Check if there are really 1,000 connections on MySQL, and exclude whether the connection is leaked
Because changing any connection pool parameters will create a new connection pool, it is likely that it has nothing to do with the number of connections.
I looked into this recently too and found there are a few solid alternatives besides IdentityServer. Keycloak is a strong open-source option from Red Hat that supports SSO, LDAP, and social logins. OpenIddict is another good one if you're staying within the ASP.NET Core ecosystem. For more enterprise-heavy needs, WSO2 Identity Server and ForgeRock are worth a look, though they come with a learning curve. IdentityServer might dominate search results, but it’s not the only game in town.
Since it seems you are not updating a column on the database records to mark them as processed, your reader should have the "saveState" property set to "true".
One thing that no one mentioned was the Automatic Variables. Commands like $env or $executioncontext will be automatically (and the fact that they are variables, hence the name) recognized by the terminal
Run php artisan storage:link in your project root to create the storage link. Make sure your folder and URL use the same case (PlantHealthy). Check that your image file actually exists in storage/app/public/PlantHealthy. Then your image link will work.
I think this is an undocumented bug on keycloak side and affects developers using flutter app auth and keycloak. Someone needs to log it in github and add it to their backlog which consists of nearly 2000 bug tickets lol.
We have gotten around this by first initiating the delete action, then ignore all responses from keycloak and try to refresh user's token:
if token refreshes, assume user cancelled flow
if token refresh fails assume user is deleted - log them out
Update, I have beemn able to install the plugin, seems to be that my OBS app was not installed properly since it was not located in applications folder, now that I have successfully installed, I don't see the plugin available in trhe "tools" menu, can somebody point me in the right direction since I dont have a clue about what can be happening.
As far as I can tell, the linked page is inaccurate and Visual Studio does not actually query the compiler set by CMake for anything. When setting the intellisenseMode as described in the docs, that tells VS to use a proto-compiler built into Intellisense as the compiler for showing intellisense info.
Some evidence for this is when changing the intellisense mode to other values magically shows values for compilers that exist nowhere on my machine.
"intelliSenseMode": "linux-gcc-x64" shows this:
"intelliSenseMode": "linux-gcc-x86" shows this:
So in conclusion, this is just not possible. Microsoft, please stop lying about what your products actually do.
Since the default order is Integer.MAX_VALUE / 2 (see org.junit.jupiter.api.Order and org.junit.jupiter.api.MethodOrderer.OrderAnnotation.getOrder(MethodDescriptor)).
We can then simply do something like @Order(Integer.MAX_VALUE), @Order(Integer.MAX_VALUE - 1), ... for the methods we want to execute last.
It is now June 2025 and this problem persists. The Polly section of
the AWS PHP SDK appears to have been last updated in June 2016. The
PHP SDK is broken for the DescribeVoices command (it does not include
the SupportedEngines field in the result) and the SynthesizeSpeech
command (it drops or ignores the Engine parameter).
I had to fall back to invoking the AWSCLI using Exec(), which does
work as expected.
I am using PHP v8.2 and installed the AWS PHP SDK yesterday. (June 9
2025).
NGINX closes long-lived HTTP/2 streams as your 4:30-5:10 minute failure window is slightly above the default 3-minute [http2\_idle\_timeout](https://nginx.org/en/docs/http/ngx_http_v2_module.html#http2_idle_timeout). This directive is also obsolete since version 1.19.7, so ensure you are using [keepalive\_timeout](https://nginx.org/en/docs/http/ngx_http_core_module.html#keepalive_timeout) directive instead. Try increasing the values set for your `proxy\_read\_timeout`, `proxy\_send\_timeout,` and `keepalive\_timeout` to keep your HTTP/2 connections alive and idle longer.
Forcing curl to use HTTP/1.1 works because HTTP/1.1 handles long-lived streaming connections better in NGINX, than HTTP/2. So another workaround you can try is disabling HTTP/2 proxying upstream, and use HTTP/1.1 instead, while keeping HTTP/2 only at the client side. See related post.
Wow see if you can spot the typo here.
I am able to pass data between these pipelines with intersink/src but my intersink was missing it's closing quote so they weren't ever linked.
Just fixed the typo and have data flowing - seeing a lot of latency but that's a different issue. Will leave this up in case anyone else has a similar stack trace and it might help them.
I need some help on this one too. Trying to work on large file issues seems to hit the same problem: how to manage the internal buffering. Thanks for your references to UnmanagedMemoryStream; but MSFT still doesn't let you easily solve that core issue.
FileStream's loading options will reveal this problem very fast. When loading a large file, you'll be forced into MemoryMappedFile, or you won't have a solution at all (!!!)
Reading a 90MB text file loads approx 300MB of (managed) memory; and locks my rendering thread while I have a background worker trying to read these 90MB sequentially. It would otherwise be that you don't need to do that pre-loading. You may just be able to read each line with some native call. MSFT doesn't like this because they've tried too hard to manage everything on their backend - now with Task (TPL) thread support. So, there's "too much overhead"; but maybe not in each situation if we already know how to manage our access to large files, or handle threading. We can space out native calls, even, if you'd let us have access to the native file stream.
My current implementation uses and UnmanagedMemoryStream - which wants a SafeBuffer, or native pointer. However, there is no (native) call that I've found to ask for a portion of the file - nor is there a simple call to get data from a SafeFileHandle. So, I'm stuck parsing MSFT's FileStream code trying to figure out when it's safe to do.
I am adding content previously remove from ProfDFrancis' post which is useful for context of the question:
Update: Summary from answers
There are excellent answers below. The key is to realize that when you write a Cypress test like this:
it("...",()=>{
callFunctionA()
b += 1;
cy.log("Hello from line C")
d = 10;
cy.get("#input-box-e")
console.log("f")
})
... what actually happens is:
callFunctionA()
b += 1;
Add to list to do LATER: cy.log("Hello from line C")
d = 10;
Add to list to do LATER: cy.get("#input-box-e")
console.log("f")
Why? I think it is because Cypress was designed specifically to interact with web pages. The overwhelming priority is therefore to queue up the cy.get() commands, and retry them until they time-out, etc. That is so important that the makers of Cypress are willing (and indeed forced) to subvert our naive expectations of what an apparently synchronous list of code statements mean.
Only by pushing the cy.get commands into a queue where they can be tried and retried, can the web page engagement work so well.
I assume cy.log was added into the set of things that get added to that queue, simply because it is part of cy..
Explanation of the behaviour I saw
I think what was happening in my cases above is that the expect is, effectively, plain Javascript (not a thing dangling off cy.), and so it was executed first, while the cy.log was pushed into a queue to do later.
When the expect was successful, the cy.log eventually printed.
When the expect failed, that aborted the queue of cy. commands for that test, so the cy.log never printed.
Solution
So yes, if we want to log output immediately and not conditionally on the test passing, we should use console.log not cy.log.
Or we can delay the testing to occur inside the cy. queue, using cy.then(), as shown by @Fody.
OK! I believe this error means "The leaderboard API has an unspecified problem with your app's signing certificate, and is going to ignore your request".
If you've been vague or disorganised with your Android signing keys and confused by this error, now is the time to fix it :)
I had *three* different signing keys in place: the Google Store key, my debug key and one other key that I clearly lost track of. Only one of them was present under the credentials in the Google Play Console.
I added the missing two keys under Grow Users → Play Games services → Setup and Management → Configuration → Credentials, one for each key:
If you've not created each key in your corresponding Google Cloud project, you'll need to find the Create OAuth Client from the Add Credential screen, and add a new OAuth client with the SHA1 key for each missing key.
Once I'd fixed this server-side, the game started recording scores correctly.
I don't yet understand why it was OK with one account posting scores and not another. But making sure the Play Games services credentials reference the right OAuth client fixed it.
<video width="720" height="720" controls autoplay loop muted> <source src="https://storage.googleapis.com/generative-ai-assets/382be57c-f9e4-4fe1-96fc-94aa0f9c21b2/Op%C3%A7%C3%A3o%203%20-%20Foco.mp4" type="video/mp4"> Seu navegador não suporta a tag de vídeo. </video>
The answer was simple in the end. The data folder cannot be under Program Files on a windows server. This folder is protected with extra security that blocks access.
It seems that a workaround for this issue is to make everything else on the page grayscale. I don't know why this is connected to image encoding on the page but it is.
I set the user as email verified. I assigned a password in the credentials section and set it as temporary off. When I realized it wasn't working, I filled in the email, firstname and lastname sections and it was fixed.
When defining a potential function, try to make it proportional to the number of small cost operations that come before each large cost operation. For example, if you are finding the amortized cost of a dynamic array, your small operation is adding an element to the array. Thus, the potential function will be some constant times the number of elements in the array. You can figure out the constant later. Keep in mind that the potential function must capture some state of your data structure.
I know it's been a bit since you've asked this question, and it'll likely change again in time, but there's now a color picker when you click the swatch next to a color in the inspector. And if you hold the shift key while you click, it'll cycle between formats for the color (RGB, hex, etc.).
I know this is an old post but did you ever resolve this issue? I'm encountering the same issue with a specific endpoint when a get request is made. I was able to determine the request is never being sent so I think it's related to the UnityWebRequest class specifically. Here's my post in case anyone is interested:
Azure Backup gives you a simple, reliable, and cost-effective way to protect your VMs and recover quickly if something goes wrong or corrupted, a solid DR solution to keep you covered until you're ready to move to Azure Functions.
To enable Azure Backup for an Azure virtual machine, open the Azure virtual machine, then navigate to the 'Backups' section and select it.
You can select an existing Recovery Services vault or create a new Recovery Services vault.
You can also create the policy based on your requirements, such as backup schedule, instant restore, and retention of daily backup points. In the backup policy, there are two types of policies the Enhanced policy, which allows you to take up to 4 backups per day and the Standard policy which supports only one backup per day based on snapshot frequency. Once selected, click on Enable Policy.
The problem disappeared since then. I suspect the "solution" was to install the newer version of Rider and clean up the residuals from the older version. I cannot guarantee this, obviously.
If I understood correctly, you only need this function that communicates with the server not to block the rest of the code? If so, using aiohttp (async requests) will solve this.
BYW asyncio is meant to replace multithreading, as it allows running io-bound tasks without blocking the system, which is exactly what you need.
I have encountered an issue I don't understand how to solve. When i change instance domain to HTTP:, coolify doesn't open: no sever available, 404, etc. I'm not sure what's wrong as in the CF tunnel I added *.domain.com HTTP:80 and coolify.domain.com:8000
@ThomasKläger I'm answering here to answer with a screenshot.
It is indeed strange! Your screenshots clearly show that the option is present in the 2025.1.2. I'm quite non-convinced to believe that I don't see this option in my setting (screenshot below) just because it is 2025.1.1. (Doesn't make sense a minor version upgrade would add this feature which was there for decades!)
You could set up a shared mailbox under an alternate email address. This would need to a proper inbox rather than just adding an alias to your existing inbox. Alternatively, if you want to send calendar invites from an alias for official announcements without having to set up a new inbox you can send emails with ics files through smtp from your preferred alternate email address or use Salepager which lets you send Outlook invites from an alias or alternate email address.
If I understand correctly, you'd probably get what you want (a titleBar-less window) by changing the BaseWindowConstructorOptions preferences to this:
mainWindow = new BrowserWindow({
height: 600,
width: 1000,
titleBarStyle: "hidden", // this hides the title bar
webPreferences: {
preload: MAIN_WINDOW_PRELOAD_WEBPACK_ENTRY,
},
});
Hoy 10 de junio de 2025 me aparecio ese error sin haber cambiado algo del proyecto, lo que hice fue actualizar la versión la springboot de la 3.4.5 a la 3.5.0 y se solucionó.
El error era:
org.hibernate.bytecode.internal.bytebuddy.BytecodeProviderImpl Unable to get public no-arg constructor
As of June 2025, the latest gdb version provided by Homebrew is 16.3, running macOS Sonoma 14.7.5. When I try to debug an executable I still see the same behavior (just hangs after the new thread message), and when I kill the process I get the warning:
warning: unhandled dyld version (17)
The source code for 16.3 (solib-darwin.c) still has a max DYLD version of 15.
I have the same problem when importing HuggingFaceEmmbeddings as well as OllamaEmbeddings:
ImportError: cannot import name 'convert_to_openai_image_block' from 'langchain_core.messages' (E:\U-DEVTOOLS\Python-3.11.9\Lib\site-packages\langchain_core\messages\__init__.py)
from langchain_core.messages.content_blocks import (
convert_to_openai_data_block,
convert_to_openai_image_block,
is_data_content_block,
)
I also discovered another Messages.py in \Python-3.11.9\Lib\site-packages\langchain\schema, also defining the same class, but exporting only a part of the methods. This looks like an interface.
When I try to load the OllamaEmbeddings interactively in my notebook, I have the same error. And when I remove the convert_to_openai_image_block function from the import statement:
from langchain_core.messages import AIMessage, BaseMessage, convert_to_messages
print(convert_to_messages)
from langchain_core import messages
print(messages)
<function convert_to_messages at 0x00000169A8185620>
<module 'langchain_core.messages' from 'E:\\U-DEVTOOLS\\Python-3.11.9\\Lib\\site-packages\\langchain_core\\messages\\__init__.py'>
Can someone explain what's going on ?
Reasons:
Blacklisted phrase (1): I have the same problem
RegEx Blacklisted phrase (2.5): Can someone explain what
RegEx Blacklisted phrase (1): I have the same error
That is because vercel issues a certificate for your deployment and you are now running on https and you are probably no longer working on http://localhost:4000 so you needto add to your cors policy
SELECT REGEXP_REPLACE(
LOWER(TRIM('Isabelle Cándido de Paula')), '(\b\w)(\w''*)', x -> upper(x[1]) || lower(x[2]))
Results in:
Isabelle cándido De Paula
Must be:
Isabelle Cándido De Paula
I just published a Swift package for developers who need to create CMS (Cryptographic Message Syntax) signatures in iOS/macOS apps using OpenSSL under the hood.
CMS signing is widely used in secure email (S/MIME), digital document signing (PDF), and many enterprise-level protocols — and yet it’s rarely supported out of the box in Swift.
With SwiftCMSSigner, you can now generate detached, base64-encoded CMS signatures using a simple, modular Swift API. No external tools. No hacks. Just clean integration via Swift Package Manager.
You're halfway there! If your websites are accessible by direct URL but not appearing in Google search results, it’s most likely a search engine indexing issue.
function useNestedFieldError(error: FieldError | undefined, fieldName: string) {
if (!error) return undefined;
if (error.type === "required") return error;
return (error as any)[fieldName] as FieldError | undefined;
}
// Usage in component:
const xError = useNestedFieldError(fieldState.error, "x");
The cleanest solution would be to properly type your form errors from the beginning. When setting up your form, you can provide a generic type that includes the nested structure:
type FormValues = {
vector: Vector3D;
};
const { formState: { errors } } = useForm<FormValues>();
// Now errors.vector will be properly typed as Vector3DError
This way, TypeScript will understand the nested error structure throughout your application.
I also encountered the issue. However, after checking if the location is enabled as Crispert suggested, I moved my startScan method from onCreate to onResume. This modification resolved the problem for me.
You should go to the root project :
in Terminal type :
cd [project paht]
For get full path you can copy path from folder in your finder ( in mac , go to the project and right cleck in your folder and push option button) and copy full path
cd /Users/..name../Desktop/FlutterExamples/name_of_project
finally got it working. the config looks like below:
from langroid.language_models import OpenAIGPTConfig
ollama_config = OpenAIGPTConfig(
chat_model="ollama/mixtral",
chat_context_length=16_000,
api_base="http://<your.ip>/v1" #the /v1 was necessary for my ollama llm
)
I'm about 12 years late on this, but I just bumped into this searching for something else, and I believe I've just seen code that does exactly what you want as part of the open source project Schism Tracker:
I had this kind of trouble with a /tmp/xxxx.cmd file. From cygwin point of view, it had x-permission, from windows point of view (ls -ls 'C:\ProgramData\cygwin64\tmp\xxxx.cmd') it had not. In https://unix.stackexchange.com/questions/549908/chmod-not-working-in-mingw64-but-working-cygwin I found that if the file contents starts with ':\n' then cygwin recognizes it as an executable thingy, and it will display the x bit.
The NCCL backend fails using under Docker + WSL2 (unstable multi-GPU communication) possibly due to the incompatibility of the NCCL, Docker and WSL2 versions that are not compatible. Also, NCCL works best with Native Linux.
The other alternative is to increase the GPU size in the system that can accommodate the model requirement.
