-
Notifications
You must be signed in to change notification settings - Fork 27.5k
Unreadable error message when a module is missing #5568
Comments
I've never seen |
@CWSpear This text is not clickable. I've just checked that the problem doesnt' exist in Chrome stable (31) but it has existed in Canary for a long time now, not sure if it's going to land in 32 or 33. I'm using the non-minified Angular version. |
@mzgol Strange. I see it, too on Canary. You should open a bug with Blink/Chromium/Canary/Chrome or whomever you need to on that end, as that seems like it may be a bug/unintended (or otherwise should be reversed). |
Can someone file a bug with chromium please and cross-reference it here. I wonder if this is intentional change or not. |
Note: the change has already propagated to Chrome stable, now the clipped message is unavoidable. :( |
workaround: http://stackoverflow.com/a/22218280/674326 |
+1 |
1 similar comment
👍 |
I can't repro this any more. Was this reverted from Chrome? |
I'm not seeing this in Canary, but still seeing it on normal Chrome. Not sure about the details. However, at least in my case the error is even less helpful than the |
👍 In canary 36 it's still there. |
Also clipped in Chrome Version 34.0.1847.116 on linux |
Still a problem for me in 36.0.1967.0 Canary. |
Still issue in 36.0.1984.1 canary |
My console just says 'Uncaught object'. This is happens when a module is missing as a dependency. No module name I just have to guess modules until the error goes away. AngularJS 1.2.16 Check the console: |
Seems to be fixed on Version 37.0.2008.2 canary |
Hmm, I can't reproduce it in 37.0.2020.0 (Official Build 273468) or a content_shell I built today. With stable chrome, I can't seem to write a failing test case for this, although I do see the error in practice. It's not easy to tell what the error actually refers to, though, since the message which causes the Error constructor to throw is quite large, and trying to narrow it down isn't working so well yet. So, maybe this is something that will just fix itself in the not too distant future (since it's working well in Canary). maybe we should just leave it, unless you can put together a failing test case for us (It doesn't seem to be coming from <CR><LF> characters, colons, quotes, unix linefeeds, or anything else special --- it's really not clear what is causing the issue) |
@caitp There is a Chrome thread here https://code.google.com/p/chromium/issues/detail?id=331971 It apparently got fixed in Canary (v37) and it was due to buggy code for traces with more than 256 characters -> https://codereview.chromium.org/271733005/patch/1/10001 So I fear the problem will persist until Canary makes it as the main release |
@adriatic no need to worry about a reproduction, I see what the issue is. This is, as has been pointed out, a browser bug which has since been fixed. I don't think there is much benefit fixing this in core |
OK, I got too dependent on Chrome, it seems. The stack trace on Firefox correctly explains the cause of the failure (misspelled module name in this case) |
I opened my jsfiddle in Firefox and I get the error link now instead of |
I verified all other browsers and they all correctly show the complete stack trace - nit just Firefox. As @caitp stated above, we need to wait for Canary if we depend on Chrome in Angular development. |
However, this would be a good opportunity to print clearer stack traces when it comes to module names... (the actual error should come first) |
That would certainly be my take - as the current "Uncaught Object" error message has caused two orders higher waste of time than the time it would take to put in a temporary replacement for Chrome's stack trace formatting. Being an Angular novice, this specific behavior has caused me days and days of grief :-) |
Same here in 1.2! |
+1 @adriatic for "days of grief". Canary (v37) works well. |
I have no doubt that Canary works well (at least in the context we are discussing here :-). The fact that this "Uncaught Object" is a known problem for nearly a year and that nobody (in Angular team) found the time and need to put in a patch for this issue is what doesn't work well. I had so many people respond to my explanation of this problem I posted at Pluralsight site that it's pretty clear I was not alone suffering with this. This is not a big issue on a more global scale, it is just my deep conviction that the customer is always right (customer = developer in this case) that kept me from just ignoring it once I found what I need to do to go around it. |
I was experiencing the same bug. After testing in Canary, I noticed I had a typo in my Gruntfile. To everyone still experiencing this, try running your code in Canary. It gave me a more verbose log. |
Upgrade from beta.7 to beta.10 (nothing else changed). Same error |
Please check out latest Chrome Canary. It's already fixed there. All we have to do is to wait or to start developing on unstable releases what have other advantages as well like latest changes in the dev tools ;) |
This was happening to me in Chrome 35.0.1916.153. I tried Canary 37.0.2058.2 and it worked (i.e. I got a better log message). Thanks for the info. |
When Angular 1.2.x added the ngMinErr utility, this Chrome issue really started to see light of day. |
@ifyify +1,I also got this error message, have u fixed it? |
As suggested by several people in this thread (myself included) the best solution is to download the Chrome Canary (from https://www.google.com/intl/en/chrome/browser/canary.html) and use it instead of your current version of Chrome. |
@adriatic thank u! |
Thanks @adriatic |
This was fixed in Chrome 36 |
Chrome clips long error messages with no way to get to their full text making most Angular "unfound dependency" messages completely useless. An example I stumbled upon now:
The text was updated successfully, but these errors were encountered: