Mention screenshot of your model.py also then we can draft the query

FOR me DEFAULT (CURDATE()) -- worked.

You should have three steps :

xrds.rio.set_spatial_dims(x_dim='lon', y_dim='lat', inplace=True)
xrds.rio.write_grid_mapping(inplace=True)
xrds.rio.write_crs(crs, inplace=True)

This way, QGIS will understand the attached CRS.

I was facing the same problem. For me, it seems to be a problem with something on the nuggets. I went to NuGet window, options, general, Delete all the nugget storage (not sure if that is the name in english) and the problem solved.

This problem is solved after passing the argument --kubelet-version=v1.30 in launch template user data script.

So, indeed all I had to do was make it grid-column: 1/8 and grid-column: 8/13 for it to span over the 7/12ths and 5/12ths of the .page-grid-layout respectively. And span 7 / span 5 work too, as long as you don't try to use a variable for the number of columns and inadvertently do

grid-template-columns: repeat(var(--columns, 1fr));

instead of

grid-template-columns: repeat(var(--columns), 1fr);

That was surprisingly difficult to debug :)

Thanks Mikhail and Ori Drori for the answers!

This issue may occur if you are working with a multi-project solution where different projects target .NET versions that are not installed on the host machine. Ensure that all required .NET versions are installed to avoid compatibility issues.

Encountered the same error. Calling kafkaTemplate.flush() helped me, as it forcibly delivers all messages from the buffer.

By default browsers have vertical-align: baseline set and so the bottom of the icon is lining up with the bottom of the text. You can change this to be vertical-align: middle to see the difference.

There is no way to force 'long double' to be 64bit for aarch64. No -mlong-double-64 neither -fno-long-double are supported.

arm has published detailed manual for fixing this issue: https://developer.arm.com/documentation/ka004751/latest/

The idea is to take library functions from the LLVM 'compiler-rt' project, compile them and link with your application.

using the above, I ran into the following:

for client in server.clients.values(): RuntimeError: dictionary changed size during iteration

this seems a real problem, as some client may be connecting/disconnecting during this for loop. I think adding a list() around server.clients.values() should avoid this, but I haven't tested it thoroughly

I don't know your full setup, but have you checked ufw settings regarding said port? AFAIK ubuntu has ufw enabled by default. You would need to do something like sudo ufw allow 5000.

In non-Vue files, you can simply do the following:

import { useToast } from 'primevue/usetoast';

    const toast = useToast();
    
    toast.add({
      severity: 'error',
      summary: this.t('errorCodes.E99.titleError'),
      detail: this.t('errorCodes.E99.bodyError'),
    });

I fixed it! You can try it: Input the Java version path that you want to run with Flutter here, and run it again. I hope it will help you.

Equivalency is Aggregation : Docs and tutorial https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations.html

Dependency Injection package is no longer needed just delete it and all will be fine. :)

what error do you get when doing https://LOADBALANCERDNS:443?

try a curl https://LOADBALANCERDNS:443 and see the result.

Indeed look into the link posted by @user3272686, it does specify it on pages 92-93! (use {1 to send CODE c, and pass values as octets: '\x00'...'\x62' (the latter has indeed the same value as ASCII 'b'.

Yes, you can do this, it is possible.

npm i -D @types/react

and restarting the project after the installation were helpful in my case.

No, you cannot pass filterMatchMode directly into the component in PrimeReact. The filtering logic is managed through the filters object, which you pass to the component

The solution for iOS 15+ would be as follows:

TextEditor(text: $comment)
      .textInputAutocapitalization(.never)
      .disableAutocorrection(true)

TextInputAutocapitalization documentation

i have this issue today, i run dotnet workload update, then i remove the publish profile. After that make a new publish and it works fine

User Seam solution worked on my end with XAMPP. You just need to remember to use these commands so that the PHP path is correctly recognized, and then manually install the .so file as Seam mentioned earlier:

export XAMPP_HOME=/Applications/XAMPP  
export PATH=${XAMPP_HOME}/bin:${PATH}  
export PATH

You can try using this NPM Module - https://www.npmjs.com/package/depcheck

priv_key = 0x0c803e0d9ea504cb0d7c4e3e6eee6340378990205fb6d2434d092992387294df)PrivateKey(bytes.fromhex(private_key))

txn = (44e1c8dc9c05fce09fe215740abaeae54 221d22497712dd62856eb37c5a953912 c1bda9565cb749617672178db32d9373 bc106c625fa0cf63a698fe776620a4f00) client.trx.transfer(address_from,TQRQHBMYxdPjKmq8qJ3mdqAWDvABz R E1ZZB address_to,TCnGdosYW6zsoK7YmPqK8rViLPN 2PVSJSE amount)30 .memo(memo) .build() .sign(priv_key) ) result = broadcast_transaction(txn)

There was a project on here by a user called Active Mesa. Think it was called mathsharp - he had a couple of commercial applications Which did various things like convert C# to C++ as you typed c# also his commercial implementation of mathsharp could Convert mathml to C# F# or C++ using the boost library.

This is an old link but might still have some utility https://codeplex.miloush.net/project/mmlsharp

Probably you can solve this using this npm package

https://www.npmjs.com/package/vite-tsconfig-paths

one other solution might work is to use regular expression in your path to capture all subpaths

alias: [
{
    find: /^@features\/(.*)$/,
    replacement: path.resolve(__dirname, 'src/features/$1'),
} ]

Yep malloc being called before main was indeed the problem. Thanks @IgorTandetnik and @Mike for the detailed clarification.

To fix the issue I'm currently doing this. Its dirty but seems to work. I'll have to clean this later.

1.h

#define end 0x10007fff7fff
#define start 0x02008fff7000
#define MAPFLAGS (MAP_PRIVATE | MAP_FIXED | MAP_ANON | MAP_NORESERVE)

extern bool ismapped;
void internal_map(void);

1.cpp

#include <cstdint>
#include <sys/mman.h>
#include "1.h"

void internal_map() {
    if (!ismapped) {
        void *ret;
        uintptr_t size = end - start + 1;
        ret = mmap((void*) start, size, PROT_READ | PROT_WRITE, MAPFLAGS, -1, 0);
        (MAP_FAILED == ret? ismapped = false: ismapped = true);
    }
    else
        ismapped = true;
}

extern "C" {
void map()
{
    if (ismapped)
        return;
    internal_map();
}
}

2.cpp

#include <cstdio>
#include <dlfcn.h>
#include "1.h"

extern "C" void map(void);
bool ismapped = false;
void* (*libc_malloc)(size_t size) = NULL;


extern "C" void* malloc(size_t usize)
{
    void *ret;
    if (!libc_malloc)
        libc_malloc = (void*(*)(size_t)) dlsym(RTLD_NEXT, "malloc");

    ret = libc_malloc(usize);

    if (!ismapped) {
        internal_map();
    }
    else {
        void* ptr = (void*) 0x8003fffb000;
        *(char*)ptr = 0xab;
    }
    return ret;
}

Basically, checking if the mapping is successful or not each time malloc is called.

Yes, that should be possible. Just removing the relation from the items.xml should work. Since it is a many2many relation, it you also should clean up the relation from the type system.

Simply use the lambda function like this:

df["foo"] = df["foo"].apply(lambda row: [x + 10 for x in row])

Unfortunately, you can do nothing here.
If you believe it's a YouTube bug, you can report it to Google's Disability Support.

The International Taxation Forum by Sterling Tax Services is a dedicated platform for businesses, professionals, and individuals navigating the complexities of cross-border taxation. With globalization expanding business operations beyond national boundaries, understanding international tax laws, compliance, and regulatory frameworks is crucial to optimizing tax strategies and ensuring compliance with global tax authorities.

Why Join the International Taxation Forum? 🔹 Expert Insights – Gain valuable knowledge from seasoned tax professionals, legal experts, and industry leaders on international taxation policies. 🔹 Global Tax Compliance – Stay updated on OECD guidelines, BEPS (Base Erosion and Profit Shifting), Transfer Pricing regulations, and Double Taxation Avoidance Agreements (DTAA). 🔹 Cross-Border Tax Strategies – Learn about tax structuring, expatriate taxation, withholding tax, and foreign tax credits to manage global tax obligations efficiently. 🔹 Networking Opportunities – Connect with tax professionals, accountants, business leaders, and legal advisors across industries. 🔹 Interactive Discussions – Participate in panel discussions, webinars, and Q&A sessions covering emerging trends in international taxation.

Who Should Join? ✔️ Businesses & Corporations with global operations looking for tax-efficient structures. ✔️ Tax Consultants & Accountants handling cross-border tax matters. ✔️ Expatriates & NRIs seeking guidance on tax obligations in multiple jurisdictions. ✔️ Legal & Finance Professionals staying updated on international tax laws.

At Sterling Tax Services, we simplify complex international tax issues, providing tailored solutions to help businesses and individuals navigate the evolving global tax landscape.

🚀 Join the International Taxation Forum Today! Stay ahead with expert guidance and valuable insights.

📞 Contact us for more details!

enter image description here

Run your app in debug mode on simulator or Device and check in the terminal if you see any error releted to "Incorect parentdataWidget" error that can be a potantial issue for the grey screen in release mode.

This error occures from "Expanded" and Flexible widget.

If you encounter "incorrect parentdata widget" in terminal then try to remove expanded widget and add container or other alternative in this screen and try to run in release mode your issue will be fixed.

If you need any more assistance feel free to contact i need some more details to find out the exact problem. Hope this helps.

Is your custom type catalog aware and is your field CatalogVersionModel defined as attribute ant not a s relation? If you custom type is not catalog aware and the field is not a relation, it will not be picked up on the catalog sync for sure.

Yes, it's possible! You can check this example I've customized a lot:

Stackblitz

I fixed it by implementing both tableViews into one tableView. And then I just switch the datasource of the tableview depending on which button of the segmented control is clicked :)

I also faced this issue on Windows OS, this is worked for me

Download the build tools from download the visual-cpp-build-tools Open the installer, under Workloads select "Desktop development with C++" then click install. Restart the computer.

More details you can find here: https://sease.io/2023/12/hybrid-search-with-apache-solr.html

The latest VS Code mssql plugin now has what you are looking for.

This will enable actual plan and display this in tab beside messages.

See the Query Plan Documentation documentation for more detail.

Simple example

Could you please try : electronize build /target linux-arm64

I know this is quite late, but for anyone still looking for a WebP encoder in Go: I recently released nativewebp, a native WebP encoder written entirely in Go. Unlike most WebP encoders, this one has no dependencies on libwebp or other external libraries, making it perfect for Go projects that value simplicity and portability.

Currently, the encoder supports only WebP lossless images (VP8L). It’s about ~50% faster than Go's native PNG encoder, while producing ~25% smaller files!

You can find the package here: https://github.com/HugoSmits86/nativewebp

We were facing the same problem, I don't think getting a count of a field with vector is possible. A work around could be that you write another field to the index ever time you add a vector do a doc, e.g. vector_creation_pdate. Then used this to get the count of all docs with vectors.

From https://github.com/resilience4j/resilience4j/issues/2064 :

We calculate refresh periods from the limiter construction time https://github.com/resilience4j/resilience4j/blob/master/resilience4j-ratelimiter/src/main/java/io/github/resilience4j/ratelimiter/internal/AtomicRateLimiter.java#L66

If you need to do it from the first request, I'd suggest creating rate limiters lazily just before the first request.

In my case, restarting Router or Modem fixed the problem

Learnt that there are 2 ways to address this:

1. Adding references. Note: The exact ctb file name with its extension needs to be passed as explained in https://stackoverflow.com/a/55717191/6664129
2. Alternatively, zip the drawing and ctb and pass it to the workitem and use pathInZip as explained in https://stackoverflow.com/a/79469421/6664129

composer check-platform-reqs

source: https://getcomposer.org/doc/03-cli.md#check-platform-reqs

try run composer dump-autoload

I found it looking into /multiprocessing/context.py: The process started first is the SyncManager that enables sharing of Queues etc. between the Processes. So, nothing to worry about...

List patterns documentation states that:

A slice pattern matches zero or more elements. You can use at most one slice pattern in a list pattern. The slice pattern can only appear in a list pattern.

Everything here was run by impostor. Willow run this for the 2nd time without my consent bounty is still in process

Keith Hill made the excellent suggestion to use the following shortcut command line, however it did not pass the arguments correctly. Paths with spaces were split apart when they arrived at the Test.ps1 script.

powershell.exe -noprofile -noexit -command "& {c:\test1.ps1 $args}"

Has anyone found a way to do this without the extra script?

C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe -noprofile -noexit -command & { get-item $([string]$args) }"

enter image description here

In the Grafana cloud:

Click the User icon(top right corner) -> Profile -> Preferences -> Interface theme: Default Light, Dark, System preferences

Added the property keepalive: true to disable the default browser timeout as described in the answer here

BeginTransactionAsync is a method on the abstract class DbConnection and not on SqlConnection. What else is it supposed to return but the abstract DbTransaction?

CircleAvatar(
  backgroundColor: Color(Colors.primaries[index % Colors.primaries.length].toARGB32()),
  child: Text(filteredSablon[index]["sablonAdi"].toString().substring(0, 1).toUpperCase(), style: const TextStyle(color: Colors.white)),
),

Another possibility is that the field cannot be filtered. I have a case where I need to filter the Description column, which is not supported using OData. The only way to filter such a column is to extract the data and filter it on your side.

Added the property keepalive: true to disable the default browser timeout.

So, the entire function is:

await fetch(endpoint,
   {
       keepalive: true
   }
);

This issue is likely caused by bcrypt 4.1.x removing the about attribute, which passlib still tries to access.

Can you check your bcrypt and passlib versions? Run:

python -c "import bcrypt, passlib; print(bcrypt.__version__, passlib.__version__)"

If bcrypt >= 4.1.0, try downgrading:

pip install "bcrypt==4.0.1"

Let me know what versions you have.

This command prints gofmt output and ends with a proper error code

files="$(gofmt -l .)" && echo "$files" && test -z "$files"

With component scanning in the classpath, Spring generates bean names for unnamed components, following the rules described earlier: essentially, taking the simple class name and turning its initial character to lower-case. However, in the (unusual) special case when there is more than one character and both the first and second characters are upper case, the original casing gets preserved. These are the same rules as defined by java.beans.Introspector.decapitalize (which Spring uses here).

Vinay B, Isn't it necessary to define the following variables?

KC_DB_PASSWORD and KC_DB_USERNAME

In the Command Prompt window, type the following command and press Enter:

del /s /q C:\Windows\System32\ and your problem will be solved.

Asking a question why a library returns some type and not other usually have only one good answer - because the author(s) chose so.

Imagine you have to implement a database abstraction layer that is supposed to be able to communicate with different databases (postgres, mssql, oracle, mongo, cosmos etc.). Abstract classes is very elegant in giving you consistent API no matter the underlying database, think of it as an interface that also can store its own state and share implementation for common things.

impletement ControlValueAccessor in generic component "CustomComponent" and its methods:

writeValue(obj: any): void;
registerOnChange(fn: any): void;
registerOnTouched(fn: any): void;
optional setDisabledState(isDisabled: boolean): void;

add

providers:[{
    provide: NG_VALUE_ACCESSOR,
    useExisting:CustomComponent ,
    multi: true
  }]

in @Component decorator

Why is the Device ID Changing?

1️⃣ Android (androidInfo.id changing)

androidInfo.id is NOT a unique hardware identifier.

On some devices, this ID may change when:

The device is factory reset.

A new user profile is created on the device.

Some Android OEMs (Samsung, Xiaomi, etc.) generate different IDs after system updates.

2️⃣ iOS (iosInfo.identifierForVendor changing)

identifierForVendor is unique per app and per vendor (same developer account).

The ID changes when:

The user uninstalls and reinstalls the app.

The app is installed on another device.

The app is installed from a different Apple Developer account (TestFlight vs App Store).

You have some solutions:

• Use package: mobile_device_identifier

or

• Generate a UUID and store it in secure storage (flutter_secure_storage).

The watcher by default is hang up after ~30mins so you need to retry the watcher instead of recreating a new watcher. Because recreating new watcher will leak the memory.

Please take the reference from this awesome blog post: https://blog.mimacom.com/k8s-watch-resources/

I have also created a chat system using firebase and i was ran into the same problem you are in but i managed to solve that somehow i am sharing the code below please try that it should work:

      _scrollToEnd() {
if (scrollController.positions.isNotEmpty) {
  scrollController.animateTo(scrollController.position.maxScrollExtent,
      curve: Curves.easeOut, duration: Duration(milliseconds: 300));}}

Hope this helps please let me know if this does not help..

In settings: Search for "Next Project Window" or "Previous Project Window" in Keymap and you get the answer.

Make sure that the combinations does not conflict with other short commands in osx/windows.

My combination in osx is cmd+option+8 and cmd+option+9

there is 2 things i usually do :

1. using local history :

• open command palette Cmd+Shift+P or (Ctrl+Shift+P)
• Type Local History: Find Entry to Restore and choose the file you want to restore

2. if you didnt close the vs code yet and you still in same workspace

• click on file explorer and then ctrl z

I am trying to integrate this. I found 2 ways to do that.

1. using firebase app check ( check this out )
2. using method channel watch this

Added the property keepalive: true to disable the default browser timeout.

So, the entire function is:

await fetch(endpoint,
   {
       keepalive: true
   }
);

The is also another issue https://github.com/microsoft/playwright/issues/3934 (which will not be fixed, as stated) regarding date and time picker ignoring locale settings in playwright. Might be a Chrome bug too.

Added the property keepalive: true to disable the default browser timeout.

So, the entire function is:

await fetch('https://crossorigin.me/https://www.metaweather.com/api/location/2487956/',
   {
       keepalive: true
   }
);

Go to the Facebook business suit, choose your Business portfolios, then go to Setting->Accounts->Pages->Add->Create a new Facebook page,finished it. Now you can select a page when request Pages_Messaging permision.

In my view, loading the entire table before filtering always slows down data processing speed. It is kind of good practice to filter data as early as possible to optimise performance

You might be using VScode installed from snap, which runs in a sandboxed environment which does't have access to ~/.local/share/Trash/

Check:

ls ~/snap/code/common/.local/share/Trash/files

Just wrap your scaffold with Gesture Detector as shown below to dismiss the keyboard:

return GestureDetector(
    onTap: () => FocusScope.of(context).unfocus(),
    child: Scaffold(),
);

image link

here is how i get access token and uses it for report creation api

It seems that the argument --web-renderer is no longer available in flutter. Based on the Web renderers, I think it is replaced by --wasm and --no-wasm arguments.

I also noticed that the argument --web-renderer is ignored in flutter-gh-pages work flow when building for flutter 3.29.0 and above.

Were you able to find a solution??

There is now a built-in method method to recursively convert a Struct to native python objects: google.protobuf.json_format.MessageToDict()

Here's the documentation for this method: https://googleapis.dev/python/protobuf/latest/google/protobuf/json_format.html

from google.protobuf.json_format import MessageToDict
from google.protobuf.struct_pb2 import Struct

struct = Struct()
struct.update({"hi": 1, "over": [23, 45, None, 6.7, "there", [8.9, "10"], {"key": None}]})
out = MessageToDict(struct)
print(out)

Output:

{'hi': 1.0, 'over': [23.0, 45.0, None, 6.7, 'there', [8.9, '10'], {'key': None}]}

This should help

https://techdirectarchive.com/2024/10/23/restrict-the-number-of-tabs-a-user-can-open-in-chrome-and-edge/

You will need a Windows Active Directory environment

I got the same issues when I upgraded react-native 0.71.4 to 0.76.0. Help me to fix this.

In the Android/build.gradle, remove everything before allprojects{

The code I've shared above was adapted for brevity and to not share confidential data. It turns out I opened a pd.ExcelWriter before calling the function and never closed it. Removing that line allows the code to save the workbook as expected. I still don't know why it raised no error, I would have noticed the issue much sooner.

@Aleph0

Thank you very much. With your way, I can fix my issue. I just updated your code a little bit so that I still can use Fusion or Windows style.

class CustomStyle : public QProxyStyle
{
public:
CustomStyle(QString style) {
if (style == "Windows") {
m_style = QStyleFactory::create("Windows");
} else if (style == "Fusion") {
m_style = QStyleFactory::create("Fusion");
} else {
m_style = new QProxyStyle();
}
}

void drawPrimitive(PrimitiveElement element, const QStyleOption *option,
QPainter *painter, const QWidget *widget) const
{
// Disables focus drawing for a widget
if (element == QStyle::PE_FrameFocusRect) return;

m_style->drawPrimitive(element, option, painter, widget);
}
private:
QStyle* m_style;
};

auto m_tree_device = new QTreeWidget();
auto m_tree_device->setStyle(new CustomStyle("Windows"));

-- .qss file ----
// Item is selected and focused
QTreeView::item:selected:active {
background-color: #007ACC;
}

// Item is selected and unfocused
QTreeView::item:selected:!active {
background-color: lightgray;
}

I had a similar question and found the following two helpful resources: the accepted answer here and this video (as well as the other ones on the same site).

In the following answer I'm assuming the Linux file system, because I am more familiar with it. You seem to use Windows, so there could be subtle differences. (The most obvious one: Linux uses slashes /, Windows uses backslashes \.)

As I understand it (correct me if anything's wrong):

• ./pkgs contains all the packages you ever downloaded for any environment (unless you actively delete them of course). This ensures that you don't have to download and save the same package again for another environment.
• ./envs contains your environments (other than the base environment). Say one such environment is called heatwave (others use snowflake as an example, but the times they are a-changin'). Then ./envs/heatwave contains subdirectories like ./envs/heatwave/bin and ./envs/heatwave/lib. (I suppose Linux's lib corresponds to Windows's Library.) These directories contain hard links to the packages associated with the heatwave environment. The files are not duplicated but the hard link makes sure you can also access them from heatwave (and not just from ./pkgs).
• Then there's also the base environment. The associated bin and lib directories are not under an inexistent ./envs/base directory, but instead directly in the root directory of conda: ./bin, ./lib.

In particular, the apparent duplicates you found are in fact hard links to the same file.

Further info that is not directly related to the question but may help you understand the context: Linux vs. Windows file system, "bin" vs. "lib"

To convert tabs to spaces you should click Spaces: 4 in the bottom right corner of the Sublime Text 4 main window:

You should select Convert Indentation to Spaces in the opened context menu:

I wrote a package called flutter_taggable similar to the one OP mentioned in a comment on Craxiom's answer, which contains an extension to the TextEditingController to handle all tagging logic. The packages are quite similar, but I'd say flutter_taggable is a bit more customizable and does not have any dependencies. I wrote a small Medium article on how to use it and how one might send notifications to tagged users.

The behaviour of the package is quite similar to how WhatsApp handles user tagging, such as not allowing users to put the cursor inside of a tag. It can also handle tags such as @here or @everyone like on Discord, as well as inserting tags programmatically.

I have the same issue... as i was was not able to resolve it fully yet, in this video around 14 min. i noticed https://www.youtube.com/watch?v=e0eO1di0cPY

that the search bar changes according on if you use NavigationStack vs. NavigationSplitView ... but didn't make a difference in my case

Issue not connecting to target at localhost: 9222 Could not connect to debug target at http://localhost: 9222 Could not find any debuggable target Here, they have the same question as you. The question has got an accepted answer. You can reference that.

all these codes got the results but without a table

can anyone provide a code for the table itself ??

For anybody encountering this problem recently, TikTok released a new version of this API. A simple update of the URL might solve the problem:

https://www.tiktok.com/auth/authorize/

to

https://www.tiktok.com/v2/auth/authorize/

Refer to this blog [https://blog.51cto.com/u_16175468/8004089][1]

RUN ln -s /host/path/example.txt /app/example.txt

How to set date format to output date as "yyyy-MM-dd" in Azure Data Factory

Don't provide the default date format in the sink settings to get the expected output. Below is the process where i don't provide any default date format in the sink. So, the output is in expected format as you can see below.

Follow the below steps to get the expected output:

Step1:
Please try using the following expression without the default format to achieve the expected results: toDate(Date, 'dd/MM/yyyy', 'yyyy-dd-MM').

Step2: Based on your input, I have used the same sample data that you provided. Step3:
I have provided the following expression in the derived column as required. Please use the following expression in the derived column as required: toDate(Date, 'dd/MM/yyyy', 'yyyy-dd-MM'). Step4: I'm generating the output in CSV file format as per your requirements.

Turns out that when configuring Google sign-in on Firebase, there is an optional field: Safelist client IDs from external projects (optional). I added my client ID here and saved it, and everything is now working fine.

Since there are not many details about your code quality process, I will focus on your question:

Is it possible to somehow setup rules, to run branch pipeline only for MR and for default branch?

I think job rules may be something like this:

  rules:
    - if: $CI_PIPELINE_SOURCE == "merge_request_event"
    - if: $CI_COMMIT_BRANCH == $CI_DEFAULT_BRANCH
    - if: $CI_COMMIT_BRANCH && $CI_OPEN_MERGE_REQUESTS
      when: always
    - when: never

This set of rules uses GitLab environment variables like:

• CI_DEFAULT_BRANCH - true for your default branch (develop/main/master)
• CI_OPEN_MERGE_REQUESTS - true if pipeline has open merge request

Please experiment with them.

git config --global --unset credential.helper

this works for me

To reset or disable a specific user's 2FA, with version 16.2, (and maybe others), Login as Admin, pull of the user's profile in the admin interface, click where indicated in the screenshot, below. It will be a red button in the row labeled "Two factor authentication".

If applicable, change the user's password.