opensta-devel — OpenSTA developer discussions

[OpenSTA-devel] Re: CodeMax 2.1.0.24

[Daniel S. <da...@Op...>]

Holger Koch wrote:
> > but it looks like it doesn't work with IE6, because MS replaces the
> > character with %7B which doesn't help much.

This problem:
  http://sourceforge.net/tracker/?group_id=10857&atid=110857&func=detail&aid=897856
Is supposed to be fixed in 1.4.3.  If you are having an issue then
please tell us exactly the symtoms and how to reproduce the problem
and I'll fix it - I know this code pretty well now ;-)

Olaf Kock wrote:
> As far as I understand http://%7B is now equivalent to http://{ and
> is interpreted by the gateway in the same way. That is: you see
> http://%7B and the gateway will connect https:// to your server and
> record https:// urls in your script.
> 
> Somebody correct me if I'm wrong - it's been a long time since I
> recorded https traffic.

You entirely are correct Olaf.

The initial posted workaround for the issue was executable patching
to replace the '{' with a '-' , I wasn't entirely happy with this
approach for a real fix so I just made the gateway see '{' and '%7B'
as equivalent.  It wasn't quite that simple but that was the outcome.

Holger: I am interested in your thoughts on a "workaround",
especially where you need Codemax to rebuild the Modeler... both of
the previous "fixes" for this issue are not in Modeler code at all.

I can make the required Codemax available to you whatever - I just
need to know that you are willing to work with us and "share" your
ideas and code before I make the effort to help you... :-)

Cheers
/dan
-- 
Daniel Sutcliffe <Da...@Op...>
  OpenSTA part-time caretaker - http://OpenSTA.org/

Re: [OpenSTA-devel] CodeMax 2.1.0.24

[Olaf K. <ok...@ab...>]

Holger Koch wrote:
> Thanks Olaf,
> 
> but it looks like it doesn't work with IE6, because MS replaces the
> character with %7B which doesn't help much.

As far as I understand http://%7B is now equivalent to http://{ and is
interpreted by the gateway in the same way. That is: you see http://%7B
and the gateway will connect https:// to your server and record https://
urls in your script.

Somebody correct me if I'm wrong - it's been a long time since I
recorded https traffic.

Cheers,
Olaf

Re: [OpenSTA-devel] CodeMax 2.1.0.24

[Holger K. <ho...@gm...>]

Thanks Olaf,

but it looks like it doesn't work with IE6, because MS replaces the
character with %7B which doesn't help much.

Holger

On 8/1/05, Olaf Kock <ok...@ab...> wrote:
> Holger Koch wrote:
> > Hi everyone,
> >
> > I'm looking for CodeMax - 2.1.0.24 to write a little work around for
> > the https://{ problem with IE6. It would be very helpful if anyone
> > could help to find this particluar version of CodeMax.
>=20
> I can't help you with CodeMax (other than point to the FAQ) - but the
> fix for the http://{ issue (note: http://{, not https://{) is included
> in release 1.4.3.
>=20
> Olaf
>=20
>=20
> -------------------------------------------------------
> SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
> from IBM. Find simple to follow Roadmaps, straightforward articles,
> informative Webcasts and more! Get everything you need to get up to
> speed, fast. http://ads.osdn.com/?ad_id=3D7477&alloc_id=3D16492&op=3Dclic=
k
> _______________________________________________
> OpenSTA-devel mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opensta-devel
>

Re: [OpenSTA-devel] New to OpenSTA, need help

[Olaf K. <ok...@ab...>]

Wel...@ao... wrote:
> I have been trying to record a script for an ecommerce site but during the  
> record, on https pages, I get a "Page Cannot Be Displayed" and the script  
> contains a comment:   ![ERROR]: Unsupported protocol packet received :  cannot 
> continue .
>  
> (note, didn't start the record in https so OpenSTA is putting in the  curly 
> brace for me on the URLs, I don't think that's the issue)

OpenSTA usually replaces https:// urls with http://{, but cannot do this
always. If the url is constructed through string concatenation within a
javascript snippet or by a flash component, a java applet or any other
kind of active component, it most likely cannot be replaced. Please look
for this kind of artifacts within your html code.

Cheers,
Olaf

Re: [OpenSTA-devel] CodeMax 2.1.0.24

[Olaf K. <ok...@ab...>]

Holger Koch wrote:
> Hi everyone,
> 
> I'm looking for CodeMax - 2.1.0.24 to write a little work around for
> the https://{ problem with IE6. It would be very helpful if anyone
> could help to find this particluar version of CodeMax.

I can't help you with CodeMax (other than point to the FAQ) - but the
fix for the http://{ issue (note: http://{, not https://{) is included
in release 1.4.3.

Olaf

[OpenSTA-devel] New to OpenSTA, need help

[<Wel...@ao...>]

I have been trying to record a script for an ecommerce site but during the  
record, on https pages, I get a "Page Cannot Be Displayed" and the script  
contains a comment:   ![ERROR]: Unsupported protocol packet received :  cannot 
continue .
 
(note, didn't start the record in https so OpenSTA is putting in the  curly 
brace for me on the URLs, I don't think that's the issue)
 
I've read as much as I can find on this mailing list and the OpenSTA  FAQs 
but it seems like everything keeps telling me to go to the FAQ about  recording 
in https.  I feel ignorant; but reading that FAQ gets me no  closer to 
understanding why this won't work.  Is there anything I can do to  set certificates 
or get past this?   Does anyone have specific  help?  
 
Thanks in advance.
 
- BJxn

[OpenSTA-devel] CodeMax 2.1.0.24

[Holger K. <ho...@gm...>]

Hi everyone,

I'm looking for CodeMax - 2.1.0.24 to write a little work around for
the https://{ problem with IE6. It would be very helpful if anyone
could help to find this particluar version of CodeMax.

Thanks in advance

Holger

[OpenSTA-devel] [ opensta-Bugs-1240930 ] Compiler crashes when bad var SET to array var that isn't

[SourceForge.net <no...@so...>]

Bugs item #1240930, was opened at 2005-07-19 11:41
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110857&aid=1240930&group_id=10857

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Script Language
Group: Crash
Status: Open
Resolution: None
Priority: 5
Submitted By: Daniel Sutcliffe (dansut)
Assigned to: Daniel Sutcliffe (dansut)
Summary: Compiler crashes when bad var SET to array var that isn't

Initial Comment:
This SCL script illustrates the problem:
  Environment
    Description "This causes the compiler to crash"
  Definitions
    CHARACTER ChrNotArray
  Code
    SET NotVar = ChrNotArray[1]
  Exit
It should produce a compliation error because NotVar
doesn't exist and and ChrNotArray is not an array - but
it actually crashes the compiler.  If ChrNotArray is an
INTEGER the compiler doesn't crash and produces the
correct error message.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110857&aid=1240930&group_id=10857

RE: [OpenSTA-devel] Re: Codemax 2

[JC D. <ub...@ya...>]

> I'm still not 100% sure of what type of things you intend to pull out
> as candidates for variables - there's lots of possibilities. Maybe a
> quick simple example would help. Whatever it sounds like something
> which would be generically useful and thus something which more
> people would be willing to help you achieve..

If you take the example of the framework I've given, the candidates should be the identifiers that have to be extracted from a call and used in subsequent calls (note that this framework has only few methods, and to step through a business process, you always have to call the same method, let's call this method "NextStepBP"). As I explained, there were 3 identifiers: the first one was used as a session ID to identifier a user job. the second one was identifing the step of  a business process, and the last one was identifing the businness process itself (more then one business process could be used at the same by the same user. For instance : Let's call a BP "search city" and another "create person". Now, a user want to create a profile for a new person. He then begin a "create peron" BP :

Call NextStepBP?BpId=1 (BpId=1 being the identifier for the "create person" BP). The framework returns in the response body something like :

- a form in wich you can enter some information about the person (name, function, etc)

- a link to NextStepBP?UserId=1&BpId=1&StepId=2 . A click on this link validate the information of the form and bring the user to the second step of profile creation.

The frramework now returns in the response body something like :

- a form in which you can enter more informatioin about the person (adress, ...)

- a link to NextStepBP?UserId=1&BpId=1&StepId=3

And this time, related to the adress information, you can also find a link like:

- NextStepBP?UserId=1&BpId=2&StepId=1 . This link launch a "search city" BP (for example, this BP helps you find a city name from its zip code).Let's say the user step through that BP. we will now have responses like :

NextStepBP?UserId=1&BpId=2&StepId=2 then NextStepBP?UserId=1&BpId=2&StepId=3, etc. and when this BP is finished, we get back to the step of the first BP, where we left it so the next step would be:

NextStep?UserId=1&BpId=1&StepId=3, then NextStep?UserId=1&BpId=1&StepId=4, etc. until completion of the very first business process.

 Note that the actual identifiers are more complicated than the one given in this example. They are uniques and distincts. And also, this example show a very simple case. It is very common to have up to 5 BP started at the same time, with overlapped stepping. Now you understand that script variabilisation is not straightforward. Therefore, it could be handled very well (I believe) by simple rules and comparison. For instance, each time you receive a response from that framework, you can apply a rule to extract those identifiers (from boundaries for example), compare them with already existing variables to find out if the current BP, user, etc., then create new variables from the remaining ids that will be used in subsequent calls (this is what I call correlation of the variables).

I hope this example clarify this concept of automatic correlation and the kind of candidate variables.

>Given the rough idea of what you want to achieve I think you might
>actually be looking at the wrong place for code to alter. The
>recording is performed by the Gateway (a completely seperate
>component to the Modeler) and produces a TOF file. The Modeler uses
>the Tof2SCL.dll to then translate this into the SCL... Maybe you'd
>be better looking at these areas, or maybe I'm grabbing the wrong
>end of the stick ;-)


And I believe you are right !!! My problem is, I don't have a clue (well yes but its very fuzzy) about OpenSTA internals. What are the components ? What do they do ? ...

By the way, thank you for this (quick) explanation. I have been able to succefully compile the Gateway with VC.NET, so I'm going to see what I can do from it now. I'll keep you posted with my experiences.

>Aha - it sounds like much of what you are doing are not really what
>I'd call "fixes" but more "porting" to a different compilation
>environment. Some times you'll find general issues here that should
>be fixed because the new environment has better checking - other
>times you'll have to break compilation in the new environment to make
>the new one work. Obviously I'm interested in the former, but the
>latter needs to be treated carefully.

Ok, sorry for the misuse of langage.

>Anyway, don't be afraid to ask about anything in the source - publish
>anything you find "strange" here, and generally keep us informed of
>what you are doing - it's much easier to get something going when you
>are not working totally alone.


I'll bear that in mind whilst working on OSTA.

Cheers


		
---------------------------------
 Appel audio GRATUIT partout dans le monde avec le nouveau Yahoo! Messenger
 Téléchargez le ici !  

[OpenSTA-devel] Re: Query regarding Opensta

[Daniel S. <da...@Op...>]

Manisha Dutta wrote:
> I have Opensta 1.4.3.2 version installed on my PC.  My
> configuration is P-IV with 512KB Ram.
> 
> I have recorded a scenario wherein I am logging in to the site and
> entering a buy order and then submitting the order.With this site
> we can also view the order which has been created.
> 
> So after I run this test for 1 user, I cannot view the new order
> created though OpenSTA doesn't show any error.

You are posting in the wrong place: this mailing list is for
discussion of the development (compiling and altering) of OpenSTA.
You are looking for user level support - places to go for this are
listed in this FAQ:
  http://portal.opensta.org/faq.php?topic=OstaSupport

Plus your question could probably be answered by simply spending some
time reading through the Frequently Asked Questions:
  http://portal.opensta.org/faq.php

> Can anybody pls let me know how I can solve this issue.

Yes!  If you post in the right place and show some evidence that
you've actually made a little bit of effort to find the answer
yourself first...

Thanks for making the effort,
/dan
-- 
Daniel Sutcliffe <Da...@Op...>
  OpenSTA part-time caretaker - http://OpenSTA.org/

[OpenSTA-devel] Re: Codemax 2

[Daniel S. <da...@Op...>]

JC Dus wrote:
> I used OpenSTA for the first time in my last mission and found out
> it's a pretty great tool for Web load testing.  In this specific
> context, I had to deal with a proprietary framework wich was using
> many identifiers (well three, but tricky enough) to ensure
> client/server conversation to be in a valid state, and I found myself
> spending more time looking in the logs to put the pieces together
> then scripting.  So I thought OpenSTA would really benefit such
> features as automatic correlation of this identifiers.  Right now, I
> am really thinking of basic stuff, like using rules (if possible) to
> automatically create variables, or comparing recorded HTTP
> conversation during a replay with the current HTTP conversation to
> find out differences.  These differences would then be a pool of
> candidate variables to be created in the script.

I'm still not 100% sure of what type of things you intend to pull out
as candidates for variables - there's lots of possibilities.  Maybe a
quick simple example would help.  Whatever it sounds like something
which would be generically useful and thus something which more
people would be willing to help you achieve...

> > > But since I haven't got VC++ 6 at my disposal, I am currently
> > > trying to compile a dotNet version of the modeller, so I could
> > > then write my addons in C#.
> 
> > That, in itself, is a project that I believe will be of interest
> > to many here - I don't have dotNet but I'll be interested in
> > hearing all this entails...
> 
> Yes, I realized that when I first try to compile the full build.
> That's why I have decided to focus on the Modeler (which is the part
> that I really need to compile succefully),  By still producing
> correct SCL scripts, I could use the "standard" OpenSTA components
> to run the tests.

Given the rough idea of what you want to achieve I think you might
actually be looking at the wrong place for code to alter.  The
recording is performed by the Gateway (a completely seperate
component to the Modeler) and produces a TOF file.  The Modeler uses
the Tof2SCL.dll to then translate this into the SCL...  Maybe you'd
be better looking at these areas, or maybe I'm grabbing the wrong
end of the stick ;-)

> > > I already fixed a few things but now I miss codemax to go
> > > further.
> 
> > Please provide more information on your "fixes" and make sure
> > that all these get back to me, however small.
> 
> Ok, my two cents :

Aha - it sounds like much of what you are doing are not really what
I'd call "fixes" but more "porting" to a different compilation
environment.  Some times you'll find general issues here that should
be fixed because the new environment has better checking - other
times you'll have to break compilation in the new environment to make
the new one work.  Obviously I'm interested in the former, but the
latter needs to be treated carefully.

> Sorry but I am working alone on this so I didn't bother use a
> versioning tool.

No need for you to work in any versioning tool - although I find it
useful even when working alone - just keep a copy of the virgin
sources around and a good diff/comparator tool can help remind you
of what you've had to change.  Another hint, and something I find
very difficult to do is to keep your changes as small as possible
and don't alter tabs to spaces (etc.) - this just helps keeping track
of the changes easier, even if you find the coding standards horrible
to read through (which I mostly do)...

Anyway, don't be afraid to ask about anything in the source - publish
anything you find "strange" here, and generally keep us informed of
what you are doing - it's much easier to get something going when you
are not working totally alone.

Cheers
/dan
-- 
Daniel Sutcliffe <Da...@Op...>
  OpenSTA part-time caretaker - http://OpenSTA.org/

[OpenSTA-devel] Query regarding Opensta

[Manisha D. <man...@xp...>]

Hi All,

I have Opensta 1.4.3.2 version installed on my PC. My configuration is =
P-IV with 512KB Ram.

I have recorded a scenario wherein I am logging in to the site and =
entering a buy order and then submitting the order.With this site we can =
also view the order which has been created.

So after I run this test for 1 user, I cannot view the new order created =
though opensta doesn't show any error.

Can anybody pls let me know how I can solve this issue.

Regards,

Manisha

RE: [OpenSTA-devel] Re: Codemax 2

[JC D. <ub...@ya...>]

>> Well, I was thinking of writing an autocorrelation tool for the
>> script's variables. 
> Sounds interesting - please tell us more.
I used OpenSTA for the first time in my last mission and found out it's a pretty great tool for web load testing. In this specific context, I had to deal with a proprietary framework wich was using many identifiers (well three, but tricky enough) to ensure client/server conversation to be in a valid state, and I found myself spending more time looking in the logs to put the pieces together then scripting. So I thought OpenSTA would really benefit such features as automatic correlation of this identifiers. Right now, I am really thinking of basic stuff, like using rules (if possible) to automatically create variables, or comparing recorded HTTP conversation  during a replay with the current HTTP conversation to find out differences. These differences would then be a pool of candidate variables to be created in the script.   

>> But since I haven't got VC++ 6 at my disposal, I am currently
>> trying to compile a dotNet version of the modeller, so I could then
>> write my addons in C#.
>That in itself is a project that I believe will be of interest to
>many here - I don't have dotNet but I'll be interested in hearing
>all this entails...

Yes, I realized that when I first try to compile the full build. That's why I have decided to focus on the modeler (which is the part that I really need to compile succefully), By still producing correct SCL scripts, I could use the "standard" OpenSTA components to run the tests.

>> I already fixed a few things but now I miss codemax to go further.
>Please provide more information on your "fixes" and make sure that
>all these get back to me, however small.
Ok, my two cents :
I am solving the errors in the order I get them from the compiler. Until now, I have successfully built :
CyrToolsOS
Msglib
HTMLParser
HttpCaptureCmd
prdlib
iecacheman
CyrODBC
SCLParserLib
XMLGentables
CECGridLib
The latest fix I made was in the CECGridlib.lib project,
In the file GridCtrl.cpp, I was having the following error (I think there was only two errors of this kind) :
q:\OpenSTA\ostaw32-1-4-3\src\BaseUI\CECGridLib\GridCtrl.cpp(3586): error C2668: 'CUIntArray::InsertAt' : ambiguous call to overloaded function
In the method int CGridCtrl::InsertRow(LPCTSTR strHeading, int nRow /* = -1 */)
at the line
        else
        {
  =>     m_arRowHeights.InsertAt(nRow, 0); // Note: I don't remember, but I think there might have been an int cast of the 0 value.
            if (!GetVirtualMode())
                m_RowData.InsertAt(nRow, new GRID_ROW);
        }

and in the method :
int CGridCtrl::InsertColumn(LPCTSTR strHeading,
                            UINT nFormat /* = DT_CENTER|DT_VCENTER|DT_SINGLELINE */,
                            int nColumn  /* = -1 */)

at the line 
        else
        {
 =>       m_arColWidths.InsertAt(nColumn, 0); // Same remark as before
            if (!GetVirtualMode())
            {
                for (int row = 0; row < m_nRows; row++)
                {
                    GRID_ROW* pRow = m_RowData[row];
                    if (!pRow)
were causing the problem.
In array_u.cpp, I found that the InsertAt method has the following two definitions :
void CUIntArray::InserAt(INT_PTR nIndex,UINT newElement, INT_PTR nCount=1)
void CUIntArray::InsertAt(INT_PTR nStartIndex, CUIntArray * pNewArray)
The fact that the 0 value in both incriminated method calls isn't cast properly makes it undecidable for the compiler to choose the right method. I tried doing a static_cast of those value to CUIntArray *, but it didn"t worked. A dirty cast did the job though... : m_arColWidths.InsertAt(nColumn, (CUIntArray *)0); 
Note that I actually didn't really now that this is the correct cast (maybe it is a UInt cast which is correct), but I didn't wanted to start reverse engineer too much code.
The next error was about CodeMax. I then tried CodeMax 4 in trial version, but found out that the code is no longer compliant. I tried the yahoo group, with the result you know. 
Sorry but I am working alone on this so I didn't bother use a versionning tool. Since I am ponctually working on this project among others at my spare time, (I began working on this for something like 2 months ago and during these 2 months, I only get to work on this project 4 or 5 times) thus I don't remember any other fix I could have done, but I don't believe there was much to do to compile the first libs. I will keep from now on a trace of any changes done, and keep you posted of any significant progress. 
 

 

		
---------------------------------
 Appel audio GRATUIT partout dans le monde avec le nouveau Yahoo! Messenger
 Téléchargez le ici !  

[OpenSTA-devel] Re: Codemax 2

[Daniel S. <da...@Op...>]

> > That's great. What are you planning - anything we could help
> > with?  At this point one thing I'd really like to see is a
> > progress bar for the Tof2Scl conversion after recording... and
> > of course getting rid of CodeMax would avoid just these sort of
> > issues.
> 
> Well, I was thinking of writing an autocorrelation tool for the
> script's variables.  

Sounds interesting - please tell us more.

> But since I haven't got VC++ 6 at my disposal, I am currently
> trying to compile a dotNet version of the modeller, so I could then
> write my addons in C#.

That in itself is a project that I believe will be of interest to
many here - I don't have dotNet but I'll be interested in hearing
all this entails...

> I already fixed a few things but now I miss codemax to go further.

Please provide more information on your "fixes" and make sure that
all these get back to me, however small.

> > I luckily joined up before they stopped accepting new members so
> > I can help you with this... just want to make sure that you're
> > going to help us in return by making sure that any work you do on
> > the Modeler gets back into the mainstream version...

Sorry to even have to bring this matter up.  It's just that so many
times I hear about fixes for various bits of OpenSTA and I so rarely
get to see anything resembling hints (never mind code) for these
fixes.  I just want to make sure that if I'm helping anyone then they
are committed to returning the favor by helping OpenSTA and its
community.

> Of course, if any of the work I will be able to complete is of
> relevance for you.  I would be happy to provide you with sources
> and instructions.

ANYTHING that you: work out, discover, achieve, fail with, regarding
OpenSTAs source code is valuable to the people here now and in the
future.  Right now the best way to get information back to other
people also interested in developing OpenSTA is by posting it to this
mailing list.  Whatever information you provide may help others to
help you out or continue your work if you move onto other things. 
Open Source doesn't work if people don't publish their knowledge.

I look forward to hearing more from you here,
Cheers
/dan
-- 
Daniel Sutcliffe <Da...@Op...>
  OpenSTA part-time caretaker - http://OpenSTA.org/

RE: [OpenSTA-devel] Re: Codemax 2

[JC D. <ub...@ya...>]

Daniel Sutcliffe <da...@Op...> a écrit :
JC Dus wrote:
> I'm planning to do some work on the script modeler

That's great. What are you planning - anything we could help with?
At this point one thing I'd really like to see is a progress bar for
the Tof2Scl conversion after recording... and of course getting rid
of CodeMax would avoid just these sort of issues.


Well, I was thinking of writing an autocorrelation tool for the script's variables.

But since I haven't got VC++ 6 at my disposal, I am currently trying to compile a dotNet version of the modeller, so I could then write my addons in C#. I already fixed a few things but now I miss codemax to go further.

I luckily joined up before they stopped accepting new members so I
can help you with this... just want to make sure that you're going to
help us in return by making sure that any work you do on the Modeler
gets back into the mainstream verstion...


Of course, if any of the work I will be able to complete is of relevance for you. I would be happy to provide you with sources and instructions.

Cheers


		
---------------------------------
 Appel audio GRATUIT partout dans le monde avec le nouveau Yahoo! Messenger
 Téléchargez le ici !  

[OpenSTA-devel] Re: Codemax 2

[Daniel S. <da...@Op...>]

JC Dus wrote:
> I'm planning to do some work on the script modeler

That's great.  What are you planning - anything we could help with?
At this point one thing I'd really like to see is a progress bar for
the Tof2Scl conversion after recording...  and of course getting rid
of CodeMax would avoid just these sort of issues.

> and, in order to compile it, I need codemax version 2.??

2.1.0.24 is the latest version but not much has really changed since
2.1.0.22 ...

> Seems that the codemax yahoo group can't be joined.  So if anybody
> could provide me with these sources, I would be very grateful.

I luckily joined up before they stopped accepting new members so I
can help you with this... just want to make sure that you're going to
help us in return by making sure that any work you do on the Modeler
gets back into the mainstream verstion...

Cheers
/dan
-- 
Daniel Sutcliffe <Da...@Op...>
  OpenSTA part-time caretaker - http://OpenSTA.org/

[OpenSTA-devel] Codemax 2

[JC D. <ub...@ya...>]

Hello everybody, 
 
I'm planning to do some work on the script modeler and, in order to compile it, I need codemax version 2.?? 
Seems that the codemax yahoo group can't be joined. So if anybody could provide me with these sources, I would be very grateful.
Thanks in advance.
 

		
---------------------------------
 Appel audio GRATUIT partout dans le monde avec le nouveau Yahoo! Messenger
 Téléchargez le ici !  

[OpenSTA-devel] [ opensta-Bugs-730313 ] Script generation error with www-authenticate

[SourceForge.net <no...@so...>]

Bugs item #730313, was opened at 2003-04-30 17:17
Message generated for change (Comment added) made by brianavid
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110857&aid=730313&group_id=10857

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: HTTP Capture
Group: Inconvenience
Status: Open
Resolution: None
Priority: 5
Submitted By: Jerome Delamarche (trickyjer)
Assigned to: Nobody/Anonymous (nobody)
Summary: Script generation error with www-authenticate

Initial Comment:
When the session being registered by the Modeler 
contains Windows authentication messages, the 
generated script forgets to declare some "blob" 
variables:

Load Response_Info Header on 3		&
	Into blob_3_0	&
	,WITH "WWW-Authenticate"

does not declare the "bloc_3_0" variable.

Workaround: fix by hand !



----------------------------------------------------------------------

Comment By: Brian Collins (brianavid)
Date: 2005-06-23 13:36

Message:
Logged In: YES 
user_id=1214182

Also note that the space declared for NTLM blobs is 256 
CHARACTERS, but HTTP www-authenticate responses have 
been seen larger than this. I manually edit the generated 
declarations to 512 CHARACTERs to get it to work (mostly).

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110857&aid=730313&group_id=10857

[OpenSTA-devel] [ opensta-Bugs-730317 ] Bad Windows Auth handling

[SourceForge.net <no...@so...>]

Bugs item #730317, was opened at 2003-04-30 17:24
Message generated for change (Comment added) made by brianavid
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110857&aid=730317&group_id=10857

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: HTTP Capture
Group: Design
Status: Open
Resolution: None
Priority: 5
Submitted By: Jerome Delamarche (trickyjer)
Assigned to: Nobody/Anonymous (nobody)
Summary: Bad Windows Auth handling

Initial Comment:
The script generated by the Modeler when the session 
requires a login/password cannot be replayed 
successfully.

The Server first challenges for a "Basic" authentification, 
then it turns this auth. into a NTLM auth.

The generated script includes statements such as:

BUILD AUTHENTICATION BLOB	&
FOR BASIC &
FROM USER "admin" PASSWORD "" DOMAIN "" &
INTO blob_2_1

and further:

Load Response_Info Header on 3		&
Into blob_3_0	&
,WITH "WWW-Authenticate"

BUILD AUTHENTICATION BLOB	&
FOR NTLM	&
FROM BLOB blob_3_0	&
INTO blob_3_0

But requests are all denied by the Web server.
IIS server just require a valid user for itself, it does not 
belong to a Windows Domain.



----------------------------------------------------------------------

Comment By: Brian Collins (brianavid)
Date: 2005-06-23 13:35

Message:
Logged In: YES 
user_id=1214182

Also note that the space declared for NTLM blobs is 256 
CHARACTERS, but HTTP www-authenticate responses have 
been seen larger than this. I manually edit the generated 
declarations to 512 CHARACTERs to get it to work (mostly).

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110857&aid=730317&group_id=10857

[OpenSTA-devel] Re: Duration of the viewstate encoding

[Daniel S. <da...@op...>]

Daniel Sutcliffe wrote:
> I'm working on some other builds at the moment - one that is similar
> to the above build but where the STLport is set up to just provide
> shells around the platform iostreams implementation; another doesn't
> use the STLport at all.

Well today I completed a build of OpenSTA without using the STLport
at all, but still using the latest PSDK.  I had to make a few
modifications and was left with plenty of warnings but it does build
and appeared to basically work...

I then ran the benchmark we have been discussing and was blown away
with the results:
                      1.4.3  STLport4.6.2 No STLport
                     Release    Build       Build
      /  URLENCODE1    5.51      1.82       0.68
VSTATE1  URLENCODE2    7.21      2.17       0.81
      \  URLENCODE3  180        70.1       33.4

      /  URLENCODE1    1.26      0.40       0.14
VSTATE2  URLENCODE2    1.38      0.42       0.15
      \  URLENCODE3  189        68.2       36.2

I've double checked this and added debugging to the script to make
sure the encoding was actually being done - and it seems kosher.

I think this shows a fairly clear path forward at this point,
obviously lots of testing needs doing.  I'd also like a 3rd party
to replicate my results.  Any thoughts on these results or what our
path forward should be?

/dan
-- 
Daniel Sutcliffe <Da...@Op...>
  OpenSTA part-time caretaker - http://OpenSTA.org/

[OpenSTA-devel] Re: Duration of the viewstate encoding

[Daniel S. <da...@Op...>]

Thierry Boullet wrote:
> I made some measurements of the duration of the viewstate
> encoding because I noted that it is not negligible and that it can
> disturb the tests.

There's no argument that doing SCL string encoding is inneficient and
that we need to add a way of doing this efficiently in SCL...

However this mail isn't about a native  SCL encoding method, instead
it's about the benchmarks that Thierry created and the possibility of
speeding OpenSTA SCL string handling up by building it differently.

I have taken Thierry's URL encoding methods, and his viewstate data,
added a third encoding method (a really inneficient one) and packaged
them up so they can be run by anyone, anywhere.

The SCL mechanism Thierry referred to as METHOD2 (the method Thierry
uses) has become my URLENCODE1
The SCL mechanism Thierry referred to as METHOD1 (the FAQ method) has
become my subroutine URLENCODE2
The SCL mechanism I created (a simple one which is written without
efficiency in mind) has become my subroutine URLENCODE3

The script that contains these 3 subroutines reads the viewstate data
from a .FVR data file and times each subroutine for each of the
viewstates, it doesn't do any HTTP.  You can download this test here:
  http://ftp.opensta.org/tests/urlencode-050620.zip
I ran the script in a test so it iterated 20 times on a machine that
had been freshly rebooted each time.

The test machine was a P4/2.8GHz with 1Gb RAM running WinXPproSP2.  I
ran the tests first with the stock 1.4.3 release and then again with
a CVS source build, built against the WinServer2003SP1 PSDK and
STLport 4.6.2 - the other externals are as recommended for 1.4.3, the
build needed some small "adjustments" to OmniORB and the STLport to
work cleanly...  STLport was set up to use its own IOSTREAMS.

The results (just like Thierry's) are pretty conclusive:
                      1.4.4        New
                     Release      Build
      /  URLENCODE1    5.51        1.82
VSTATE1  URLENCODE2    7.21        2.17
      \  URLENCODE3  180          70.1

      /  URLENCODE1    1.26        0.40
VSTATE2  URLENCODE2    1.38        0.42
      \  URLENCODE3  189          68.2

I'm working on some other builds at the moment - one that is similar
to the above build but where the STLport is set up to just provide
shells around the platform iostreams implementation; another doesn't
use the STLport at all.  I'd like to try an STLport 5.0 build but
it's driving me insane ... ;-)

Thoughts, comments, etc. all welcome
Cheers
/dan
-- 
Daniel Sutcliffe <Da...@Op...>
  OpenSTA part-time caretaker - http://OpenSTA.org/

[OpenSTA-devel] Re: New build environment

[Daniel S. <da...@Op...>]

Mark Elam wrote:
> Just a a thought from my own time playing with the build
> environment, the code will build fine without using *any* version
> of STL_Port (a couple of minor changes)

The point is well taken Mark and this has cross my mind more than a
few times over the last couple of days...

> and although, If memory serves, there was a reason for using
> STL_Port 5 years ago because of some problems with the MS STL I do
> not know if this is still the case.

This is as far as my memory stretches as well.  I knew there *were*
bugs in the M$ STL implementation and that was one of the reasons
STLport was used - the other being a common STL for cross platform
coding (the standard wasn't ratified back then) which became
meaningless once the  switch to Win32 only was made...

My big problem is I can't find any documentation as to what the STL
bugs were (or even what their symptoms were) that prompted STLport
usages.  Which would be nice for piece of mind.

I think that I will just produce a build and give it a try...

> As for future cross-platform compatibility etc. The STL certainly
> was part of the C++ ansi standard and is, to my knowledge, quite
> complete in most major platform's compilers as for as the sort of
> code that OpenSTA requires.

As far as I can tell; todays reason for using the STLport over the
native implementation is really more for speed than anything else...
This is nice (I guess) but some benchmarking would be nice to prove
it useful for us - there'll be more on this in Thierry's viewstate
thread, later.

There are also the 2 major modes that STLport can be used in:
 - with their compiled iostreams implementation.
 - were their iostreams just provides a shell to the native.

I think that the goal of simplifying OpenSTAs recommended build
environment means that we should definitely consider making sure
the build will work without STLport.  Although it would be nice to
leave it as a build option (maybe used in the Release versions) if it
does give a reasonable speed boost.

Thanks for your feedback Mark, I'm interested in all others thoughts
on this and related matters,
Cheers
/dan
-- 
Daniel Sutcliffe <Da...@Op...>
  OpenSTA part-time caretaker - http://OpenSTA.org/

Re: [OpenSTA-devel] New build environment

[Mark E. <ma...@bo...>]

Daniel Sutcliffe wrote:
> 
> This is a problem!  The current recommended PSDK is the Windows 2003
> sp1 PSDK and causes issues - it specifically clashes with the STLport
> 4.6.2 version.  These aren't major problems (basically the PSDK
> #define's __out to nothing which gives some templates a headache) but
> they were enough to make me read further...  Basically the STLport
> developers current attitude towards the 4.6.x release appears to be
> that it is bad and dead and the upcoming 5.0 is better even in its
> pre-release status!
>   http://www.stlport.com/dcforum/DCForumID6/1757.html#3
> However the RC3 and CVS versions of 5.0 have many issues building
> OpenSTA, although none seem to be really major.  4.6.2 can be fixed
> but ...
> 
> I've tried building with many different versions of STLport (released,
> beta, CVS) and all have some issues at different points.  I'd really
> just like to hear what other peoples views are on which direction and
> priorities we ought to have in this area at this time...
> 

Just a a thought from my own time playing with the build environment, 
the code will build fine without using *any* version of STL_Port (a 
couple of minor changes) and although, If memory serves, there was a 
reason for using STL_Port 5 years ago because of some problems with the 
MS STL I do not know if this is still the case. As for future 
cross-platform compatibility etc. The STL certainly was part of the C++ 
ansi standard and is, to my knowledge, quite complete in most major 
platform's compilers as for as the sort of code that OpenSTA requires.

Thought I'd mention it :-)
Mark

[OpenSTA-devel] New build environment

[Daniel S. <da...@Op...>]

Inspired by Thierry's speed increase findings and the overall goal
of giving OpenSTA an up-to-date build environment I've been
experimenting...

Unfortunately I've not been having too much luck - so I'm going to
post my findings here in the hope of getting feedback, inspiration or
just ideas from others.

Thierry's good build environment uses a PSDK from Feb 2003 and
STLport 4.6.2 (the latest non-beta according to the Web site).  The
other are left as per the the current recommended elements; they
either have no updates, no good reason for updating or the switch is
a very major one.  This is a known good build, but it didn't fulfill
my goals of an *up-to-date* build environment - specifically I wanted
to use the current M$ PSDK as a priority...

This is a problem!  The current recommended PSDK is the Windows 2003
sp1 PSDK and causes issues - it specifically clashes with the STLport
4.6.2 version.  These aren't major problems (basically the PSDK
#define's __out to nothing which gives some templates a headache) but
they were enough to make me read further...  Basically the STLport
developers current attitude towards the 4.6.x release appears to be
that it is bad and dead and the upcoming 5.0 is better even in its
pre-release status!
  http://www.stlport.com/dcforum/DCForumID6/1757.html#3
However the RC3 and CVS versions of 5.0 have many issues building
OpenSTA, although none seem to be really major.  4.6.2 can be fixed
but ...

I've tried building with many different versions of STLport (released,
beta, CVS) and all have some issues at different points.  I'd really
just like to hear what other peoples views are on which direction and
priorities we ought to have in this area at this time...

The other little problem occurs when linking to the htmlhelp.lib in
this new PSDK - give these errors:
  htmlhelp.lib(init.obj) : error LNK2001:
    unresolved external symbol ___security_cookie
  htmlhelp.lib(init.obj) : error LNK2001:
    unresolved external symbol @__security_check_cookie@4
Explained here:
  http://blogs.msdn.com/nikolad/archive/2005/01/27/362214.aspx

All thoughts and comments on where we should go are very welcome.

Cheers
/dan
-- 
Daniel Sutcliffe <Da...@Op...>
  OpenSTA part-time caretaker - http://OpenSTA.org/

[OpenSTA-devel] [ opensta-Bugs-1095283 ] Starting non-existent Test using TestInit fails badly

[SourceForge.net <no...@so...>]

Bugs item #1095283, was opened at 2005-01-03 21:56
Message generated for change (Settings changed) made by ghall
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110857&aid=1095283&group_id=10857

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Architecture
Group: Behavioral
Status: Open
Resolution: Accepted
Priority: 5
Submitted By: Daniel Sutcliffe (dansut)
>Assigned to: Geoff Hall (ghall)
Summary: Starting non-existent Test using TestInit fails badly

Initial Comment:
If a Test that does not exist in your Repository is
attempted to be started using the command line
tool TestInit will look like a test is started.  An error
message should be reported.  It will also creating a
looping TestManager that can only be killed manually
and this will create directories in the Repository for
Test results for the non-existent Test - which
creates confusing non-sensical displays in the
Commander.

eg. TestInit -start -T NONSENSE -S 10
Will create a Repository/Tests/Nonsense dir with
a dated result dir below this.  If the Test Nonsense
never existed this will create a mess in the Tests
tree in the Commander.

To recover from this stituation the TestManager.exe
process must be ended from the Windows Task
Manager and the Nameserver should be restarted.


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110857&aid=1095283&group_id=10857