I'm creating an Firefox extension which is required to load a list of regular expression from a certain URL. But, if this called every time user load a page, it will be too painful for the server. So, I want the program to run once every time a user is starting Firefox.
I've found this solution:
How to detect first loading of NPAPI plugin/Extension in firefox and chrome
But it has a downside. If the Firefox starts with multiple tabs, it will not work (gBrowser.browsers.length don't have "1" as value).
So, what should I do?
Where did you read that extension code is executed every time a user loads a page? By default, extension code on Firefox only runs on
install
enable
startup
upgrade
downgrade
So, just make the request in main.js and you should be fine.
If you really want the list to only load on startup (which doesn't sound like what you would actually want), then add the line
if (require("sdk/self").loadReason==='startup') loadRegexs()
Related
While running VS Code Live Server (Go Live), the browser is crashing with the below mentioned error. I am using Google Chrome as default. I re-installed VS Code after uninstalling completely (from cache even), restarted browser, and then restarted PC as well with no use. When I checked "Use Local Ip" in LiveServer>Settings:, server is taking too long to respond and eventually fails to load the page (Error: "This site can't be reached"). Another important thing is that Chrome is opening every other website properly. This problem exists only when opening VS Code Live Server. Below is the error code:
Chrome Browser Error:
"Aw, Snap!
Something went wrong while displaying this webpage.
Error code: STATUS_BREAKPOINT"
Another popup is also showing up simultaneously:
"DevTools was disconnected from the page.
Once page is reloaded, DevTools will automatically reconnect."
(This didn't seem like a bug in the code of script.js)
Found the Solution:
Indeed it is a bug in the code of script.js.
If you are having exactly the same problem as I described, let us first see what we don't need to do, as I found these suggestions everywhere, and none will work when both VS Code and your browser are perfectly fine (they only appear to be not working):
No need to restart or reinstall VS Code.
No need to restart or reinstall browser.
No need to restart the PC
No need to disable browser extensions or VS Code extensions.
No need to change LiveServer Custom Browser; you can keep it
default.
No need to change LiveServer>Settings>Use local ip or Use browser
preview (you can keep them both unchecked if they are already
unchecked by default)
However, if you use firefox, console shows the error: "Uncaught out of memory"
Here, I got the first hint. Firefox at least told me to look at memory, but it is not about the memory we think. It is not about multiple tabs opened, lack of disk space, unstable disk, etc. It is the accidentally generated infinite loop in the code which is causing the memory problem (check your code carefully to find it).
In my case, I accidentally didn't tell it to execute the next iteration (i.e., missing counter update, increment, i++) and it is stuck in the same while loop forever, as the condition is satisfied forever without any counter update.
After fixing the code, the below error might show up in the console. Just refresh the page and it will go away.
"favicon.ico:1 Failed to load resource: the server responded with a status of 404 (Not Found)"
Note: Unless you fix the code or comment it out, other linked files (like .html) in the folder wouldn't work either.
The Chrome dev team apparently just rolled out a new "feature" called SuppressDifferentOriginSubframeJSDialogs, which makes it so alert+confirm boxes from an iFrame on a different domain than the parent does not show up, by default: https://www.chromestatus.com/feature/5148698084376576
This means if you have an embedded form, and you relied upon the standard JavaScript alert to inform the user of errors (or a confirm to ensure they want to make a change), it now does nothing, completely destroying the UX.
Obviously we can overwrite the built-in alert and confirms with custom ones, but moving away from what is built into the browser now introduces re-designs for every UI (mobile, tablet, laptop, desktop, etc.) instead of just relying upon proven, built-in technology.
Is there a way to change this cross-domain setting, or to whitelist specific domains to override this feature? We do have access to the parent site with an external JS include file (obviously, otherwise this would be a wide-open back door).
To Disable SuppressDifferentOriginSubframeJSDialogs , which block js Dialog box pop up
If you are single users ( Windows ) :
Right click on desktop and create a shortcut
Windows will pop up a screen and ask you "type the location of the item:"....now paste those value into the textbox
"C:\Program Files\Google\Chrome\Application\chrome.exe" --disable-features=SuppressDifferentOriginSubframeJSDialogs
Click "next" then "finish" button to create the shortcut
Now you shall see a chrome shortcut created in your desktop .Close all active Chrome instance ( if exists ) then launch the shortcut and you will found Js pop up wont block anymore.
Tips :
If you want push this fix to all your cooperate domain users , you may consider use Domain group policy to update Chrome shortcut value or use custom domain logon script logic to do so
I found a setting in the latest Chrome GPO template that creates a reg entry that will deal with this and the same reg entry placed in the Edge location also resolves the error there. (at least it worked for the problem that came up for us yesterday FWIW)
[HKEY_LOCAL_MACHINE\SOFTWARE\Policies\Microsoft\Edge]
"SuppressDifferentOriginSubframeDialogs"=dword:00000000
[HKEY_LOCAL_MACHINE\SOFTWARE\Policies\Google\Chrome]
"SuppressDifferentOriginSubframeDialogs"=dword:00000000
The latest Edge GPO templates do not have the setting yet, but I imagine it's only a matter of time.
Today I tried it again with Google Chrome Version 92.0.4515.131 and it suddenly worked again. Can anyone confirm this?
Add for Mac users.
If you use mac OS:
Quit all running instances of Chrome.
Run the terminal application.
In the terminal, run the command:
/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome --disable-features="SuppressDifferentOriginSubframeJSDialogs"
You can see more details here :
https://trailblazer.salesforce.com/issues_view?title=salesforce-functionality-impacted-in-[%E2%80%A6]oss-origin-iframe-javascript-dialogs&Id=a1p4V000002BRMX
Once a node.js program has run to completion in the context of an --inspect session (i.e. via the Chrome dev tools debugger) is it possible to re-start it without having to re-issue the --inspect command from the command-line?
The issue with re-issuing an --inspect command is that it generates a different chrome url every time and one has to then copy-paste this into Chrome each time. Ideally I want to be able to push F5 to re-start the chrome debug session.
So two issues:
I cannot restart the debug session without killing the current (i.e. no way to just refresh).
I have to copy paste the url into chrome each time I start a new session. (not as bad as issue 1.)
Here's a couple of options for you, though neither will provide you with a simple F5 refresh, both are significantly better than copy/pasting the new URL generated by the --inspect flag.
The most optimal solution is installing this extension for Chrome or Opera: https://chrome.google.com/webstore/detail/nim-node-inspector-manage/gnhhdgbaldcilmgcpfddgdbkhjohddkj
This will manage the node inspector for you. Just click the resulting toolbar icon and select "Auto" from the toggle switch. Your browser will then open the Chrome DevTools in inspection mode whenever your node server generates an inspection URL.
If you want to go the low-tech (and more manual) route, or don't want to install a Chrome extension, just open your Chrome to "chrome://inspect", wait a moment, and you'll get a list under Remote Target that will include your Node server. Just click the "inspect" link there, and the DevTools will open with the current URL. The downside of this method is you'll need to reclick that "inspect" link every time your server restarts. It avoids copy/pasting URLs, but still involves manual labor.
It's also significant to note that if you simply update the url of your inspector with the new ID, it will also work.
When you restart node, you'll get something like this:
Debugger listening on ws://127.0.0.1:9222/72c791b7-178f-47e8-93b1-d1be4d5ffe1e
The bit after the port/ is what you want. Replace that code in your inspector's url and it will connect to the latest session.
In the top of my website pages, I like to recommend to view my site with chrome and suggest to change browser add a button that when the user click it, one of the 2 options will happen:
If the chrome is already installed -> the page will be opened on the chrome (if the user's default browser isn't chrome).
If the chrome isn't installed -> popup message that says: "chrome hasn't been installed yet. For free download:link"
I assume that these two functions are already exist but I don't even know where to start searching....
Any help will be appreciated!!!
There is no way to detect what applications are installed on a computer using client side JavaScript (or any information that it sent to the server).
A webpage cannot trigger the loading on an arbitrary application on the client.
So no, neither of this functions exist and they cannot be written.
I wrote a very simple extension for Safari 5 that only outputs a single log message from a start script. This is the start script:
console.log('start script running');
If I go to any internet page (eg. http://www.yahoo.com) the log message appears in the error console. But if I open any local html page no log message appears. By local I mean on my local machine. It's like my start script never gets run for local pages (eg. C:/blank.html).
Is there a permission setting I'm overlooking or is this a feature of Safari 5?
Brent
Caveat: I only have a passing familiarity with the Safari extension system. I work on Google Chrome which is based on Webkit. Safari is also based on webkit.
In a Google Chrome extension, you can inject a script into a local page using the file:// permission. But I’m pretty sure there’s no analogue for Safari extensions.
The Safari page on script injection also seems to say you can’t interact with local files:
“You cannot access resources on the user’s hard drive outside of the extensions folder.”
There is no way to inject scripts on local files.
If you feel like it, you may complain at bugreport.apple.com.