc.NativeWebClientStub.close(NativeWebClientStub.j
ava:69)

      at
com.zeborg.webunit.util.DefaultWebClient.close(DefaultWebClient.java:105
)

      at com.purisma.test.WebunitTest.testPCR(WebunitTest.java:248)

=20

E

Time: 7.5

There was 1 error:

1) testPCR(com.purisma.test.WebunitTest)java.rmi.RemoteException: Bad
Stuff happened

      at com.zeborg.iorpc.Stub.inokeMethod(Stub.java:46)

      at
com.zeborg.webunit.iorpc.NativeWebClientStub.close(NativeWebClientStub.j
ava:69)

      at
com.zeborg.webunit.util.DefaultWebClient.close(DefaultWebClient.java:105
)

      at com.purisma.test.WebunitTest.testPCR(WebunitTest.java:248)

      at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)

      at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.jav
a:39)

      at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessor
Impl.java:25)

      at junit.extensions.TestDecorator.basicRun(TestDecorator.java:22)

      at junit.extensions.TestSetup$1.protect(TestSetup.java:19)

      at junit.extensions.TestSetup.run(TestSetup.java:23)

      at com.purisma.test.WebunitTest.main(WebunitTest.java:257)

=20

FAILURES!!!

Tests run: 1,  Failures: 0,  Errors: 1

=20

Anyone have any ideas?


Thanks,

-jacqui

=20

[Webunitproj-users] Re: Replacement for ExplorerNativeWebClient

[Jacqui G. <jgu...@pu...>]

Brian,=20

=20

Can you post instructions on how to run tests using this newer version
of the ExplorerNativeWebClient?  I am unable to use Webunit because I am
on XP SP2, which no longer has the Microsoft VM installed.

=20

Thanks

Jacqueline Guerrero

=20

[Webunitproj-users] IOException during newWebClient()

[Jacqui G. <jgu...@pu...>]

I just downloaded Webunit and I the following error is thrown for
DefaultWebFactory.newWebClient():

=20

.args:wjview /vst -cp:p E:\tools\webunit\webunit-1.02\webunit-ms.jar
com.zeborg.webunit.ie.ExplorerIORPCServer 9000

Execing:wjview /vst -cp:p E:\tools\webunit\webunit-1.02\webunit-ms.jar
com.zeborg.webunit.ie.ExplorerIORPCServer 9000

toStrign(Th)

E

Time: 0.063

There was 1 error:

1)
testClickLink(com.purisma.test.WebunitTest)java.lang.RuntimeException:
Error getting ExplorerWebClient:java.io.IOException: CreateProcess:
wjview /vst -cp:p E:\tools\webunit\webunit-1.02\webunit-ms.jar
com.zeborg.webunit.ie.ExplorerIORPCServer 9000 error=3D2

=20

=20

I am running my test within Eclipse with the argument:
-Dzeborg.webunit.msvm.cp=3DE:\tools\webunit\webunit-1.02\webunit-ms.jar

=20

Does anyone have any idea what I'm doing wrong?  Please help.

Thanks,

Jacqueline

=20

[Webunitproj-users] IOException during newWebClient()

[Jacqui G. <jgu...@pu...>]

I just downloaded Webunit and I the following error is thrown for
DefaultWebFactory.newWebClient():

=20

.args:wjview /vst -cp:p E:\tools\webunit\webunit-1.02\webunit-ms.jar
com.zeborg.webunit.ie.ExplorerIORPCServer 9000

Execing:wjview /vst -cp:p E:\tools\webunit\webunit-1.02\webunit-ms.jar
com.zeborg.webunit.ie.ExplorerIORPCServer 9000

toStrign(Th)

E

Time: 0.063

There was 1 error:

1)
testClickLink(com.purisma.test.WebunitTest)java.lang.RuntimeException:
Error getting ExplorerWebClient:java.io.IOException: CreateProcess:
wjview /vst -cp:p E:\tools\webunit\webunit-1.02\webunit-ms.jar
com.zeborg.webunit.ie.ExplorerIORPCServer 9000 error=3D2

=20

=20

I am running my test within Eclipse with the argument:
-Dzeborg.webunit.msvm.cp=3DE:\tools\webunit\webunit-1.02\webunit-ms.jar

=20

Does anyone have any idea what I'm doing wrong?  Please help.

Thanks,

-jacqui

=20

[Webunitproj-users] Replacement for ExplorerNativeWebClient

[Brian K. <bt...@ne...>]

Since the Microsoft Java VM is no longer supported, I went ahead and
implemented my own NativeWebClient to control Internet Explorer.
Instead of going through the ms java runtime, it utilizes generated java
code (via njawin) to access IE COM interfaces (mshtml and shdocvw).  You
just need to place jawin.dll in your system path, and jawin.jar in your
classpath.  All the unit tests pass and it seems to run a little faster.
Would you all be interested in replacing ExplorerNativeWebClient with
this version?

Also here is a link to njawin if you're interested...
http://www.simtel.net/product.php?url_fb_product_page=3D60701

Thanks,

Brian

[Webunitproj-users] image submit button

[Ahmet T. S. <ah...@20...>]

    Hello,
    I know this has been asked before but I havent seen any answers posted for them.  So I wanted to give it a shot too.  How can I make it click the <input type"image" .....> tag?

I have tried this.

    	 SubmitButton but =(SubmitButton)currentPage.getRootTag().findFirst(
     	  new String[]{"type", "border", "height"},
          new String[]{"image", "0", "22"},
          new int[]{WebTag.FIND_EQUALS, WebTag.FIND_EQUALS, WebTag.FIND_EQUALS},
          SubmitButton.class);
          
          but.click();



and it is giving me a "notfound" error.  However if I change submitButton to WebTag, It finds the tag that I am trying to click, but this time it is telling me that I cant cast it that way.  thanks

[Webunitproj-users] Using Ant to run tests

[Robert P. <rp...@ne...>]

Currently, our QA folks are running their tests from within IntelliJ. I
would like to have those tests run without user intervention from an Ant
task. I figure that would be the junit task since that's how IntelliJ is
treating it.

=20

Does anybody have a sample junit task that I could take a peek at?

=20

Thanks,

=20

Robert

RE: [Webunitproj-users] random errors - please help!

[Brian K. <bt...@ne...>]

I figured out what is happening... there is a bug in the way an onchange
event is fired for a select tag.  Currently when you set an option on a
select drop-down, ExplorerNativeWebClient only has a handle to
WebOptionField, with this it tries to get it's parent (the select tag)
by calling el.getParentElement().  Unfortunately the parent is sometimes
null which is causing the null pointer exception on line 657.  Don't
know why it's randomly null though.

I was able to create a hacked fix locally, by firing the onchange event
directly on the select tag, rather than getting the option's parent.
The clean fix will be much more involved though, and I will probably
need to get some feedback on the best implementation approach in order
to create a patch.

Thanks,

Brian



-----Original Message-----
From: Doron [mailto:dg...@sh...]=20
Sent: Monday, July 21, 2003 8:54 AM
To: Brian Knorr
Cc: web...@li...
Subject: Re: [Webunitproj-users] random errors - please help!

Hi Brian,

	Sounds like there maybe a spurious timing issue (hard to solve
as it is
spurious). It would hard to debug it without actually being there. Is
there some javascript that changes the page while web unit is trying to
do something?

	It looks like the error is here:

com/zeborg/webunit/ie/ExplorerNativeWebClient.setOption
(ExplorerNativeWebClient.java:65

	You may want to look at that line of code and see what would
cause a
null pointer exception there.

			-Doron


=09

On Sun, 2003-07-20 at 16:49, Brian Knorr wrote:
> We are encountering some very strange problems with webunit... we can
> run a webunit script with no problems and then run it again an
> encounter errors ("Bad Stuff happened")... it's very random - we
cannot
> cause something to fail consistently.  Here is one of the stack
> tracing we get when these random errors occur... please help...
thanks.
>=20
> =20
>=20
> =20
>=20
> java.lang.NullPointerException
>=20
>             at com/zeborg/webunit/ie/ExplorerNativeWebClient.setOption
> (ExplorerNativeWebClient.java:657)
>=20
>             at java/lang/reflect/Method.invoke (Method.java)
>=20
>             at com/zeborg/iorpc/RemoteProcessServer.process
> (RemoteProcessServer.java:125)
>=20
>             at com/zeborg/iorpc/RemoteProcessServer.run
> (RemoteProcessServer.java:179)
>=20
>             at com/zeborg/webunit/ie/ExplorerIORPCServer.main
> (ExplorerIORPCServer.java:35)
>=20
> toStrign(Th)
>=20
> toString(Throwable)
>=20
> toString(Throwable)
>=20
> Error invoking method:setOption er:java.rmi.RemoteException: Error
> invoking method:java.rmi.RemoteException: Error on invoking:
>
com.zeborg.webunit.ie.ExplorerNativeWebClient.setOptionjava.lang.reflect
.InvocationTargetException
>=20
>             at java/lang/reflect/Method.invoke (Method.java)
>=20
>             at com/zeborg/iorpc/RemoteProcessServer.process
> (RemoteProcessServer.java:125)
>=20
>             at com/zeborg/iorpc/RemoteProcessServer.run
> (RemoteProcessServer.java:179)
>=20
>             at com/zeborg/webunit/ie/ExplorerIORPCServer.main
> (ExplorerIORPCServer.java:35)
>=20
> =20
>=20
> RootCause:java.lang.RuntimeException: Error in
> click:java.lang.NullPointerException
>=20
>             at com/zeborg/webunit/ie/ExplorerNativeWebClient.setOption
> (ExplorerNativeWebClient.java:657)
>=20
>             at java/lang/reflect/Method.invoke (Method.java)
>=20
>             at com/zeborg/iorpc/RemoteProcessServer.process
> (RemoteProcessServer.java:125)
>=20
>             at com/zeborg/iorpc/RemoteProcessServer.run
> (RemoteProcessServer.java:179)
>=20
>             at com/zeborg/webunit/ie/ExplorerIORPCServer.main
> (ExplorerIORPCServer.java:35)
>=20
> =20
>=20
>             at com/zeborg/webunit/ie/ExplorerNativeWebClient.setOption
> (ExplorerNativeWebClient.java:660)
>=20
>             at java/lang/reflect/Method.invoke (Method.java)
>=20
>             at com/zeborg/iorpc/RemoteProcessServer.process
> (RemoteProcessServer.java:125)
>=20
>             at com/zeborg/iorpc/RemoteProcessServer.run
> (RemoteProcessServer.java:179)
>=20
>             at com/zeborg/webunit/ie/ExplorerIORPCServer.main
> (ExplorerIORPCServer.java:35)
>=20
> =20
>=20
>             <<no stack trace available>>

Re: [Webunitproj-users] random errors - please help!

[Doron <dg...@sh...>]

Hi Brian,

	Sounds like there maybe a spurious timing issue (hard to solve as it is
spurious). It would hard to debug it without actually being there. Is
there some javascript that changes the page while web unit is trying to
do something?

	It looks like the error is here:

com/zeborg/webunit/ie/ExplorerNativeWebClient.setOption
(ExplorerNativeWebClient.java:65

	You may want to look at that line of code and see what would cause a
null pointer exception there.

			-Doron


	

On Sun, 2003-07-20 at 16:49, Brian Knorr wrote:
> We are encountering some very strange problems with webunit… we can
> run a webunit script with no problems and then run it again an
> encounter errors (“Bad Stuff happened”)… it’s very random – we cannot
> cause something to fail consistently.  Here is one of the stack
> tracing we get when these random errors occur… please help… thanks.
> 
>  
> 
>  
> 
> java.lang.NullPointerException
> 
>             at com/zeborg/webunit/ie/ExplorerNativeWebClient.setOption
> (ExplorerNativeWebClient.java:657)
> 
>             at java/lang/reflect/Method.invoke (Method.java)
> 
>             at com/zeborg/iorpc/RemoteProcessServer.process
> (RemoteProcessServer.java:125)
> 
>             at com/zeborg/iorpc/RemoteProcessServer.run
> (RemoteProcessServer.java:179)
> 
>             at com/zeborg/webunit/ie/ExplorerIORPCServer.main
> (ExplorerIORPCServer.java:35)
> 
> toStrign(Th)
> 
> toString(Throwable)
> 
> toString(Throwable)
> 
> Error invoking method:setOption er:java.rmi.RemoteException: Error
> invoking method:java.rmi.RemoteException: Error on invoking:
> com.zeborg.webunit.ie.ExplorerNativeWebClient.setOptionjava.lang.reflect.InvocationTargetException
> 
>             at java/lang/reflect/Method.invoke (Method.java)
> 
>             at com/zeborg/iorpc/RemoteProcessServer.process
> (RemoteProcessServer.java:125)
> 
>             at com/zeborg/iorpc/RemoteProcessServer.run
> (RemoteProcessServer.java:179)
> 
>             at com/zeborg/webunit/ie/ExplorerIORPCServer.main
> (ExplorerIORPCServer.java:35)
> 
>  
> 
> RootCause:java.lang.RuntimeException: Error in
> click:java.lang.NullPointerException
> 
>             at com/zeborg/webunit/ie/ExplorerNativeWebClient.setOption
> (ExplorerNativeWebClient.java:657)
> 
>             at java/lang/reflect/Method.invoke (Method.java)
> 
>             at com/zeborg/iorpc/RemoteProcessServer.process
> (RemoteProcessServer.java:125)
> 
>             at com/zeborg/iorpc/RemoteProcessServer.run
> (RemoteProcessServer.java:179)
> 
>             at com/zeborg/webunit/ie/ExplorerIORPCServer.main
> (ExplorerIORPCServer.java:35)
> 
>  
> 
>             at com/zeborg/webunit/ie/ExplorerNativeWebClient.setOption
> (ExplorerNativeWebClient.java:660)
> 
>             at java/lang/reflect/Method.invoke (Method.java)
> 
>             at com/zeborg/iorpc/RemoteProcessServer.process
> (RemoteProcessServer.java:125)
> 
>             at com/zeborg/iorpc/RemoteProcessServer.run
> (RemoteProcessServer.java:179)
> 
>             at com/zeborg/webunit/ie/ExplorerIORPCServer.main
> (ExplorerIORPCServer.java:35)
> 
>  
> 
>             <<no stack trace available>>

[Webunitproj-users] random errors - please help!

[Brian K. <bt...@ne...>]

We are encountering some very strange problems with webunit... we can
run a webunit script with no problems and then run it again an encounter
errors ("Bad Stuff happened")... it's very random - we cannot cause
something to fail consistently.  Here is one of the stack tracing we get
when these random errors occur... please help... thanks.

=20

=20

java.lang.NullPointerException

            at com/zeborg/webunit/ie/ExplorerNativeWebClient.setOption
(ExplorerNativeWebClient.java:657)

            at java/lang/reflect/Method.invoke (Method.java)

            at com/zeborg/iorpc/RemoteProcessServer.process
(RemoteProcessServer.java:125)

            at com/zeborg/iorpc/RemoteProcessServer.run
(RemoteProcessServer.java:179)

            at com/zeborg/webunit/ie/ExplorerIORPCServer.main
(ExplorerIORPCServer.java:35)

toStrign(Th)

toString(Throwable)

toString(Throwable)

Error invoking method:setOption er:java.rmi.RemoteException: Error
invoking method:java.rmi.RemoteException: Error on invoking:
com.zeborg.webunit.ie.ExplorerNativeWebClient.setOptionjava.lang.reflect
.InvocationTargetException

            at java/lang/reflect/Method.invoke (Method.java)

            at com/zeborg/iorpc/RemoteProcessServer.process
(RemoteProcessServer.java:125)

            at com/zeborg/iorpc/RemoteProcessServer.run
(RemoteProcessServer.java:179)

            at com/zeborg/webunit/ie/ExplorerIORPCServer.main
(ExplorerIORPCServer.java:35)

=20

RootCause:java.lang.RuntimeException: Error in
click:java.lang.NullPointerException

            at com/zeborg/webunit/ie/ExplorerNativeWebClient.setOption
(ExplorerNativeWebClient.java:657)

            at java/lang/reflect/Method.invoke (Method.java)

            at com/zeborg/iorpc/RemoteProcessServer.process
(RemoteProcessServer.java:125)

            at com/zeborg/iorpc/RemoteProcessServer.run
(RemoteProcessServer.java:179)

            at com/zeborg/webunit/ie/ExplorerIORPCServer.main
(ExplorerIORPCServer.java:35)

=20

            at com/zeborg/webunit/ie/ExplorerNativeWebClient.setOption
(ExplorerNativeWebClient.java:660)

            at java/lang/reflect/Method.invoke (Method.java)

            at com/zeborg/iorpc/RemoteProcessServer.process
(RemoteProcessServer.java:125)

            at com/zeborg/iorpc/RemoteProcessServer.run
(RemoteProcessServer.java:179)

            at com/zeborg/webunit/ie/ExplorerIORPCServer.main
(ExplorerIORPCServer.java:35)

=20

            <<no stack trace available>>

[Webunitproj-users] WebCheckField always returns false

[Brian K. <bt...@ne...>]

After getting a handle to a checkbox through the class WebCheckField,
regardless of whether or not the checkbox is checked on my page,
WebCheckField always returns false when I call the method getValue().
Please confirm this is a bug.  If so, I will prepare a patch.

=20

Thanks,

=20

Brian

[Webunitproj-users] testing 1-2-3

[Ostermueller, E. <Eri...@fn...>]

[Webunitproj-users] test

[Tidwell, R. <Rob...@fn...>]

test
=20
Robert Tidwell
Wintel Directory Services
Fidelity Information Services
A division of Fidelity National Financial
work-501-220-6229
rob...@fn...
=20

RE: [Webunitproj-users] major performance issue with large selectlist

[Brian K. <bt...@ne...>]

We have been considering the idea of refactoring our large select boxes =
to a search/select component for a while now, but currently the pages =
load just fine, even over a WAN.  I guess the webunit performance issue =
will help push us towards investing the time/resources to make the =
changes.  And I agree with you - in the long run moving away from super =
large select lists in the UI is a good thing to do.

Thanks for the info and advice...

Brian

-----Original Message-----
From: Doron [mailto:dg...@sh...]=20
Sent: Friday, May 23, 2003 7:54 AM
To: Brian Knorr
Cc: web...@li...; Jake Dempsey
Subject: Re: [Webunitproj-users] major performance issue with large =
selectlist

Hi Brian,

	I understand your issue. In general I would say that 8K choices in a
drop down is usually not a good UI thing to do because eventually it
wont scale and it will be slow going over a WAN. I would suggest to redo
this so you have a filter to reduce the list size.

On Thu, 2003-05-22 at 18:34, Brian Knorr wrote:
> In our application we have select boxes (drop-downs) with over 8000
> options.  The IE browser comes up instantaneously, but webunit is
> tremendously slow. =20
>=20
> =20
>=20
> At first we didn=FFt know what the performance issue was, but =
suspected
> it was due to the 8000+ options.  So we ended up creating a simple
> html page with a single select box containing 8500 options.  With this
> we tested webunit and found the following time statistics:
>=20
> =20
>=20
> 1) Retrieving the WebResponse with WebClient.getCurrentState()  took 9
> seconds
> ->> This requires IE to transfer all the HTML to webunit and then for
> webunit to parse the HTML to create its own DOM tree since getting the
> DOM tree directly from IE is way too slow.

> 2) Getting a handle to the WebSelectField with
> WebResponse.getRootTag.findFirst() took 18 seconds
->> Webunit does a linear search of the XML DOM tree generated in the
above step. It should still not take that long so its possible that
parsing takes place at this point. If you put a profiler on web unit,
you'll be able to tell.
>=20
> 3) Setting the value of the select box with
> WebSelectField.setSelectOption() took 3 seconds
->> The reason for this is that web unit tries to find the option tag in
the IE DOM. This DOM is slightly different then the one Web Unit has and
searching the IE DOM for the right element takes some time.
>=20
> =20
>=20
> As you can see it=FFs taking about 30 seconds to set a single select
> box.  Any ideas why this might be taking so long?  This is becoming a
> showstopper for us -- any ideas on when a fix could be produced?=20
> Please advise.
->> So the question is, can this be improved with IE? I don't see
anything obvious to fix in Web Unit. You could get through this by
reducing the list size in your select box and this is probably better
GUI as well.

>=20
> =20
>=20
> Thanks,
>=20
> =20
>=20
> Brian=20
>=20
> =20
>=20
> =20

Re: [Webunitproj-users] major performance issue with large select list

[Doron <dg...@sh...>]

Hi Brian,

	I understand your issue. In general I would say that 8K choices in a
drop down is usually not a good UI thing to do because eventually it
wont scale and it will be slow going over a WAN. I would suggest to redo
this so you have a filter to reduce the list size.

On Thu, 2003-05-22 at 18:34, Brian Knorr wrote:
> In our application we have select boxes (drop-downs) with over 8000
> options.  The IE browser comes up instantaneously, but webunit is
> tremendously slow.  
> 
>  
> 
> At first we didn’t know what the performance issue was, but suspected
> it was due to the 8000+ options.  So we ended up creating a simple
> html page with a single select box containing 8500 options.  With this
> we tested webunit and found the following time statistics:
> 
>  
> 
> 1) Retrieving the WebResponse with WebClient.getCurrentState()  took 9
> seconds
> ->> This requires IE to transfer all the HTML to webunit and then for
> webunit to parse the HTML to create its own DOM tree since getting the
> DOM tree directly from IE is way too slow.

> 2) Getting a handle to the WebSelectField with
> WebResponse.getRootTag.findFirst() took 18 seconds
->> Webunit does a linear search of the XML DOM tree generated in the
above step. It should still not take that long so its possible that
parsing takes place at this point. If you put a profiler on web unit,
you'll be able to tell.
> 
> 3) Setting the value of the select box with
> WebSelectField.setSelectOption() took 3 seconds
->> The reason for this is that web unit tries to find the option tag in
the IE DOM. This DOM is slightly different then the one Web Unit has and
searching the IE DOM for the right element takes some time.
> 
>  
> 
> As you can see it’s taking about 30 seconds to set a single select
> box.  Any ideas why this might be taking so long?  This is becoming a
> showstopper for us -- any ideas on when a fix could be produced? 
> Please advise.
->> So the question is, can this be improved with IE? I don't see
anything obvious to fix in Web Unit. You could get through this by
reducing the list size in your select box and this is probably better
GUI as well.

> 
>  
> 
> Thanks,
> 
>  
> 
> Brian 
> 
>  
> 
>  

[Webunitproj-users] major performance issue with large select list

[Brian K. <bt...@ne...>]

In our application we have select boxes (drop-downs) with over 8000
options.  The IE browser comes up instantaneously, but webunit is
tremendously slow. =20

=20

At first we didn't know what the performance issue was, but suspected it
was due to the 8000+ options.  So we ended up creating a simple html
page with a single select box containing 8500 options.  With this we
tested webunit and found the following time statistics:

=20

1) Retrieving the WebResponse with WebClient.getCurrentState()  took 9
seconds

2) Getting a handle to the WebSelectField with
WebResponse.getRootTag.findFirst() took 18 seconds

3) Setting the value of the select box with
WebSelectField.setSelectOption() took 3 seconds

=20

As you can see it's taking about 30 seconds to set a single select box.
Any ideas why this might be taking so long?  This is becoming a
showstopper for us -- any ideas on when a fix could be produced?  Please
advise.

=20

Thanks,

=20

Brian=20

=20

=20

Re: [Webunitproj-users] RE: radio tag

[Doron <dg...@sh...>]

Brian,

	You should have access to all the attributes of a tag. The INPUT tag
has a type="radio" (if I recall correctly). So that should allow you to
filter your results.


				-Doron

On Mon, 2003-05-12 at 10:32, Brian Knorr wrote:
> Just following up to see if anyone has any thoughts on the email
> below… thanks,
> 
>  
> 
> Brian
> 
>  
> 
> -----Original Message-----
> From: Brian Knorr 
> Sent: Wednesday, April 02, 2003 1:03 PM
> To: 'web...@li...'
> Subject: radio tag
> 
>  
> 
> Does anyone know how we could use the find method to get a handle on
> all the radio tag on a page?  Currently webunit models the input type
> radio as a WebTextField, so we end up getting radios and text fields. 
> We would to just get the radios.  If there isn’t a work around, maybe
> the best solution is for webunit to model radios with a class like
> WebRadioField.  Thoughts?
> 
>  
> 
> Thanks,
> 
>  
> 
> Brian

[Webunitproj-users] patch 736457 submitted - fix for closing child webclients

[Brian K. <bt...@ne...>]

Patch # 736457 was submitted to the patches section... here are the
details... thanks.

=20

Previously, calling close() on a WebClient would dispose=20
the browser and exit the application, even if it was a=20
child webclient. The submitted patch fixes this by not=20
exiting the application when calling close() on child=20
webclients. The following file was changed to support=20
this:=20

com.zeborg.webunit.ie.ExplorerNativeWebClient=20

Thanks,=20

Brian

[Webunitproj-users] RE: radio tag

[Brian K. <bt...@ne...>]

Just following up to see if anyone has any thoughts on the email
below... thanks,

=20

Brian

=20

-----Original Message-----
From: Brian Knorr=20
Sent: Wednesday, April 02, 2003 1:03 PM
To: 'web...@li...'
Subject: radio tag

=20

Does anyone know how we could use the find method to get a handle on all
the radio tag on a page?  Currently webunit models the input type radio
as a WebTextField, so we end up getting radios and text fields.  We
would to just get the radios.  If there isn't a work around, maybe the
best solution is for webunit to model radios with a class like
WebRadioField.  Thoughts?

=20

Thanks,

=20

Brian

RE: [Webunitproj-users] Issues with child webclients

[Brian K. <bt...@ne...>]

Thanks for the reply... I was able to find the bug with
WebClient.close() - it turns out the code is calling exit() on the
RemoteProcessServer even if it is a child webclient.  I added a fix for
this and will submit it as a patch.

Where would I go to learn more about the Internet Explorer API?  I would
really like to see if it supports this event, that way our test scripts
don't have to emulate javascript closing a window.

Thanks,

Brian

-----Original Message-----
From: Doron [mailto:dg...@sh...]=20
Sent: Monday, May 12, 2003 7:25 AM
To: Brian Knorr
Cc: web...@li...
Subject: Re: [Webunitproj-users] Issues with child webclients

Hi Brian,

	See comments below:

On Fri, 2003-05-09 at 19:18, Brian Knorr wrote:
> We are having issues with child webclients... some of our web
> application's workflows require many browsers windows to be popped up,
> so in our tests we have to keep track of all the child webclients.=20
> One thing we have noticed is that all the child webclients stay open,
> even if the javascript executed on the browser closes it.=20
->> The web browser used by web unit (Internet Explorer) is actually the
COM GUI component container inside a Java window. Hence the close
javascript command may be ignored in this situation. You may need to
investigate the Internet Explorer API (callbacks) to see if there is an
even for this and if so call dispose() on the "Frame" object in java to
close the browser window for that client.
>  Is this a bug?  A work around we tried to conduct in our tests was to
> close the child webclient with WebClient.close(), but instead of
> closing just the child webclient, all the open webclients are closed
> and the test stops.=20
>  Is this a bug too?=20
->> This is probably a bug. Check the close method and see what it does.
It should just do a dispose one the java Frame object.

>  Also, with all these child webclients open (sometimes around 10) the
> process begins to eat up 100% of the cpu and then it starts to take
> about 10 seconds just to enter one value on the screen... sounds like
a
> threading/memory issue - thoughts?
->> Internet explorer will start consuming a good amount of CPU when
there are a lot of windows open and active. It is hard to say what is
causing it without actually profiling the app under your test scenario.
I suggest you look at the task manager and see where the CPU cycles are
spent (Internet explorer or Java). Also check the memory available to
see if that is the issue (although this is not likely).

You could also profile the app using something like optimize it to find
out where it slow.
>=20
> =20
>=20
> Please advise... thanks,
>=20
> =20
>=20
> Brian

Re: [Webunitproj-users] Issues with child webclients

[Doron <dg...@sh...>]

Hi Brian,

	See comments below:

On Fri, 2003-05-09 at 19:18, Brian Knorr wrote:
> We are having issues with child webclients… some of our web
> application’s workflows require many browsers windows to be popped up,
> so in our tests we have to keep track of all the child webclients. 
> One thing we have noticed is that all the child webclients stay open,
> even if the javascript executed on the browser closes it. 
->> The web browser used by web unit (Internet Explorer) is actually the
COM GUI component container inside a Java window. Hence the close
javascript command may be ignored in this situation. You may need to
investigate the Internet Explorer API (callbacks) to see if there is an
even for this and if so call dispose() on the "Frame" object in java to
close the browser window for that client.
>  Is this a bug?  A work around we tried to conduct in our tests was to
> close the child webclient with WebClient.close(), but instead of
> closing just the child webclient, all the open webclients are closed
> and the test stops. 
>  Is this a bug too? 
->> This is probably a bug. Check the close method and see what it does.
It should just do a dispose one the java Frame object.

>  Also, with all these child webclients open (sometimes around 10) the
> process begins to eat up 100% of the cpu and then it starts to take
> about 10 seconds just to enter one value on the screen… sounds like a
> threading/memory issue - thoughts?
->> Internet explorer will start consuming a good amount of CPU when
there are a lot of windows open and active. It is hard to say what is
causing it without actually profiling the app under your test scenario.
I suggest you look at the task manager and see where the CPU cycles are
spent (Internet explorer or Java). Also check the memory available to
see if that is the issue (although this is not likely).

You could also profile the app using something like optimize it to find
out where it slow.
> 
>  
> 
> Please advise… thanks,
> 
>  
> 
> Brian

[Webunitproj-users] Issues with child webclients

[Brian K. <bt...@ne...>]

We are having issues with child webclients... some of our web
application's workflows require many browsers windows to be popped up,
so in our tests we have to keep track of all the child webclients.  One
thing we have noticed is that all the child webclients stay open, even
if the javascript executed on the browser closes it.  Is this a bug?  A
work around we tried to conduct in our tests was to close the child
webclient with WebClient.close(), but instead of closing just the child
webclient, all the open webclients are closed and the test stops.  Is
this a bug too?  Also, with all these child webclients open (sometimes
around 10) the process begins to eat up 100% of the cpu and then it
starts to take about 10 seconds just to enter one value on the screen...
sounds like a threading/memory issue - thoughts?

=20

Please advise... thanks,

=20

Brian

[Webunitproj-users] testing 4-5-6

[Erik O. <eos...@ya...>]

 
 

__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com

RE: [Webunitproj-users] META-Refresh support..

[Erik O. <eos...@ya...>]

Nic,

When I added IFrameTest.java, I think I used this
method, but also don't remember its location/name.

If you locate all files touched in CVS on or about the
same date as IFrameTest, and do some diffs, you should
be able to find it.

--Erik


--- Nic Pottier <Ni...@in...> wrote:
> 
> Hrmm.. 
> 
> Ya, I'm not seeing anything about clearing the cache
> in either WebClient, WebResponse or WebFactory.
> 
> My build is a couple weeks old though, is there a
> more recent release maybe?
> 
> -Nic
> 
> > -----Original Message-----
> > From: Doron [mailto:dg...@sh...]
> > Sent: Thursday, April 24, 2003 6:15 PM
> > To: Nic Pottier
> > Cc: web...@li...
> > Subject: Re: [Webunitproj-users] META-Refresh
> support..
> > 
> > 
> > Hi,
> > 
> > 	You need to flush the page cache. (there is a
> method for that
> > somewhere). Essentially the first page you get is
> cached by webunit in
> > order to give you fast responses. A method is
> provided to 
> > flush this (as
> > I recall). Its been a while since I played with
> it. If this is not it
> > let me know.
> > 
> > 
> > 				-Doron
> > 
> > 
> > On Thu, 2003-04-24 at 17:53, Nic Pottier wrote:
> > > Howdy.
> > > 
> > > I've had great luck using the enterprise webunit
> package to 
> > test a project I'm working on.  The interface is
> simple and 
> > clean, and we're able to test our product despite
> it's heavy 
> > use of javascript which did not work without
> testers.  So in 
> > short, kudos!
> > > 
> > > Now of course, with any kudos there is the
> obligatory 
> > caveat.  One problem we have is that we use
> meta-refresh'es 
> > in order to poll for time consuming operations. 
> IE, the user 
> > takes an action which might take 20 seconds or so,
> and they 
> > are taken to a 'processing' page or the like,
> which polls by 
> > reposting to itself every 2 seconds until the
> process is 
> > complete, at which time it again uses a meta
> refresh tag to 
> > push itself to the final destination page.
> > > 
> > > Although IE correctly deals with this while
> running in the 
> > webunit context, I am unable to get the
> 'refreshed' page from 
> > webunit.  The WebResponse I get is always the
> original page.  
> > I've poked around in the API and haven't seen any
> obvious way 
> > to work around this.   Am I missing something
> here, or is 
> > this just an 'undocumented feature'. ;)
> > > 
> > > Thanks again for such a great tool,
> > > 
> > > -Nic
> > > 
> > > 
> > >
>
-------------------------------------------------------
> > > This sf.net email is sponsored by:ThinkGeek
> > > Welcome to geek heaven.
> > > http://thinkgeek.com/sf
> > > _______________________________________________
> > > Webunitproj-users mailing list
> > > Web...@li...
> > >
>
https://lists.sourceforge.net/lists/listinfo/webunitproj-users
> > 
> > 
> 
> 
>
-------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Webunitproj-users mailing list
> Web...@li...
>
https://lists.sourceforge.net/lists/listinfo/webunitproj-users


__________________________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
http://search.yahoo.com