前幾天有同學問到selenium是怎麽樣啟動瀏覽器的(selenium啟動瀏覽器的原理)，當時稍微講解了一下，不過自我感覺不夠具體，現在特地把啟動原理通過代碼和一系列操作給串聯起來，希望可以幫助大家更好的理解。

以chrome瀏覽器為例，selenium啟動chrome瀏覽器的代碼如下：

源碼

def __init__(self, executable_path="chromedriver", port=0,
                 options=None, service_args=None,
                 desired_capabilities=
 
None, service_log_path=None,
                 chrome_options=None):
        """
        Creates a new instance of the chrome driver.
        Starts the service and then creates new instance of chrome driver.
        :Args:
         - executable_path - path to the executable. If the default is used it assumes the executable is in the $PATH
 

         - port - port you would like the service to run, if left as 0, a free port will be found.
         - desired_capabilities: Dictionary object with non-browser specific
           capabilities only, such as "proxy" or "loggingPref".
         - options: this takes an instance of ChromeOptions
 

        """
        if chrome_options:
            warnings.warn('use options instead of chrome_options', DeprecationWarning)
            options = chrome_options

        if options is None:
            # desired_capabilities stays as passed in
            if desired_capabilities is None:
                desired_capabilities = self.create_options().to_capabilities()
        else:
            if desired_capabilities is None:
                desired_capabilities = options.to_capabilities()
            else:
                desired_capabilities.update(options.to_capabilities())

        self.service = Service(
            executable_path,
            port=port,
            service_args=service_args,
            log_path=service_log_path)
        self.service.start()

        try:
            RemoteWebDriver.__init__(
                self,
                command_executor=ChromeRemoteConnection(
                    remote_server_addr=self.service.service_url),
                desired_capabilities=desired_capabilities)
        except Exception:
            self.quit()
            raise
        self._is_remote = False

其中跟瀏覽器啟動密切相關的是這幾句

self.service = Service(
    executable_path,
    port=port,
    service_args=service_args,
    log_path=service_log_path)
self.service.start()

通過查看跟Service相關的代碼復盤得到啟動邏輯: 調用chromedriver可執行文件運行chromedirver。這也是為什麽我們需要把chromedriver放到系統PATH裏的原因。

所以selenium先啟動了chromedriver。當然，我們可以手工啟動chromedriver來模擬這個啟動過程。

在命令行中運行下面的命令chromedirver

你應該可以看來類似的結果

Starting ChromeDriver 2.38.552518 (183d19265345f54ce39cbb94cf81ba5f15905011) on port 9515
Only local connections are allowed.

這樣我們就手工啟動了chromedriver。driver監聽的端口是9515.

啟動了driver之後，我們需要告訴driver打開瀏覽器。selenium的源碼裏這一過程如下

def start_session(self, capabilities, browser_profile=None):
        """
        Creates a new session with the desired capabilities.
        :Args:
         - browser_name - The name of the browser to request.
         - version - Which browser version to request.
         - platform - Which platform to request the browser on.
         - javascript_enabled - Whether the new session should support JavaScript.
         - browser_profile - A selenium.webdriver.firefox.firefox_profile.FirefoxProfile object. Only used if Firefox is requested.
        """
        if not isinstance(capabilities, dict):
            raise InvalidArgumentException("Capabilities must be a dictionary")
        if browser_profile:
            if "moz:firefoxOptions" in capabilities:
                capabilities["moz:firefoxOptions"]["profile"] = browser_profile.encoded
            else:
                capabilities.update({'firefox_profile': browser_profile.encoded})
        w3c_caps = _make_w3c_caps(capabilities)
        parameters = {"capabilities": w3c_caps,
                      "desiredCapabilities": capabilities}
        response = self.execute(Command.NEW_SESSION, parameters)
        if 'sessionId' not in response:
            response = response['value']
        self.session_id = response['sessionId']
        self.capabilities = response.get('value')

        # if capabilities is none we are probably speaking to
        # a W3C endpoint
        if self.capabilities is None:
            self.capabilities = response.get('capabilities')

        # Double check to see if we have a W3C Compliant browser
        self.w3c = response.get('status') is None
        self.command_executor.w3c = self.w3c

這一過程的核心就是就是向localhost:9515/session發送1個POST請求，並發送1個json對象，默認情況下，這個對象應該是下面這個樣子。

{
    "capabilities": {
        "alwaysMatch": {
            "browserName": "chrome",
            "goog:chromeOptions": {
                "args": [],
                "extensions": []
            },
            "platformName": "any"
        },
        "firstMatch": [
            {}
        ]
    },
    "desiredCapabilities": {
        "browserName": "chrome",
        "goog:chromeOptions": {
            "args": [],
            "extensions": []
        },
        "platform": "ANY",
        "version": ""
    }
}

簡單理解就是告訴remote driver打開什麽瀏覽器，上面的例子裏我們打開的是chrome瀏覽器。

我們可以手工還原這個過程。

確保chromedriver是在運行中的，然後打開postman，構造1個POST請求，路徑是localhost:9515/session。在Body裏選擇raw和JSON(application/json), 貼入上面的json字符串，如下圖所示

點擊send，幾秒之後chrome瀏覽器應該可以正常啟動，並且postman的response裏會有大致如下的返回值

{
    "sessionId": "ad4407e133cfd5f3f49bff4c2f1f087a",
    "status": 0,
    "value": {
        "acceptInsecureCerts": false,
        "acceptSslCerts": false,
        "applicationCacheEnabled": false,
        "browserConnectionEnabled": false,
        "browserName": "chrome",
        "chrome": {
            "chromedriverVersion": "2.38.552518 (183d19265345f54ce39cbb94cf81ba5f15905011)",
            "userDataDir": "/var/folders/s6/f2_brc114wv2g8w0qggk_m2c0000gn/T/.org.chromium.Chromium.NMsAKJ"
        },
        "cssSelectorsEnabled": true,
        "databaseEnabled": false,
        "handlesAlerts": true,
        "hasTouchScreen": false,
        "javascriptEnabled": true,
        "locationContextEnabled": true,
        "mobileEmulationEnabled": false,
        "nativeEvents": true,
        "networkConnectionEnabled": false,
        "pageLoadStrategy": "normal",
        "platform": "Mac OS X",
        "rotatable": false,
        "setWindowRect": true,
        "takesHeapSnapshot": true,
        "takesScreenshot": true,
        "unexpectedAlertBehaviour": "",
        "version": "66.0.3359.181",
        "webStorageEnabled": true
    }
}

上面的返回裏最重要的就是sessionId，因為後面所有跟瀏覽器的交互都是基於該id進行的。

總結

• selenium裏，selenium client先打開chromedriver
• chromedirver創建session時打開了瀏覽器，所以瀏覽器的打開跟selenium無關，完全是chromedriver的能力

更多

其實上面的例子裏我們手工調用了webdriver協議裏的new session協議，創建了webdriver session。具體的細節大家可以參考協議了解更多。

selenium是如何啟動瀏覽器的