From dev-return-3389-apmail-jmeter-dev-archive=jmeter.apache.org@jmeter.apache.org Tue Sep 9 11:23:40 2014 Return-Path: X-Original-To: apmail-jmeter-dev-archive@minotaur.apache.org Delivered-To: apmail-jmeter-dev-archive@minotaur.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 8841511FB3 for ; Tue, 9 Sep 2014 11:23:40 +0000 (UTC) Received: (qmail 52518 invoked by uid 500); 9 Sep 2014 11:23:40 -0000 Delivered-To: apmail-jmeter-dev-archive@jmeter.apache.org Received: (qmail 52490 invoked by uid 500); 9 Sep 2014 11:23:40 -0000 Mailing-List: contact dev-help@jmeter.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: dev@jmeter.apache.org Delivered-To: mailing list dev@jmeter.apache.org Received: (qmail 52475 invoked by uid 99); 9 Sep 2014 11:23:39 -0000 Received: from nike.apache.org (HELO nike.apache.org) (192.87.106.230) by apache.org (qpsmtpd/0.29) with ESMTP; Tue, 09 Sep 2014 11:23:39 +0000 X-ASF-Spam-Status: No, hits=0.3 required=5.0 tests=FREEMAIL_REPLY,RCVD_IN_DNSWL_LOW,SPF_PASS X-Spam-Check-By: apache.org Received-SPF: pass (nike.apache.org: domain of sebbaz@gmail.com designates 74.125.82.49 as permitted sender) Received: from [74.125.82.49] (HELO mail-wg0-f49.google.com) (74.125.82.49) by apache.org (qpsmtpd/0.29) with ESMTP; Tue, 09 Sep 2014 11:23:13 +0000 Received: by mail-wg0-f49.google.com with SMTP id m15so2891533wgh.32 for ; Tue, 09 Sep 2014 04:23:12 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type; bh=f6IkxmspbZ6PkujJrtD1O+hQsyHRzkBFNxy+wMqSNd0=; b=uiEt3vXFn+ytUdSux6tyIe9BCSL7ZPxfIk+EXu+EHcc8gMbPzoXaW/HY53RSIEibr7 Gvs/DyDxXohSJarYk29F1Em9uixfcpiUmDka2+GH1vMwjfDKHEK5+yPXUA72xri4MKY2 rJEJgjLI4eNudZc4j8QnnB13nBGouEjygNK0WlOqW4hfXaGVlWHwAxQm9HXSb2e4XMPM GbNGDPVofYNlj5mXkjmEw0rnuENAPz9VoxQUTDSLB29v6iWWWyhp0jbtAOA4XUwxQa9v ONaYMybzuUAaxlolh4mjEY8MVnm8d9dT4Uq/JEJ0fFVekR5/+8qhy6ZXBR+fpzrqoSzN T9HQ== MIME-Version: 1.0 X-Received: by 10.180.86.65 with SMTP id n1mr29740412wiz.41.1410261792618; Tue, 09 Sep 2014 04:23:12 -0700 (PDT) Received: by 10.194.162.231 with HTTP; Tue, 9 Sep 2014 04:23:12 -0700 (PDT) In-Reply-To: <540EB396.1090700@ya.ru> References: <51DFB939.1080306@apache.org> <51E19C5B.2040609@apache.org> <51E1C753.50406@apache.org> <51E1DA23.3070704@apache.org> <540564EA.2000301@gmail.com> <540EB396.1090700@ya.ru> Date: Tue, 9 Sep 2014 12:23:12 +0100 Message-ID: Subject: Re: Add an icon form Templates in toolbar From: sebb To: dev@jmeter.apache.org Content-Type: text/plain; charset=UTF-8 X-Virus-Checked: Checked by ClamAV on apache.org On 9 September 2014 09:00, Andrey Pohilko wrote: > Hi, > > My opinion on the cached sample handling: > > 1. Not returning sample is wrong. Once user has Sampler he expects > sample to be generated in that place. Without a sample we will have > questions "what's happening" from users. If we look into Firebug or > Developer Tools, we'll see that there are "samples' for cached content, > having "from cache" as size qualifier. > 2. Response code 204 is also wrong, it means different situation in RFC. > The response code should be one from cached response. I suppose server > can set caching headers for any of 2xx coded responses and client should > respect that. > > I would suggest following: > a) return the code that were present in original cached response. > and the response message also > b) in case of cached response add " (cached)" string to response > message. Response message string is not defined as required by RFC and > may contain any string in it, it is 100% informational. For example, > original sample "200 OK" will look as "200 OK (cached)" > c) enable this by default, no property needed since the behavior is > unambiguous It's a bit harder to detect the sample, which is why I was thinking about defining a specific non-HTTP return code. This could be done using a property to override the 200 code. Also JMeter does not cache the content, so the sample won't have any data. Post-Processors may need to be aware of this. Though given that 204 status is also success, and was not excluded previously, perhaps this is not an issue. > Andrey Pokhilko > > On 09.09.2014 00:41, sebb wrote: >> On 8 September 2014 21:21, Philippe Mouawad wrote: >>> On Mon, Sep 8, 2014 at 10:08 PM, sebb wrote: >>> >>>> On 8 September 2014 11:56, Philippe Mouawad >>>> wrote: >>>>> On Mon, Sep 8, 2014 at 12:33 PM, sebb wrote: >>>>> >>>>>> On 7 September 2014 19:49, Philippe Mouawad >>>>> wrote: >>>>>>> On Sun, Sep 7, 2014 at 2:33 PM, sebb wrote: >>>>>>> >>>>>>>> On 6 September 2014 22:44, Philippe Mouawad < >>>> philippe.mouawad@gmail.com >>>>>>>> wrote: >>>>>>>>> Hello All, >>>>>>>>> >>>>>>>>> We could make a poll with 3 items: >>>>>>>> [This should have been a new thread, and all these should be separate >>>>>>>> user threads] >>>>>>>> >>>>>>>>> 1) Ask Early users to test and give feedback on Undo/Redo feature >>>> in >>>>>>>>> nightly build >>>>>>>> +1 >>>>>>>> >>>>>>> Let me just finish the Toolbar enabling/disabling >>>>>>> >>>>>>>>> 2) Make a poll on the "Template" icon ? I agree with you and I >>>> think >>>>>> sebb >>>>>>>>> had raised this when initially icon was a wizard. My concern is >>>> that >>>>>>>>> current icon may not be clear or "known" in other applications, so >>>> I >>>>>>>> wonder >>>>>>>>> if new user or even usual user have the idea of using it >>>>>>>> Perhaps see if we could use the Templates icon from OpenOffice? >>>>>>>> This looks like a piece of paper with a star in the top left corner. >>>>>>>> >>>>>>>> Here's a screenshot: >>>>>>>> >>>>>>>> http://people.apache.org/~sebb/JMeter/Templates.png >>>>>>>> >>>>>>> Nice one >>>>>>> >>>>>>>> I don't think we necessarily need a poll on this, so long as we all >>>>>>>> agree that the new icon is better. >>>>>>>> >>>>>>>>> 3) Make a poll (As sebb proposed in another thread) on the need to >>>>>> add a >>>>>>>>> user property for bugzilla 54778, to allow users to revert to < >>>> 2.12 >>>>>>>>> behaviour. My opinion based on previous versions behaviour is that >>>> we >>>>>>>>> should introduce this , asking community may be confusing >>>>>>>> If they are confused by the choice, won't they be confused by the >>>>>>>> change in behaviour? >>>>>>>> >>>>>>>> It will depend on how the information is presented. >>>>>>>> >>>>>>> What about this: >>>>>>> >>>>>>> When using Cache Manager in JMeter. If a resource is found in cache >>>> then: >>>>>>> - Up to version 2.11, no request hits the server and a 204 response >>>> code >>>>>>> with empty content was returned as a SampleResult >>>>>>> - Starting from 2.12, no request hits the server, but no more >>>>>> SampleResult >>>>>>> will be created (this is inline with what browser do) >>>>>>> >>>>>>> This change will affect: >>>>>>> - the number of Samples in results >>>>>>> - the average, median response time reported >>>>>>> >>>>>>> We propose 2 options to handle this: >>>>>>> (1) Add the possibility through a JMeter property to revert back to <= >>>>>> 2.11 >>>>>>> behaviour >>>>>>> (2) No JMeter property to revert back to <= 2.11 behaviour >>>>>>> >>>>>>> Please vote depending on what would be better for you. >>>>>>> >>>>>>> Or maybe we could use a doodle ? >>>>>> On reflection, I agree that providing a property to select the >>>>>> behaviour is the obvious thing to do. >>>>>> Adding the property does not need a vote. >>>>>> >>>>>> Ok >>>>>> However, I'm not sure we should change the default. >>>>>> There have been no external bug reports, so presumably users are not >>>>>> unhappy with the current behaviour. >>>>>> Or perhaps hardly anyone uses the Cache Manager. >>>>>> >>>>> No clearly there has been reports but not on mailing list. >>>> It would help to link to these on the Bug report. >>>> >>>>> As far as I remember I opened this bug after a twitter discussion: >>>>> https://twitter.com/Scooletz/status/317252989003915264 >>>>> >>>>> And see this: >>>>> >>>> http://stackoverflow.com/questions/17694151/jmeter-response-code-204-and-cant-get-content >>>>> https://flood.io/blog/18-understanding-the-jmeter-cache >>>>> >>>> http://mail-archives.apache.org/mod_mbox/jmeter-dev/201303.mbox/%3CCAH9fUpYXZ8+jXZP48L37yXA_vXnToQ4ghMHibG6kOpFOFe-Zag@mail.gmail.com%3E >>>> >>>> This suggests returning a sample with a different status, rather than >>>> no sample at all. >>>> >>>> You mean the last one ? >>> >>>>> So for me returning 204 is clearly a bug that must be fixed. >>>>> >>>>> >>>>>> Also, I wonder whether the fix for the bug report is the only possible >>>> one. >>>>>> It seems odd to drop the sample result entirely. >>>>>> >>>>>> >>>>> For me dropping sample is fine as in this case , a browser would lookup >>>>> cache and not hit server and not return 204. >>>>> 204 is a particular code that does not meet this case. >>>> Perhaps, but JMeter is not a browser, and anyway a browser would >>>> return the cache entry. >>>> >>> Yes but it would not return 204. 204 is a code that means server returned >>> an empty response , but server was hit so clearly wrong in this case. >>> >>>> Not returning a sample is not necessarily the correct thing to do. >>>> >>> Why ? for me it means no request occured which is what happens for the >>> browser and on client <=> server dialog (which jmeter mimics) >>> >>>> There is no unique HTTP status that corresponds to a local cache hit. >>>> >>> Yes because in this case , HTTP is not involved as no Request occurs >>> >>>> I assume that it was decided to use 204 as being the closest. >>>> >>>> JMeter is not a browser; it is primarily intended for testing servers. >>>> So I think what is important here is to generate results that can be >>>> analysed to show the appropriate performance figures. >>>> >>> Not returning any sample when cache is hit is the right option for me for >>> the reason I described above. >>> >>> >>> >>>> If you are only interested in the samples that hit the server, then I >>>> agree that cached responses can be ignored. >>>> But if one is interested in the client load that the server can >>>> support, it would be useful to be able to record cache hits. >>>> >>> Yes this could be useful. >>> >>> >>>> An alternative would be to return 200 (cache) for a local cache hit. >>>> Though this is not ideal given that JMeter does not currently cache the >>>> content. >>>> >>>> Maybe JMeter could return a currently unused status? >>>> >>>> Suppose JMeter were enhanced to cache the response content. >>>> >>> You mean https://issues.apache.org/bugzilla/show_bug.cgi?id=53540 ? >> Not really. >> That issue is a consequence of not caching the content, not actually >> an enhancement request to add such caching. >> Indeed one solution was to cache just the embedded resource URLs, >> which is quite different. >> >>>> What status should it return? How would one distinguish it from an >>>> actual server request? >>>> >>>> Could you give more details on this through an example ? >> Seems too obvious to need explaining, but here goes: >> >> Assume JMeter has been enhanced to cache the full content of a URL. >> Request a URL. >> Assume it is cached. >> Request the same URL, assume the cache entry has not expired. >> JMeter can then return the entire contents, exactly as it did originally. >> >> If it returns 200, how can one tell from the response whether it was >> served from cache or from the server? >> Note that a server request can have an elapsed time of zero if the >> server is faster than the clock granularity >> >> >>>>>>>>> Regards >>>>>>>>> Philippe >>>>>>>>> >>>>>>>>> >>>>>>>>> On Tue, Sep 2, 2014 at 8:34 AM, Milamber >>>>>>>> wrote: >>>>>>>>>> Le 31/08/2014 13:22, Philippe Mouawad a ecrit : >>>>>>>>>>> Hello, >>>>>>>>>>> I wonder if the current icon for Templates is the right one. >>>>>>>>>>> >>>>>>>>>>> The first icon which was a kind of wizard seems to me maybe a >>>>>> better >>>>>>>>>> option >>>>>>>>>>> even if it's not really a wizard. >>>>>>>>>>> The current icon is kind of "unknown" in the mind of users I >>>> think >>>>>>>> which >>>>>>>>>>> does not encourage its usage. >>>>>>>>>>> >>>>>>>>>>> Thoughts ? >>>>>>>>>> No problem for me for change icon, but the Wizard icon seems >>>>>> introduce >>>>>>>>>> some confusion to the users (the template box is just 1 box, not >>>>>> several >>>>>>>>>> steps to construct a test plan like a "wizard" in a software). >>>>>>>>>> Perhaps ask on user mailing list (make a poll to choice the best >>>>>> icon?) >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> Regards >>>>>>>>>>> Philippe >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On Sun, Jul 14, 2013 at 12:52 AM, Milamber >>>>>>> wrote: >>>>>>>>>>>> Le 13/07/2013 22:15, sebb a ecrit : >>>>>>>>>>>> >>>>>>>>>>>> On 13 July 2013 22:32, Milamber wrote: >>>>>>>>>>>>>> Le 13/07/2013 20:58, sebb a ecrit : >>>>>>>>>>>>>> >>>>>>>>>>>>>>> +1, but I think it should be after Open; >>>>>>>>>>>>>>> >>>>>>>>>>>>>> Sincerely, I prefer the template icon after New icon. It's >>>> more >>>>>>>>>>>>>> user-friendly in my mind. New=>New test plan from scratch, >>>>>>>>>> Template=>New >>>>>>>>>>>>>> test(or merge) to help end / beginner user, Open=>Open a test >>>>>> plan >>>>>>>>>> that I >>>>>>>>>>>>>> create (I'm pass the first steps to use JMeter). >>>>>>>>>>>>>> >>>>>>>>>>>>>> Unless there are plenty of templates for the user, they will >>>>>>>> likely >>>>>>>>>>>>> ignore the Template button very quickly.. >>>>>>>>>>>>> >>>>>>>>>>>> The Template behavior is a new functionality of JMeter, I can't >>>>>> know >>>>>>>> if >>>>>>>>>> we >>>>>>>>>>>> will have a plenty of templates (I hope this). >>>>>>>>>>>> I would like commit with "New|Template|Open" in toolbar. If in >>>>>>>> future, >>>>>>>>>> the >>>>>>>>>>>> users brings some misunderstand on this order/toolbar, we can >>>>>> change >>>>>>>>>> easily >>>>>>>>>>>> the order. >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> Milamber >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>>> and we need to drop the New >>>>>>>>>>>>>>> icon (Bug 55258) - so it will still be second. >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> On 13 July 2013 19:37, Philippe Mouawad < >>>>>>>> philippe.mouawad@gmail.com> >>>>>>>>>>>>>>> wrote: >>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> +1 for app office >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> On Saturday, July 13, 2013, Milamber wrote: >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> Le 13/07/2013 16:29, sebb a ecrit : >>>>>>>>>>>>>>>>> On 13 July 2013 17:05, Milamber >>>>>> wrote: >>>>>>>>>>>>>>>>>> Le 12/07/2013 14:10, sebb a ecrit : >>>>>>>>>>>>>>>>>>> On 12 July 2013 09:07, Milamber >>>>>>> wrote: >>>>>>>>>>>>>>>>>>>> Hello, >>>>>>>>>>>>>>>>>>>>> I thinks the new ability of Templates can have a >>>>>> dedicated >>>>>>>>>> button >>>>>>>>>>>>>>>>>>>>> in >>>>>>>>>>>>>>>>>>>>> Toolbar. >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> Agreed. >>>>>>>>>>>>>>>>>>>> I saw the Open Icon Library [1] to find a icon to >>>> this >>>>>>>> action >>>>>>>>>>>>>>>>>>>> (with >>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> license >>>>>>>>>>>>>>>>>>>>> CC-BT-SA 3.0). >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> CC-BY-SA >>>>>>>>>>>>>>>>>>>> I found two icons (the same but with two different >>>>>>>> colors), >>>>>>>>>>>>>>>>>>>> which >>>>>>>>>>>>>>>>>>>> is a >>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> magic >>>>>>>>>>>>>>>>>>>>> wand. In the sense of "magician". >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> Not sure the Template system is anything like a >>>> Wizard >>>>>> in >>>>>>>> the >>>>>>>>>>>>>>>>>>>> normal >>>>>>>>>>>>>>>>>>>> computing sense. >>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>> I think it's misleading to use a wand here. >>>>>>>>>>>>>>>>>>>> Basically it's a selection from a list. >>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>> I've update with other icon (CC-BY-SA) : >>>>>>>>>>>>>>>>>>> http://www.milamberspace.net/**toolbar-template/< >>>>>> http://www. >>>>>>>>>>>>>>>>>>> milamberspace.net/toolbar-template/> >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> The open icon library provide great icons, but it's very >>>>>>>>>> difficult >>>>>>>>>>>>>>>>>>> to >>>>>>>>>>>>>>>>>>> find a >>>>>>>>>>>>>>>>>>> good icon with the good license for the "templates" >>>>>> bahavior. >>>>>>>>>>>>>>>>>>> Yes, indeed. >>>>>>>>>>>>>>>>>> What about one of the following? >>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>> http://openiconlibrary.**sourceforge.net/gallery2/?./** >>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>> Icons/categories/applications-**office.png>>>>>>>>>>>>>>>>> openiconlibrary.sourceforge.net/gallery2/?./Icons/ >>>>>>>>>>>>>>>>>> categories/applications-office.png> >>>>>>>>>>>>>>>>>> http://openiconlibrary.**sourceforge.net/gallery2/?./** >>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>> Icons/actions/archive-insert-**3.png< >>>> http://openiconlibrary >>>>>> . >>>>>> sourceforge.net/gallery2/?./Icons/actions/archive-insert-3.png >>>>>>>>>>>>>>>>>> http://openiconlibrary.**sourceforge.net/gallery2/?./** >>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>> Icons/actions/document-import-**2.pngve>>>>>>>>>>>>>>>>> >>>> openiconlibrary.sourceforge.net/gallery2/?./Icons/actions/ >>>>>>>>>>>>>>>>>> document-import-2.pngve>update >>>>>>>>>>>>>>>>>> : >>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>> I've update the page: >>>>>>>>>>>>>>>>> http://www.milamberspace.net/**toolbar-template/< >>>> http://www. >>>>>>>>>>>>>>>>> milamberspace.net/toolbar-template/> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> The application-office icon seems very good ? +1 for me at >>>>>> this >>>>>>>>>>>>>>>>> position. >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> You can see the render here and 2 proposal of >>>> position in >>>>>>>> the >>>>>>>>>>>>>>>>>>> toolbar: >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> http://www.milamberspace.net/**toolbar-template/< >>>>>> http://www >>>>>>>> . >>>>>>>>>>>>>>>>>>>>> milamberspace.net/toolbar-template/> >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> I think any icon belongs either just before or just >>>>>> after >>>>>>>>>>>>>>>>>>>>> "Open", >>>>>>>>>>>>>>>>>>>>> i.e. >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>> New, Template, Open >>>>>>>>>>>>>>>>>>>> or >>>>>>>>>>>>>>>>>>>> New, Open, Template >>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>> It does not seem logical to have it between Save >>>> Selection >>>>>>>> and >>>>>>>>>> Cut. >>>>>>>>>>>>>>>>>>>> What is your opinion? need to find a new icon? the >>>> best >>>>>>>>>>>>>>>>>>>> position/color? >>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> Milamber >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> [1] >>>>>>>>>>>>>>>>>>>>> http://openiconlibrary.**sourceforge.net/gallery2/?./ >>>>>>>> >>>>>>>>>>>>>>>>>>>> /openiconlibrary.sourceforge.net/gallery2/?./> >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> -- >>>>>>>>>>>>>>>> Cordialement. >>>>>>>>>>>>>>>> Philippe Mouawad. >>>>>>>>>>>>>>>> >>>>>>>>>> >>>>>>>>> >>>>>>>>> -- >>>>>>>>> Cordialement. >>>>>>>>> Philippe Mouawad. >>>>>>> >>>>>>> >>>>>>> -- >>>>>>> Cordialement. >>>>>>> Philippe Mouawad. >>>>> >>>>> >>>>> -- >>>>> Cordialement. >>>>> Philippe Mouawad. >>> >>> >>> -- >>> Cordialement. >>> Philippe Mouawad. >