I've got a weird problem using MVVM-Light in a WP8.1 universal app. When the download folder has been changed, I am sending out a message via Messenger that contains the new StorageFolder download folder.
I am registered to receive this message in two different viewmodel constructors (one for the main page, one for a custom file manager page). Instances of both have already been created.
When I run this in my emulator, everything works as it should. I get the message on both viewmodels. However, when I run this on my phone (Lumia 920 with developer preview), I only get the message on the main page, not the file manager page that even sends the message (from the page's own codebehind).
At a glance I'm wondering if there's a thread problem here due to the difference in speeds... but I also wonder if there's a bug on the ARM side of the MVVM-light toolkit.
Based on how I was using the MessengerInstance.Register method, I was the victim of a limitation of the Messenger system (not a bug exactly). Someone else described it here exactly:
https://mvvmlight.codeplex.com/workitem/7640
Basically, I was using a lambda that included a parameter from outside the lambda statement, making it a class held with a weak reference. This was garbage collected and then gone forever. Fixed by rewriting the entire thing...
Related
Problem summary
I'm attempting to establish a connection between HoloLens and an Android device, which worked sporadically in the beta-version of the MRTK.
However since moving to the MRTK RC1 (Also Refresh), I've encountered issues regarding the QR-scanning. When pressing connect, the two devices seemingly finds eachother, however when the wearer of the HoloLens1 looks directly into the QR-code, nothing happens. (the white dot and "Locating marker..." text is showing)
Background summary
1. The Setup:
Implemented working MRTK RC1 Refresh
Cloned Feature-SpectatorView separately, copying only the "MixedRealityToolkit.Extensions" folder to the MRTK project.
"Spectator View - HoloLens" prefab added to scene.
First pressing "HoloLens" in the PlatformSwitcher, building for HoloLens1, then switching to "Android" and exporting the project to Android Studio.
Building the .apk from Android Studio
(opencv binaries are downloaded and implemented since beta version, I haven't since changed them from when they worked the last time.)
2. The Proces:
On the HoloLens, I press the "Connect" button in which a white text appears saying "Locating Marker..."
The Android phone presses connect and it goes to "Waiting for User" then as soon as a HoloLens is connected, it switches immediately to a QR code that should be readible from said HoloLens.
Looking directly at the QR-code and nothing new happens, connection does not establish further.
I checked if something was not ticked in Player Settings/Capabilities, but I can't seem to find what the culprit would be. Did I forget something in this proces?
There are a few things that could be causing this issue.
If the Android device is showing a marker, this means the two devices have established a network connection and are communicating with one another. Typically, when I run spectator view I enable the following capabilities: "Internet (Client & Server), Internet (Client), Microphone, Pictures Library, Private Networks (Client & Server), Spatial Perception, Videos Library, Webcam" in the Package.appxmanifest in visual studio. Pressing "HoloLens" on spectator view's unity platform switcher should typically achieve enabling these capabilities, but sometimes the package.appxmanifest doesn't get updated correctly in the visual studio project with subsequent builds in Unity. You can fix this by deleting your visual studio directory and rebuilding the visual studio project in unity.
If these capabilities are checked in the package.appxmanifest, it may be that you rejected a capability request when first running the application. If you open Settings -> Privacy -> Camera on the HoloLens, you can check whether your deployed spectator view application has camera access granted. You should be able to enable the camera functionality here if it is disabled.
There have been changes to both MixedRealityToolkit and MixedRealityToolkit-Unity spectator view logic, so cloning these items at different points in time may cause functions to no longer resolve (We're hoping to consolidate this code into the same repo/commit history in the future to prevent this from continuing to happen). Typically, in the Unity logs there will be errors stating that a function was not found for SpectatorViewPlugin.dll if the dll functionality is not resolving correctly. It sounds like this is not the issue you are hitting if things worked previously. But if it does turn out the case, it may be that you need to rebuild the SpectatorViewPlugin.dll to match the feature/spectatorView code you are using.
If you recently copied the SpectatorViewPlugin.dll and its dependencies to a new unity project, it may be that they aren't getting registered as usable by the windows uwp unity player. Make sure these binaries are in a Plugins\WSA\x86 folder within your assets folder. Also check the *.dll.meta definitions in the unity inspector to ensure the dll's are declared as usable for the unity wsa player/x86 builds.
Relatively new to working with .net but learning a lot while working with a developer at work who at times can be stubborn. Recently the asp.net web form we are developing has periodically started crashing such that every time a link is clicked that retrieves an object (an in process form that the user had previously filled out) the user gets the following application error: "Sequence contains more than one element."
All the sources I looked at online suggested that this was an InvalidOperationException thrown by the Single method and could be avoided by (a) eliminating the source of duplicate elements or (b) using the First method instead of the Single method.
The developer told me that I had no idea what I was talking about and that this was a caching error that could only be solved via an IIS reset.
Since this has been happening periodically I'm a bit concerned that doing the IIS reset doesn't address the underlying issue that creates the exception. Any one have any ideas? Is my developer right to say that an IIS reset is the best way to address this?
Thanks!
According to me your developer might have used LINQ. If he used single() or SingleOrDefault() or First() , tell him to replace those method with FirstorDefault() method.
A Qt project gives a number of errors - that I do not actually consider errors I think:
Some examples:
Fontconfig error: "/etc/fonts/conf.d/65-khmer.conf", line 32: out of memory
// ?
QGraphicsItem::ungrabMouse: not a mouse grabber
// i think happens when the mouse grab didn't quite happen, user moved too fast...
A few others
the first one seems it has an answer here https://askubuntu.com/questions/421891/fontconfig-error-out-of-memory
So I cannot fix it through software, the user must fix it... since it is system dependent...
Even built in release, these errors will still show up on command line... Since they do not cause the application to behave abnormally, I don't think the user should see them.
How can I build/deploy the application without these errors showing ?
ideally of course there would be no errors in a deployed application, but I don't know what I can do about those types of errors...
It is possible to redirect a stream (in this case stderr) after the fact using
freopen("/dev/null", "w", stderr);
Just place this in your main before any other code runs and you should be good. Keep in mind that this might need more work on Windows.
If the error message is posted through the Qt Message infrastructure (qDebug etc.) you can set a new message handler through the function qInstallMsgHandler.
I'm running into a very weird issue. I have created 4 environments:
frontend_dev
frontend_prod
backend_dev
backend_prod
In frontend_dev, the profiler is shown at the bottom of the page as it should be. However, on the backend_dev environment it is not shown. Upon investigation I have found out that the piece of code to load the profiler is actually there at the body end tag, however the call to get the profiler content doesn't return anything at all.
Further investigation showed that the loadProfile method failed because $this->storage->read($token) returns NULL.
I wonder what is happening, seems like somehow the storage couldn't save the profile and thus could not read it upon request?
Edit 1:
I think I know why: previously both my frontend and my backend were using the same environment (dev) and the profiler cached the toolbar content to the same place. However, now that we are using 2 environments the profiler caches the toolbar content to backend_dev but attempts to load from frontend_dev cache folder.
I have found the reason as described above, in the end for the sake of simplicity (and also other reasons) I have separated my original app into 2 apps using this guide
So far it has worked out very well for me.
Using dojo's TreeGrid (v. 1.9.0) served by my local machine (mvc.net) the grid loads fine with my sample, hard-coded data. However, when served from a remote machine, I get the un-debuggable "Sorry, an error occurred".
Breakpoints within my formatters are not even hit, suggesting the problem does not lie with any unloaded dependencies within a formatter. This fails not only on initial startup, but also if at runtime I try to reload the model using .setModel(myNewModel) after everything has definitely loaded. I have the firebug flag set in the dojo config object, and no informative warnings or errors are showing up in firefox...just that annoying and useless message in the grid itself.
The Sorry, an error occurred message is a result of a problem with your store and usually the data you are putting into the store.
The message is shown in DataGrid._onFetchError and TreeGrid mixes in the DataGrid.
You can
put a break point on the query engine of what ever store you are using,
analyze the data being passed into the store