Extended Robot Framework libraries to make testing GUI, REST/SOAP API, Mobile, and Windows Desktop easier.
Home Page: https://accruent.github.io/robotframework-zoomba/
License: Apache License 2.0
![dependabot-preview[bot] avatar](https://avatars.githubusercontent.com/in/2141?v=4 "dependabot-preview[bot]")
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot]")
![mend-bolt-for-github[bot] avatar](https://avatars.githubusercontent.com/in/16809?v=4 "mend-bolt-for-github[bot]")
![sourcery-ai[bot] avatar](https://avatars.githubusercontent.com/in/48477?v=4 "sourcery-ai[bot]")
How do I use Basic Authentication with Zoomba.SOAPLibrary ?
With SeleniumLibrary deprecating 'Select Window':
https://robotframework.org/SeleniumLibrary/SeleniumLibrary.html#Select%20Window
We need to use 'Switch Window' now:
https://robotframework.org/SeleniumLibrary/SeleniumLibrary.html#Switch%20Window
I installed openCV4NodeJS on top of Zoomba robot-framework library, so I could get the image locator into use.
However, even the image gets located, and "clicks" are done to the correct place of the application (close button), they seem to have no effect as they seem to not come from LMB, instead some kind of touch emulation?
Basically what happens, is that the test "clicks on air", never hitting the located spot
How to pass download directory with options?
I tried the below code: But it did not work
${options}= Get Edge Options
${options.use_chromium}= Set Variable True
${options.add_experimental_option["download.default_directory"]} Set Variable C:\Downloads
Open Browser http://www.google.com browser=Edge options=${options}
Could you please help ?
Hi,
I would like to get some help to get faster test runs.
We a have a big (and slow) executable to test.
After logging in we have a lots of different options - windows and child windows (...and sub sub.. windows) and we had very slow test case running in some complex test cases. ( like over 20 minutes!)
Evolution in test speed:
Great speed improvements happened when we :
But in some cases - when sub-sub windows is the actual,- this actions took 20-60 seconds.
I think this is because we have too many objects in hour application.
So here is Lazy Loading concepts comes in picture:
One of our colleagues used WAD with other language, C# or Java, and he used the Lazy Loading "concept"
https://github.com/RAM0803/WinAppDriver/tree/master/Java/CalculatorTest
Some references related to Lazy Loading :
https://www.geeksforgeeks.org/lazy-loading-design-pattern/
https://pypi.org/project/lazy-load/
https://stackoverflow.com/questions/36274/what-is-lazy-loading
https://www.infoworld.com/article/2077568/java-tip-67--lazy-instantiation.html
Can we use this Lazy Loading here in Zoomba Desktop Library?
How I can use function Double Click?
When a keyword fails in the Desktop Library a screenshot is created. For some reason currently this is duplicated.
Using the 'Get Webelement' example from the Desktop Library documentation results in an error.
AttributeError: 'WebElement' object has no attribute 'startswith'
robotframework-zoomba version: 2.11.0
*** Settings ***
Library Zoomba.DesktopLibrary
Suite Setup Start App
*** Variables ***
${REMOTE_URL} http://127.0.0.1:4723
${APP} Microsoft.WindowsCalculator_8wekyb3d8bbwe!App
*** Keywords ***
Start App
Open Application ${REMOTE_URL} platformName=Windows deviceName=Windows app=${APP} alias=Main
*** Test Cases ***
Click Webelement
${element} Get Webelement name=One
Click Element ${element}
Improve error messaging and error handling in the Desktop Library. This is mainly focused around keywords that utilize ActionChains (such as the mouse keywords) but is needed in other keywords that needed modification and the private methods (perhaps).
Taking a look at https://bettercodehub.com/results/Accruent/robotframework-zoomba
The section 'Write Code Once' pointed out a few areas where desktop and mobile libraries have some duplicated code. This makes sense as they both extend from AppiumLibrary but we should still make these into helper methods to avoid this reuse.
Many of the wait keywords we have set up are just using the global timeout. This is fine but the user should be allowed a timeout override per keyword use if they want.
Brought up here: https://forum.robotframework.org/t/unable-to-run-multiple-testsuite-using-zoomba-desktop-library/1348/3
Essentially if we have two test files just like this in a directory:
*** Settings ***
Documentation Zoomba Desktop Library Tests.
Library Zoomba.DesktopLibrary
Suite Setup Start App
Test Setup Launch Application
Test Teardown Quit Application
Suite Teardown Driver Teardown
Force Tags Windows
*** Variables ***
${REMOTE_URL} http://127.0.0.1:4723
${APP} Microsoft.WindowsCalculator_8wekyb3d8bbwe!App
*** Keywords ***
Start App
[Documentation] Sets up the application for quick launching through 'Launch Application' and starts the winappdriver
Driver Setup
Open Application ${REMOTE_URL} platformName=Windows deviceName=Windows app=${APP}
Maximize Window
Quit Application
*** Test Cases ***
Switch To Desktop Test
Close Application
Switch Application Desktop
Wait For And Click Element name=Start
Wait For And Click Element name=Start
And we run the whole directory (so file1, then file 2) and the 2nd file attempts to run 'Switch Application' it is referencing the old desktop session:
Need to investigate why the session data is being reused after the applications have been quit and the driver has been torn down.
Remove the previously deprecated keywords in the DesktopLibrary.
testing 123
// int waitingRequestsCount = 0;
// Matcher matcher = pendingRequests.matcher(htmlContent);
// if (matcher.find()) {
// String queueCount = matcher.group(1).replaceAll("[^0-9]", "");;
// waitingRequestsCount = Integer.parseInt(queueCount);
// }Window 10
2.13.1
Hi,
We have a test script to Open and launch the application( desktop app )
The next file is to perform subsequent checks on the opened application.
The first file executes successfully ( means application is opened ), but when the control is moved to the second script "no open application " is thrown.
How can I resolve this ?
My setup includes ( DesktopLibrary - Zoomba )
Please update the version of robotframework-requests to the latest (0.8.2 at the moment)
Changes to Travis-ci have cut off our ability to test. Emails to support have not been answered to get an OSS credit allotment so we need to transition to Github Actions.
After adding the robot tests to the travis-ci build it was revealed that a few of the APILibrary Robot tests were not functioning as intended.
Validate Response Negative Tests - Broken
Validate Response List Negative Tests - Broken
Key By Key Validator Negative Tests
These tests are all producing the correct error messages. The issue is that the compare done by 'Run Keyword And Expect Error' is saying that the error doesn't match the intended.
From a quick glance and test I cannot figure out the difference between the actual and expected.
Each has been tagged 'Broken' and some have been moved into their own test cases since they were part of a larger set. When they are fixed they should be untagged and moved back to the test case they came from (same test case name but without the '- Broken' part).
While trying to get the Desktop Library up and running, I ran into an issue with example remote_url in https://accruent.github.io/robotframework-zoomba/DesktopLibraryDocumentation.html
Most remote_url example is given as http://localhost:4723/wd/hub, but that didn't work for me. I had to use 'http://localhost:4723`.
TestZoomba.robot that worked for me:
*** Settings ***
Library Zoomba.DesktopLibrary run_on_failure=No Operation
*** Test Cases ***
Test Zoomba Desktop Library
[Tags] Zoomba
[Setup] Driver Setup
Log to console ...
Switch Application By Name http://localhost:4723 Untitled - Notepad
[Teardown] Driver Teardown
RunTestZoomba.bat:
CD /D "%~dp0"
robot -d Output/Zoomba -T ^
--include Zoomba ^
-s Suites.TestZoomba .
Does the documentation need to be updated?
Is my setup wrong?
C:\>pip show robotframework-zoomba
Name: robotframework-zoomba
Version: 2.13.1
Summary: Robot Framework mini-framework.
Home-page: https://github.com/Accruent/zoomba
Author:
Author-email:
License: GPL-3.0
Location: c:\program files\python37\lib\site-packages
Requires: msedge-selenium-tools, psutil, python-dateutil, requests, robotframework, robotframework-appiumlibrary, robotframework-requests, robotframework-seleniumlibrary, robotframework-sudslibrary-aljcalandra, selenium
Required-by:
C:\>python
Python 3.7.9 (tags/v3.7.9:13c94747c7, Aug 17 2020, 18:58:18) [MSC v.1900 64 bit (AMD64)] on win32
WinAppDriver installed at: C:\Program Files (x86)\Windows Application Driver\WinAppDriver.exe
Windows 10
In attempting to make a Type check here we are not properly comparing type, but instead the Type object.
We can see this issue using the following code:
def check(obj):
if type(obj) is unicode or str:
print "Good"
else:
print type(obj)
obj = {'one': 'value'}
check(obj)
Hi. Please Kindly tell me how to get value from elemnet in window app to variable for calculate.
could you pls update dependency to psutil 5.8.0? at least python 3.9 compatible wheels are available now.
Error from Robot test
JavascriptException: Message: javascript error: $ is not a function
This keyword was implemented as a workaround for this issue:
SeleniumHQ/selenium-google-code-issue-archive#3604
It has been some time since then so perhaps this method is no longer necessary and can be removed? If it indeed is not needed lets remove it otherwise fix the current issue.
Hi @Wolfe1,
We have our automated tests using zoomba for GUI and Android.
We felt occasionally to write certain functions in Python instead of robotframework. The problem we see is that we do not know how to extend it using robotframework-zoomba.
For e.g., we can write the following and then call the keyword where required, but the issue would be that this does not know the current browser that is launched and so the error message would be No browser open. Could you help us with how to let the keyword know about the browser instance?
from Zoomba.GUILibrary import GUILibrary
web_browser = GUILibrary()
wait_time = 30
class LoginPage:
login_email_textbox = "id=email"
def login_with_enter_email_address(self, email_address):
web_browser.wait_for_and_input_text(self.login_email_textbox, email_address, timeout=wait_time)
We need a way to create a more uniquely named screenshot. To do so we have implemented the following in Robot framework:
Save Selenium Screenshot
${screenshot_index}= Get Variable Value ${screenshot_index} ${0}
Set Global Variable ${screenshot_index} ${screenshot_index.__add__(1)}
${time}= Evaluate str(time.time()) time
Capture Page Screenshot selenium-screenshot-${time}-${screenshot_index}.png
We need to implement this in the library so that we don't need to copy this code to every repository we use.
Hi,
First of all this Zoomba library is amazing. thank you very much for this. :)
My task to test a windows Desktop application, and i stucked with two issues. (i will open a separate issue)
My issue is with slow running when starting Switch Application Desktop.
If our app started only with Open Application,
Login page opens and I was able to type username and pw. then click on login button. (this is fast)
After login was success, application loading and main form appears. (login disappears)
On this MAIN form i was not able to interact any of the elements!
I was hopeless for one week and searching the internet for help then i found an advise to open app=root.
and then i found Zoomba. :)
So I started using the Switch To Desktop Test example
Our app loaded after login and I was able to Click, Input, Send keys, etc.
But each command issue takes 5-15 seconds (robot framework waits to issue this to WAD driver.)
So this is very very long, almost unusable because lots of interaction needed for each test case.
Can you pls advise how to solve the slowness issue?
Calculator example is fast.
CPU usage less than 20% (mostly winappdriver ) with my app. Memory is 50%. So these should not be the reason.
test case
Do you have any advise?
Thanks in advance.
Best Regards,
Robert
Hi Team,
I have an application that is developed using Webview2 .net component. Basically, we have a desktop container for our Angular based app. This is run using the microsoft edge (chromium).
I tried using Zoomba library for writing to a text field but the test case would fail giving an exception
No such elementException.
Any help on this would be appreciated
I have written very basic test case as shown below. The app opens but the element is not recognized.
Open Application ${RemoteURL} platform=Windows deviceName=Windows app=${myapp} window_name=${mywindowname} exact_match=false
Input Text xpath=//[@id="myTextBoxID"] MysampleText
This keyword is supposed to check for the window title (up to the timeout) and report back that it is open.
Currently it only attempts this once and relies on other keywords to keep it in check. This is misleading and not how this keyword should function.
This issue was brought to my attention via an email that it would be useful to be able to start and stopp the winappdriver session at will via some keywords. This would probably be done much like in WADLibrary's Driver.py.
Probably would default the winappdriver location to the default install.
For an actual and expected dictionary the following may happen:
{ "a" : [1] } - Actual
{ "a" : [1] } - Expected
This will break when it sends 1, 1 to the recursive call on line 286.
Changing the 1 to a string resolves this problem.
The primary issue is that instead of checking for equality we are checking if string, then checking for inequality.
After that we assume the inner object is a dictionary.
A Set or a Tuple would also break this.
Most of this is exceptions since this method is dependent on robot framework which deals nearly only with strings, dicts, and lists, but it is not impossible to encounter ints or floats or other objects with regard to custom robot keywords in other libraries.
Running my RF scripts with docker on internal servers I get lots of warnings:
/usr/local/lib/python3.8/dist-packages/urllib3/connectionpool.py:1013: InsecureRequestWarning: Unverified HTTPS request is being made to host 'proxy'. Adding certificate verification is strongly advised. See: https://urllib3.readthedocs.io/en/1.26.x/advanced-usage.html#ssl-warnings
In requests library, keyword Create Session there is a possibility to use the disable_warnings argument, see here.
There is no Create Session keyword in Zoomba.APILibrary.
And the Call Post Request has no disable_warnings argument.
So is there another way to suppress warnings?
Add information about the DesktopLibrary and how to run it into the readme.
Currently I need to automate an excel spreadsheet that contain macros, therefore i need to automate using the excel application. But I am unable to extract the text from a cell, I can find the text in attribute ValuePattern.Value.
Does anyone knows how to extract this data?
This has been asked for twice in the last few days:
https://forum.robotframework.org/t/zoomba-windows-desktop-regular-expression-on-window-name/908/2
https://forum.robotframework.org/t/how-to-read-accessibility-tool-inspected-elements-using-zoomba-library/903/2
Something like:
window_name = “wel*” [partial tile] instead of “welcome 10.3.2.4.2”
Zoomba has a big delay to process a act. I can't fix this problem. Can you help me?
I use locator to click on button or check if the button is enabled. But, these process it takes too long.
Hi,
I figured out how to get Checkbox value/status, it takes a few hours, so i would like to share this with others :)
How to add it to Wikipage, or to zoomba/DesktopLibraryDocumentation ?
Attribute to evaluate should be Toggle.ToggleState (and not Value.Value as in your example)
${CBV}= Zoomba.DesktopLibrary.Get Element Attribute AutomationId_ofthechkbox Toggle.ToggleState
Zoomba.DesktopLibrary.Element Attribute Should Match AutomationId_ofthechkbox Toggle.ToggleState 1
Best Regards,
Robert
Hi,
Buttons and CheckBoxes needs to be clicking twice for working. (Manual tests needs to be pushed once)
The code in Robot Framework:
POST /session/5166D5D3-27CE-41B7-BC31-9722D6388F42/elements HTTP/1.1
Accept: application/json
Accept-Encoding: identity
Connection: keep-alive
Content-Length: 115
Content-Type: application/json;charset=UTF-8
Host: 127.0.0.1:4723
User-Agent: appium/python 1.0.2 (selenium/3.141.0 (python windows))
{"using": "xpath", "value": "//Button[@AutomationId='btnOK']", "sessionId": "5166D5D3-27CE-41B7-BC31-9722D6388F42"}
HTTP/1.1 200 OK
Content-Length: 98
Content-Type: application/json
{"sessionId":"5166D5D3-27CE-41B7-BC31-9722D6388F42","status":0,"value":[{"ELEMENT":"42.5505976"}]}
==========================================
POST /session/5166D5D3-27CE-41B7-BC31-9722D6388F42/element HTTP/1.1
Accept: application/json
Accept-Encoding: identity
Connection: keep-alive
Content-Length: 115
Content-Type: application/json;charset=UTF-8
Host: 127.0.0.1:4723
User-Agent: appium/python 1.0.2 (selenium/3.141.0 (python windows))
{"using": "xpath", "value": "//Button[@AutomationId='btnOK']", "sessionId": "5166D5D3-27CE-41B7-BC31-9722D6388F42"}
HTTP/1.1 200 OK
Content-Length: 96
Content-Type: application/json
{"sessionId":"5166D5D3-27CE-41B7-BC31-9722D6388F42","status":0,"value":{"ELEMENT":"42.5505976"}}
==========================================
POST /session/5166D5D3-27CE-41B7-BC31-9722D6388F42/element/42.5505976/click HTTP/1.1
Accept: application/json
Accept-Encoding: identity
Connection: keep-alive
Content-Length: 73
Content-Type: application/json;charset=UTF-8
Host: 127.0.0.1:4723
User-Agent: appium/python 1.0.2 (selenium/3.141.0 (python windows))
{"id": "42.5505976", "sessionId": "5166D5D3-27CE-41B7-BC31-9722D6388F42"}
HTTP/1.1 200 OK
Content-Length: 63
Content-Type: application/json
==========================================
POST /session/5166D5D3-27CE-41B7-BC31-9722D6388F42/element HTTP/1.1
Accept: application/json
Accept-Encoding: identity
Connection: keep-alive
Content-Length: 115
Content-Type: application/json;charset=UTF-8
Host: 127.0.0.1:4723
User-Agent: appium/python 1.0.2 (selenium/3.141.0 (python windows))
{"using": "xpath", "value": "//Button[@AutomationId='btnOK']", "sessionId": "5166D5D3-27CE-41B7-BC31-9722D6388F42"}
HTTP/1.1 200 OK
Content-Length: 96
Content-Type: application/json
{"sessionId":"5166D5D3-27CE-41B7-BC31-9722D6388F42","status":0,"value":{"ELEMENT":"42.5505976"}}
==========================================
POST /session/5166D5D3-27CE-41B7-BC31-9722D6388F42/element/42.5505976/click HTTP/1.1
Accept: application/json
Accept-Encoding: identity
Connection: keep-alive
Content-Length: 73
Content-Type: application/json;charset=UTF-8
Host: 127.0.0.1:4723
User-Agent: appium/python 1.0.2 (selenium/3.141.0 (python windows))
{"id": "42.5505976", "sessionId": "5166D5D3-27CE-41B7-BC31-9722D6388F42"}
HTTP/1.1 200 OK
Content-Length: 63
Content-Type: application/json
{"sessionId":"5166D5D3-27CE-41B7-BC31-9722D6388F42","status":0}`
Any advise why we have this ?
Zoomba Version 2.8.0
If keyword "Switch Application By Name" is used in combination with exact_match=false it is much slower than using it with exact_match=true (by factor 2-4 if no other application is running but by factor 10-15 if some other applications are running).
You can test it by the follwing suite:
*** Settings ***
Library Zoomba.DesktopLibrary
Suite Setup Suite Initialization
Suite Teardown Suite Finalization
*** Variables ***
${REMOTE_URL} http://localhost:4723
${APP} Microsoft.WindowsCalculator_8wekyb3d8bbwe!App
*** Test Cases ***
AppChange
# ${calculator_handle}= Open Application ${REMOTE_URL} platformName=Windows deviceName=Windows app=${APP}
${calculator_handle}= Switch Application By Name ${REMOTE_URL} window_name=Rechner exact_match=true platformName=Windows deviceName=Windows app=${APP}
Wait For And Click Element accessibility_id=clearEntryButton
Wait For And Click Element accessibility_id=num1Button
Wait Until Element Contains accessibility_id=CalculatorResults 1
# Switching to Word but don't know the name of the currently open word document (means I have to use exact_match=false)
Switch Application By Name ${REMOTE_URL} window_name=Word exact_match=false platformName=Windows deviceName=Windows
Send Keys Hello World 1 \ue007
Switch Application ${calculator_handle}
Wait For And Click Element accessibility_id=num2Button
Wait Until Element Contains accessibility_id=CalculatorResults 2
# Switching to Word but using the exact document name
Switch Application By Name ${REMOTE_URL} window_name=Dokument1 - Word exact_match=true platformName=Windows deviceName=Windows
Send Keys Hello World 2 \ue007
Switch Application By Name ${REMOTE_URL} window_name=Rech exact_match=false
Wait For And Click Element accessibility_id=num3Button
Wait Until Element Contains accessibility_id=CalculatorResults 3
# Switching to Word againg using exact_match=false to see whether there is some caching
Switch Application By Name ${REMOTE_URL} window_name=Word exact_match=false
Send Keys Hello World 3 \ue007
*** Keywords ***
Suite Initialization
Driver Setup
Suite Finalization
Driver Teardown
This results in the follwing log
What can be done to improve the performance?
Hello,
Greeting!!!
When i try to open a app, i'm getting "Webdriverexception: Failed to locator opened application window with appId" error and application is not getting launched.
I have the below start driver keyword in my suite setup,
Start Driver
driver setup
OPEN APPLICATION ${REMOTE_URL} platformName=Windows deviceName=Windows app=${APP}
REMOTE_URL = 'http://127.0.0.1:4723'
I have my in local storage (Plug and play app. not installable).
It was working fine couple of month before. I'm not sure what's going on.
Could you please help?
Hi,
is there any plan to add touch support in Desktop Library?
Thanks
'Press Key' is deprecated, replace with 'Press Keys'
Hi everyone,
I was trying to automate a windows app where button1 is visible from the start, but disabled. Button2, when pushed, starts an operation at the end of which button1 becomes enabled and thus clickable. The keyword Wait Until Element Is Enabled looked perfect for the job, but it kept failing. Looking at the Python code behind the library, the reason appeared clear:
self._wait_until_page_contains_element(locator, timeout, error)
self.element_should_be_enabled(self.current_element)
The keyword (like its sister Wait Until Element Is Disabled) doesn't actually wait for the element to switch to Enabled, rather it waits until the "page" contains the element and then summons the test-failing Should Be Enabled. So, in my case, the disabled button being already present from the start, Wait Until Page Contains Element passes, then Should Be Enabled instantly fails the test.
The keyword does not really do what it says it does, i.e. it doesn't wait for the button to become Enabled, it waits for it to be present and then fails if it's not enabled. I think this needs to be fixed, because if they worked as they should, they would be truly useful indeed.
Not sure when this appeared but its seems that Appium (1.20.2 in my tests) is no longer functioning when it comes to mouse and pointer inputs:
Cool Idea Here
Just do it
Hi ,
We have implemented Zoomba library to automate windows desktop application , is there a way we can extend this support to mac ?
Hi,
Since DesktopLibrary overrides the implementation of _element_find from AppiumLibrary and returns only a single WebElement, it does not work with the “Xpath Should Match X Times” keyword from AppiumLibrary which expects _element_find to return a collection so that it can apply the len() function on it.
DesktopLibrary#open_application does not start the app with webdriver#remote when the attribute window_name is given.
The result is that the application may be started with an other user than webdriver is running with and as it with other user rights than intended.
robotframework-zoomba/src/Zoomba/DesktopLibrary.py
Lines 280 to 299 in 69012dd
We need to provide more information for users first landing on the repository. Right now we just rely on user knowledge, the docs, and the example tests. However for a library aimed at helping not only existing teams, but new teams get set up we need a more comprehensive guide to importing and using the libraries.
Hi
I have a usecase where I need to grab the auth token from webUI and then Pass it on to API test-cases as header.
I have implemented and get the auth-token from Keyword (Get AuthToken). But in Zoomba.APILibrary we set the headers in Variables section as
*** Variables ***
${endpoint}= "some url"
&{headers_dictionary}= Authorization=Bearer ${token} Content-Type=application/json charset=UTF-8
${query_string}= none
In the test case file I am using Set Variable to update the ${token} but it's not getting updated.
Get Basic API Test
${token1}= Run Keyword Get AuthToken #keyword to get auth-token and assign it
${token}= Set Variable ${token1}
Log ${token}
${response}= Call Get Request ${headers_dictionary} ${endpoint} ${query_string}
Log ${response}
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google ❤️ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.