We have chrome and FF extensions which works with a native messaging app. Recently we developed edge extension and a UWP app which works in the same way. Now the user has to install two separate native messaging apps if they want use any chrome/FF/Edge browsers.
My question is, is it possible for Chrome/FF extension talk to windows UWP app?
Unfortunately no. On Windows, both Chrome and Firefox use the registry to locate the native applications manifest, and Windows Store apps are forbidden from writing to the registry. (ref: Prepare to package an app (Desktop Bridge)
If something changes (Store apps gain the ability to alter the registry, or Chrome and Firefox introduce an alternative way to locate the manifest), then it might be possible. Though likely not via the UWP app directly. UWP apps appear to support standard input and output, but the way they are run prevents access to it. It might however be possible to create an intermediary Win32 app that can communicate with the UWP app via the AppService and the browser extension via stdio.
Afterthought: Enpass Password Manager (win32) was ported to the Windows Store reportedly because of API issues, and has a browser extension for Chrome and Firefox. Might be worth asking them how they pulled it off. I did some more digging, and figured it out. localhost loopback, a custom url scheme, web sockets, and browser verifications is how they are doing it. Not an ideal solution, but it seems to work.
Related
This feels like a shot in the dark but...
Should a single Native Messaging host be able to communicate with the same extension installed on multiple chrome profiles?
I'm working on an extension which is installed to both my personal and work profiles. But it seems that the Native Messaging host only sends messages to the most recently connected instance of the extension.
I don't believe this is addressed in the Native Messaging documentation and I've run out of search ideas, thanks in advance for any help!
When nativeMessaging API is used it starts an instance of native app each time a connection is created by the extension so such an instance can communicate with its "parent" extension only. Consequently, there should be no problem.
In case you want to use chrome.runtime.onConnectNative to do the reverse (to connect from a native app to an extension which will work even when Chrome is closed) see crbug.com/967262 for more info or create a new issue there asking for details. Judging by the bits I see this feature is available only on ChromeOS and it's even disabled by default.
I am currently using phantomjs to generate screenshots of web pages.
Due to various issues including web fonts, and video tag support, I would like to try switching to headless chrome.
Has anyone had success running headless chrome in an azure web app?
I cannot find a .net api, and assume I need to run node.js with puppeteer If I want a javascript api similar to phantomjs.
Right now most of the big libraries are made for NodeJS. It's possible to run Chrome headless without it, but you'll need a library for your .NET API to effectively communicate with Chrome. If none exist, then the protocol itself is documented here, and you could build one for .NET.
If you're struggling to get Chrome running on Azure I have a service that offers Docker images, which could be as simple as a few commands to have up and running (located here). Other than that there's a few open-source options out there, but they have differing issues and uses-cases.
I have a Chrome extension that uses the chrome.management API to get a list of installed Chrome extensions and apps. The problem starts because I also want to use the chrome.syncFileSystem API which only seems to be available to Chrome apps.
If I switch the extension to an app, I can no longer use the chrome.management API. I haven't been able to find an API to access the installed extensions from an app. Any tips?
I don't think you can, not without having both an extension and an app.
Google has a private API to do that, but for public API, they want apps to be as independent as possible from the browser.
Lets say i have an app and i want it to be executable standalone. I've found that you can do that with normal webpages, webapps, using https://github.com/rogerwang/node-webkit.
But if its chrome packaged app how do you do that.
I know it's possible as you can see in slides it uses webkit.
https://speakerdeck.com/u/zcbenz/p/node-webkit-app-runtime-based-on-chromium-and-node-dot-js
Theoretically it is possible but you will have to implement by hand all of the chrome apis you use in your chrome app
I have created an App for google chrome which just opens a web page. I am new in the topic
of Apps. I was thinking if it is possible to execute some linux command by using an App
in google chrome. For instance, an App which can open a terminal or open a program
installed in my machine like Gimp, Kate, Libre Office ...
From the instructions in the web page of google I saw that the only actions for an App
are limited to open a web address but I dont know if it is possible to extend the capabilities
of the Apps,
Regards.
Aren't apps sandboxed into the Google Chrome Process to ensure they can't affect other processes and for other security reasons. If so, you won't be able to execute programs/commands or view the User's Files unless you use some workaround such as Google's Native Client.
Google has locked down capabilities to stop malicious Web Apps from executing code and bringing malware and exploits.
You can write a NPAPI Plugin or a custom URI scheme associated to Unix terminal.