Reading the Goole 2010 IO pdf of : "Architecting for Performance with Google Web Toolkit", I found this approach to reduce startup time :
Better: 2 round trips:
First:
- Host Page
- Script URL
- Initial Data
Second:
- Compiled Script
It is not the Default GWT setup:
4 round trips
- Host page
- Selection Script
- Compiled Script
- Initial Data
My problem is that I do not how to do to acheive the "2 round trip solution". Could you help me ?
It's actually rather easy:
inline the selection script in your page (can be done with a special GWT Linker, or using a JSP include directive for instance –I've successfully been using the latter, much easier– and add a <meta name='moduleName::gwt:property' content='baseUrl=moduleName/'> to tell the inlined selection script where to find the compiled permutation.
Technically, you could go even farther and replace the selection script with server-side content negotiation. I believe Google does that (e.g. for Google Groups) but it's a bit harder (made way simpler with modularization of the CrossSiteIframeLinker, but still not that easy).
for the initial data, see https://developers.google.com/web-toolkit/articles/dynamic_host_page and http://jectbd.com/?p=1174 (I've been using AutoBeans to pass user information and application settings to the GWT app)
Related
I'm thinking about developing an Excel add-in as described here.
Would that be possible with Svelte - and do you know of any guides/help if yes?
I have looked through the video here, and I'm about worried about the usage of webpack.
Well... let's break it down
Is it possible?
Short answer: yes
Long answer: the documentation clearly states that Excel add-in still uses jQuery for logic manipulations. If your question was about Angular or react it would probably be a hard NO since those frameworks use an engine that should be included as part of solution. This kind of dependencies when dealing with plugins development are pretty hard to implement and maintain as a function of time so it's better to use very lightweight, non-core dependencies instead. Since you are asking about svelte - it is "compiled" into a bundle that contains pure code (based on your app logic). So - as long as your app rely on the load event sequence described in the docs - you are good to go.
Do you really need Webpack?
Short answer: no
Long answer: svelte can be deployed using rollup instead - which is more suitable for micro-applications (such as yours). So, if you feel that webpack (somehow) is blocking your work pipeline - just use svelete default configuration with rollup and you are ready to go
Your workflow
Create a very simple svelte app (my suggestion - try to take the example in the docs and implement it using svelte)
Test it locally (just verify it works)
Build it (you should ended up with 3 files - 1 html file in public directory and 2 other files in public/build directory - 1 js file and 1 css file (both called bundle)
Here's the tricky part - the html file does nothing - just loading the css and js files. In your case - you don't really need it
Change the bundle.css file to Home.css file
Change the bundle.js file to Home.js file and put your app inside the add-in main function
'use strict';
(function () {
Office.onReady(function() {
// Office is ready
YOUR SVELTE APP CODE SHOULD BE PLACED HERE
});
})();
Pack your add-in and test it
Technical notes
If Excel blocks the creation of new elements (using dynamic injection) - this approach will NOT work (since your entire app is generated by your js file)
Please refer to this article for more information about packing your app
Try to make your app as lightweight and small-size as possible just to avoid the risk of exceeding the limits allowed for add-ins
GWT is used and the application is deployed on WebLogic using HTTPS.
The performance is poor and with F12 Developer Tools, we could see that blank.html is downloaded multiple times. This is clearly related to GWT but we have not been able to figure out why.
The following is from javascript:
defineSeed(2613, 2614, makeCastMap([Q$BaseModelData, Q$ModelData, Q$Theme, Q$Serializable]), Slate_0);
var SLATE;
function $clinit_GXT(){
$clinit_GXT = nullMethod;
IMAGES = new XImages_generatedBundle_0;
MESSAGES = new XMessages__0;
SSL_SECURE_URL = getModuleBaseURL() + 'blank.html';
}
This is from GWT.java:
/**
* URL to a blank file used by GXT when in secure mode for iframe src to
* prevent the IE insecure content. Default value is 'blank.html'.
*/
public static String SSL_SECURE_URL = GWT.getModuleBaseURL() + "blank.html";
Does anyone know under what circumstances blank.html is called?
Thanks!
This is from GWT.java:
This is actually from GXT.java.
This is used in a few cases when creating an <iframe> element, so that IE won't give errors if your site is hosted from SSL. I can actually only find one case (as of GXT 3.1.1) which uses this, in Layer.java. Only IE pages loaded from https urls will make use of this.
The Layer class uses this as a "shim", a way to prop up some DOM elements above overs, and work around some browser bugs (typically plugin or iframe related). Menus and popup dialogs use this to ensure that they don't appear "underneath" content that they should be "above".
This file is very small - just enough HTML to convince IE than the iframe has correctly loaded, and no more. It never changes, and should load nearly instantly.
As far as performance goes, this should only happen when a Menu or Window/Dialog/Tooltip is shown - these shouldn't be happening on app startup usually, at least not more than a window or two. Additionally, the browser should recognize that it is loading the same element and cache it correctly, and not load it multiple times (though it might be listed several times as hitting the cache). If the server has instructed the browser to never cache the file, that is something you should look at changing.
In short, this is very unlikely to be the cause of any performance issues, at least in GXT itself. If somehow you have the shim enabled on every single widget in your project, this should not be required. If the file is loading slowly, something may be very wrong with your server configuration.
For reference, here is the entire file:
<html></html>
the problem is: my web application uses ZK, which automatically generates random UUID for each web element.
When I try to record some basic test-case with Selenium IDE, it automatically tries to use these randomly-generated ID's, without even giving me a good alternative.
Is there a way to forbid Selenium IDE to use IDs while locating elements?
Possible workaraounds:
Implement ID generator in ZK: I've thrown away this possibility, because the application GUI is too complex for this task, and ID should be unique for whole sesion, which make this workaraound really hard to implement, when you have same elements on different page.
Find another recording tool: I've only found XLT script developer, which does the work by writing DOM-path using classes (which zk gives plenty) - but sometimes the location strategy gives false path, which is then not reproducible. Any good alternatives here?
You can change locater builder by changing the order of the locater in options>locater builder.
For example if you want to give first preference to css: name drag it on the top so when you start recording it will first give the preference to css name
Hope this will help you
I have a D2006 app in which the DPR file has had numerous edits (yes, I know - you shouldn't mess with the DPR file) to accommodate such things as a splash screen, preventing a second instance of the app being started, handling of command line options that need to be processed before any forms are created, etc.
One day, I noticed that the auto-create forms list in the project options is empty - but the DPR file still has the code in there to create some of the forms.
If I try to restore all the forms that should be auto-created from the dialog, it complains Error - Call to Application->CreateForm is missing or incorrect and doesn't do anything.
How can I restore this connection - apart from rebuilding the DPR from scratch?
is it safe to manually add the CreateForm calls?
are there any documented rules as to what you can do in the DPR file?
I have a suspicion that try..except and if..else clauses in the DPR upset Delphi. Will moving as many functions as possible to a separate unit and calling them be helpful?
I haven't really seen any documented rules as to what you can do in the DPR file, because I guess there are no strict rules.
The problem begins when you create a "Forms" application. (No problems with console or non-GUI applications I've noticed).
The IDE will automatically change the DPR any time you add a new Form or a DataModule to it, by assuming you want to auto-create them.
This can mess up your DPR, if it has a lot of code/compiler-directives/if-blocks/try-catch blocks etc...
So I'll tell you what my rules are, and in a short line:
Keep it as simple as you can.
My DPR contains only a call to some init code and auto creates the main form only:
MyAppInit; // in AppInit unit
Application.Initialize;
Application.CreateForm(TMainForm, MainForm);
Application.Run;
However in the uses section I add (or keep what the IDE added) all the forms my application uses (and also application related units) - this is useful when I want to view->forms or view-units.
In-fact when I add a new form to the application, the first thing I do is go the DPR and remove the line:
Application.CreateForm(TMyNewForm, MyNewForm);
NOTE (EDIT): The IDE can be configured to NOT auto create forms (No Application.CreateForm entry will be created in the DPR). In older version of Delphi this option is under: Tools/Environment Options/Preferences -> Auto create forms. In newer versions: Tools/Options/VCL Designer/Module creation options -> Auto create forms & data modules.
At run-time, I create all my forms dynamically when I need them, and destroy them when they no longer needed. DataModules/Splash (etc...) are created on the MainForm.OnCreate event.
This method has worked for me nicely for the past few years maintaining a large scale DB application. This will probably not cover all cases, but it worked fine for my needs.
P.S: "Is it safe to manually add the CreateForm calls" - Yes. But think twice if you really need them to be auto-create by the application.
IMHO You don't really need that AutoCreate forms from Delphi, but maybe in the casual test project.
And the dpr is just another source code file, where you're meant to write code to make things happen (or prevent it to happen), so don't worry if you lost that sincronization, which IMHO is buggy if can't read your pascal code to work properly.
If you still want to create some forms from the DPR, add the Application.CreateForm or TMyForm.Create calls manually to the file, AFAIK there's no rules against doing it that way.
Since Delphi owns the .DPR, I put my startup logic into a separate unit for each project.
That works really well, only very few entries need to be in the .DPR.
Since that unit controls the Application.CreateForm logic too, the IDE has an empty list for that: I'm fine with that.
The only things left in the .DPR are:
the big uses that indicates which modules are part of your project
a call to Main in the startup logic unit
Doing some research with regards to application framework used on the SUT am going to LoadTest using Loadrunner I fond the application is developed using wickets.
I have generated script using web http/html protocol against a wicket
application and there was some calls recorded in the following
format,script is failing at this URL when i ran the script in VUGen.
http://somem/nnnweb/main/ ?
wicket:interface= :1:someSearchForm :someSearchForm :searchInfo: :IActi
vePageBehaviorListener :0:&wicket: ignoreIfNotActiv e=true&random=
0.038901654740178815",
I find out like when i generate the script which has just views(
viewing tabs) is working fine, but when i edit somefileds and submit
the script is generating the above calls ( http://xxx...) and failing
This guy has explained just the same issue here:
http://tech.groups.yahoo.com/group/LoadRunner/message/27295
I hope it is ok to refer other sites in Stackoverflow?
General question. Does loadrunner support testing of wicket application and is TruClient the best choice here? I actually got it to work with TruClient, but as I understand there are drawbacks with memory footprint using TruClient, but maybe it is time to move on to next generation protocol given my project has decided to use wicket framework?