发表新帖
发表新帖

How to get selected text in chrome extension development?

后端 未结
关注
2 472
星月不相逢
星月不相逢 2021-01-05 20:45

I\'m developing a chrome extension which involves getting selected text of the current tab. Here is the html file that I use:
相关标签:
2条回答
  • 悲哀的现实
    2021-01-05 21:20

    As @Xan suggested, the method mentioned before (you can find it here) is overcomplicated. To get it to work, there are only two things to do:

    1. Change value to innerHTML in document.getElementById("output").value

    2. Add an activeTab permission in manifest.json file

    Here is the complete source code, three files in total.

    manifest.json

    {
    "manifest_version": 2,
    "name": "sample",
    "description": "A sample extension to get the selected text",
    "version": "1.0",
    "icons": { 
        "16": "img/icon16.png",
        "48": "img/icon48.png",
        "128": "img/icon128.png" 
    },
    "browser_action": {
        "default_popup": "popup.html"
    },
    "permissions": [
        "activeTab"
    ]
}

    popup.html

    <!DOCTYPE html>
<html lang="en">
    <head>
        <meta charset="UTF-8">
        <script src="popup.js"></script>
        <title></title>
    </head>
    <body>
        <div id="output"></div>
    </body>
</html>

    popup.js

    chrome.tabs.executeScript( {
    code: "window.getSelection().toString();"
}, function(selection) {
    document.getElementById("output").innerHTML = selection[0];
});
    0 讨论(0)
  • 我寻月下人不归
    2021-01-05 21:41

    Your original approach was correct (and probably taken from this question).

    It had 2 problems:

    1. Inline scripts (<script>...</script>) are not allowed; it's fixed by putting code in a separate file, say, popup.js.

    2. You need permission to access a page for content script injection; in your particular case, there's a specific permission, "activeTab", that does it in a transparent and painless way with no security warnings. When your extension is invoked (by clicking on the button), you are given access to the current tab.

    With those fixed, your "direct" approach works.

    By the way, to debug such problems in future, you need to inspect the popup page.

    As for your own answer, you are over-complicating things a lot.

    You don't need an Event page is this particular case; your popup can call executeScript and listen to messages. There are cases when you do need it, specifically when you can't guarantee that popup is open when content script sends messages; but here you can guarantee it.

    Supposing you need the event page, consider not using getBackgroundPage to call a method in it - it tightly couples code; you can instead send a message. sendResponse can be used to pass the results back, even asynchronously.

    Finally, your schema is confusing with respect to content.js. Only one copy of it executes, in a special context that is attached to the target page (and not in the extension context).

    0 讨论(0)
 看不清?
提交回复
热议问题