we are developing an extension, hosted in the Google chrome web store.recently - we've got complaints from our users that sometime they get a notification window saying "the extension crashed, click here to reload".
after a short research we found out that this is happening only when we upload a new version to the Chrome Web Store.
we started to look it up online and found no documentation what so ever for this, so we started to check for it ourselves.
we tried to see what exactly can cause this problem and if we can identify a distinctive cause.
our tries included updating only the manifest.json file, a css file, a js file or not changing nothing at all but the version number, and on each change we've uploaded a new version and update it in about 10 different machines.
the results were the same, when on each update we made, it caused the extension to crash on just a few of the machines, while updating perfectly fine on the others. each time different machines acted differently.
then, we thought it might be related to the fact we have a timer working in the background page, and it might be happening just at the time it is working.
so we tried to raise the timer's frequency (from 5 seconds to 100 millisecond), and it still acted the same, crashing on only 3 out of the 10 machines.
we ran out of ideas now, and it really causing a problem in terms of user experience to our extension's users.
did someone had this problem, or came across any extension crashes on version update?
is it a known bug in chrome's extension engine or are we doing something wrong?
I am having the same problem and I think I found the cause. Do you by chance, override the new tab page?
I am able to reproduce the problem 100% of the time and when I remove the new tab override from the manifest, the problem goes away.
I opened an issue: Issue 104401
Related
I'm trying to help a client with their slow-ish website https://www.dp-tools.de. If I use Google page speed for mobile I can see that it takes 7 seconds to be interactive but nothing in the found problems really tells me why it is actually that slow. I also tried chrome lighthouse and couldn't really see all that much.
Is there another way of checking or maybe anyone here sees why it is so slow?
Open up the client's Shopify Admin, and examine the theme. Does it have any strange slow code in it? Examine the Apps installed that directly affect theme. Are any of them crap, old, broken junk?
The best way to debug a slow theme is to just start hacking out any junk the client may have added to their theme. A lot of themes are so bad, for example, they load jQuery 3 times. Likely you have one bad apple in there, a call that is blocking and takes way to long to timeout or respond. Developer tools can point these out to in your console. Mobile versions have to come with a developer console you can inspect too right?
I made this URL public a few years ago as a way to retrieve data. It was working yesterday. Today my site is crashing as it relies on this data, but google is now returning a 500 error if I try to retrieve it programmatically (through axios).
https://spreadsheets.google.com/feeds/list/1DLAVN3q758sPohCFeZlVSVRZKXzEser1SIsQnH2mvrw/ogwtdyp/public/basic?hl=en_US&alt=json
I have also created a code sandobox with react to demonstrate: https://codesandbox.io/embed/xjmy9vl2z4
However, I am still able to view the response in the browser if I am logged in.
Did google recently change their API so that it is no longer possible to retrieve data this way? Is there a way to configure my axios request in such a way to continue retrieving this data?
UPDATE:
Issue is being tracked here: https://issuetracker.google.com/issues/131613284
According to the support team: "The engineering team believes they have located the source of the problem and are rolling back to a previous build in order to resolve issue. Please stand by."
https://issuetracker.google.com/issues/131613284#comment109
UPDATE:
It seems like the issue has been resolved by rolling back to a previous build. I am no longer experiencing these issues. The link above, and the code sandbox example are both functioning. Reading the forum, it seems like others are also no longer experiencing issues. If you are still experiencing issues, I would encourage opening another (perhaps more specific) post or issue.
I have been trying to upload an app with a corresponding watch kit app for the last 4 hours (not even an exaggeration) and I am at my wits end... I have deduced that there is something not recognising the bundle ID in iTunes connect but I have been unsuccessful in figuring out what it is or how to fix it and as consequence unsuccessful in uploading the app, This is the first time uploading this app from my current computer, as I recently upgraded. However in the passed I have uploaded successfully from my previous machine. I cannot validate the application either it just gives me the error 'No suitable application records were found. Verify your bundle identifier 'XXXXXX.watchkitapp' is correct. I have tried with application loader, same error. I hope someone can help me, as I said I am no closer to figuring it out than I was 4 hours ago.
Edit***
I have looked at this link stack overflow question: no suitable records were found verify your bundle indentifier is correct but it hasn’t helped, my issue is not with the ios app but the Apple Watch counterpart, I do not know where to find the applewatch expected bundle id in itunes connect further more I don’t believe either bundle IDs in Apple watch, watch app or watch kit extension is incorrect.
Edited (again)****
So I decided that if the appwatch app has been causing so much trouble as I have been trying to figure it out for 3 days now, why not just remove it for the time being.... so I did then I got a new error No suitable application records were found. Verify your bundle identifier 'org.cocoapods.GTMSessionFetcher' is correct. I tried everything that I tried on the previous issue to no avail. HELP PLEASE? This is really frustrating, maybe there is some file that didn't copy over with the project that controls the various bundle id's I don't know, really any help would be greatly appreciated I guess at present I'm just one step away from completely remaking the app in a new Xcode project....
With a different Mac you probably need to download your certificates again. Also make sure you registered the bundle id in developer.apple.com
I'm working on an OS X app that most users choose to "launch at login", the kind you'd find at the menu bar.
In order to launch it at login I'm using SMLoginItemSetEnabled to launch a LoginHelper app that will open the main app, as described in this tutorial.
The app is failing to start up at login for just a handful of our users.
I was unable to reproduce this or to track the cause but I found (on a user's machine) that:
Deleting /Root/_com.apple.SMLoginItemBookmarks/[myapp] and /Root/[myapp] on /var/db/launchd.db/com.apple.launchd.peruser.$UID/overrides.plist and then resetting the Launch at login on the app fixed the issue. Also, we diff'd the files and the Data on the bookmark had changed.
For reference, I found about the overrides.plist here.
Since the app has both a Mac App Store and a Direct download version, I'm suspecting multiple copies of the app setting themselves as Launch at login may be the reason for this to fail, maybe these "bookmarks" are trying to open another instance of the app, that may or may not be deleted.
Now the questions, provided that this app needs to remain sandboxed:
Is there a way of updating that Data on the overrides.plist bookmarks?
Is there any way of deleting one self's app from the overrides.plist to start clean?
Is this maybe a known issue?
Any other suggestions on why the bookmark seems to point nowhere or how to fix it will be appreciated.
Note: This is my first question on StackOverflow, please excuse me if I failed to follow some of the suggested etiquette.
I don't know a definitive solution to this, I wasn't even aware of the overrides.plist. It could be related to multiple copies. As far as I understand, adding login through SMLoginItemSetEnabled sets a metadata flag that this Bundle ID should be launched on start. Then Spotlight, on start, will go do a metadata search on the file system and see which Bundle IDs need to be launched. Then, I guess, it will initiate the launch using the bundle ID. In my own application, Trickster, which uses the same technique for launching, I see that if I enable launch-on-login through the app itself, it might pop-up this strange message which refers to a debug build. I'm not even sure why it says about the first time. Very strange.
So, to have them launched, you have to make sure that the relevant bundles are in locations where Spotlight indexes (that the user hasn't disable Spotlight for these locations). Usually users don't disable Spotlight, especially for /Applications/ but I'm just saying.
What I usually suggest when support comes my way (and how I have it set up for me because I have multiple copies), is I to disable launch from within the app and instead add the correct one (from /Applications) manually in Login Items in System Preferences.
Many of our users are reporting that they are getting a blank page when using IE11 to access our website. Sometimes they don't even get a blank page, the browser just stays on the last page visited. These users can access other domains (such as google.com) without a problem.
For the browsers that are failing, if those users disable Protected Mode in IE, then they can again access our site, but this isn't a good solution because we have thousands of users and we can't go telling each of them to reconfigure their browsers, not to mention that we're completely losing the business of those potential customers who just surf by, see a blank page, and then keep on going without filing an error report.
Firefox and Chrome work fine. Also, even when using IE11 in protected mode, some users have zero problem, their computers just seem to work.
We are running the site off of IIS7. Other sites on the same server run fine, it's just the one particular site that is having the problem, and it's intermittent, affecting some computers and not others.
There must be something I can do on my side in the server or network settings to keep this problem from happening, but I'm baffled as to what it might be. When I look at the network traffic on a browser that's failing, the GET request is just being aborted with no explanation. When looking at the traffic with Wireshark, I see no errors, and no errors are showing up in the IIS logs. Of the browsers that fail, they are not even opening a connection to our web server to make a GET request, the request is just aborting immediately.
Any advice appreciated!
(followup): Another test we did: We can reproduce the problem on our development server, with the same pattern of which computers "work" and which don't. We tried turning off the webserver, and the problem stayed consistent, we'd still get a blank page error. So it's obviously not to do with anything that's in our page content? I've run tests on 9 computers on our office network: 5 worked, 4 failed. We're all baffled. :/
(January 24 followup): We figured out what's causing the problem, but not how to fix yet. Deep in the Windows registry, a key is being set in a Zonemap folder, adding a "play.net" domain with a key value of 2. IE sees that key when someone types play.net, and quivers and dies without an error message (Firefox and Chrome handle it fine).
So next question, what is setting that key in the first place? Probably an ActiveX control somewhere, but we haven't had any luck finding it in this 15 year old site, as many of the coders who may have created ActiveX controls in the past are long gone.
Does anyone know of a way to scan an entire domain and list anything that might be twiddling a registry key?
Followup mail threads from Elonka indicated that some users had the play.net domain mapped to a specific / non-default Internet Explorer security zone, and those users were the ones having problems.
Try add
<meta http-equiv="X-UA-Compatible" content="IE=10" />
into the header. It fixed IE11 for my website. It just forces it into IE10 mode (A complete joke though)
I accessed your domain in Internet Explorer 11, with Protected Mode enabled. I was unable to get a fully broken experience, but I did see something perhaps related to what your customers are experiencing. Some directories on your domain would hang for me, resulting in a white screen if they've not yet been loaded, or staying on the present page while the forthcoming page is loaded.
When I navigated to /dr, I found that my browser hung for over 7 seconds while your files loaded. Note, there were roughly 60 items, totaling about .78mb. This isn't much, but I'm presently on a relatively slow connection, and this resulted in a hang for me.
I would inquire what connection speeds your clients are on when experiencing this issue - it could very well be nothing more than some directories being served up more slowly than others. Regardless, you could optimize the experience considerably by compressing and concatenating your CSS files, as well as all of your JavaScript files. This results in fewer parallel requests, and quicker downloads.
If you have a reproducible example, please feel free to share and I'm sure myself and many others here will be more than happy to assist you.
For me box-shadow on an element caused this behavior. I removed the box-shadow(but left -webkit-box-shadow and -moz-box-shadow) and the problem disappeared. The interesting thing is that I still have box-shadow on another element...
If you are using 'offline.js' in your project, and you have version less than 0.7.19, then IE11 latest security update will see this code as potential security threat and block pages which are using 'offline.js' code.
Solution: Update to latest 'offline.js' version.