if ( $from === $to ) {
// Single day: Woo interprets this as "during that day in site timezone"
$args['date_created'] = $from;
} else {
// Range: Woo interprets this as "between those days in site timezone"
$args['date_created'] = $from . '...' . $to;
}
ALL SCRIPTS INTO my APP load:
web/config.php
-> this file define:
time ZONE
directory for images
ect...
ONLY when users OPEN:
web/tools_14
web/tools_21
web/tools_26
I need use the class NumberFormatter()
then:
is a bad practice ADD into {web/config.php} the var
$n2text = new NumberFormatter('eng', NumberFormatter::SPELLOUT);
???
if I do then the WHOLE web use Xram with this CLASS,
if I not do (my current practice) then I need call into each "TOOL" the class
in your opinion which is the more efficient?
( if we need something as "function_very_hard()" then we only ADD this function/class into the fragment code who require that, and not into "config.php"
Yep, super normal. The first WebClient call pays for DNS, SSL handshake, and spinning up the connection pool, so it’s slow. After that everything is already warm, so it flies...
Fix is basically: keep the WebClient as a singleton "wich you already do" and trigger a tiny warm-up call on startup so your users never see that 5 to 9s delay...
A simple search for "php test memory usage" returns the memory_get_usage function as its first result. There are even AI examples provided by some search engines to give you a head start (YMMV). There are also examples for you in the documentation itself.
Was there something in particular you were having trouble with to determine how much memory your scripts need?
You might want to check out some tools that simplify these kinds of Google Sheets customizations go to website for options that could make applying colored chips across dynamic dropdowns much easier.
Yep, TestStand nuked your types and there’s no undo for that.
Just grab the modified type-palette .ini or .seq files, restore them from Git, restart TestStand, and your custom types are back !
The security-related entities, e.g., definition of username, database roles, role memberships, permissions, are determined by a specific server configuration, and they do not belong to programming asserts, a.k.a., programmability of stored procedure, function, etc.
We have resolved this issue by removing the security-related entities from the SQL Server Database project.
We also configured Visual Studio's Schema Compare to ignore the following object types to prevent synchronizing on the security-related entities in future:
@Sergey A Kryukov thank you for the help, it's a male-only family tree, no wifes. It's a popular topic in the Arabic world and a historical science there, thanks again.
Best thing to do would be to just get the .className at runtime. Crappy fix, but it'll work. Personally I have issues trying to programmatically apply animations because it transforms the css name but not any other reference to it in the file. The keyframes become s-VSwDfhj2prWB-opacityvisibleanimation but my code is element.style.animation = "1s forwards opacityvisibleanimation", which fails. Even including it in the HTML segment fails, such as style="animation: {expanded ? "1s forwards opacityvisibleanimation" : "1s forwards opacityinvisibleanimation"}". I'm not sure how Rich intended us to reference animation names?
You'll need to do this at runtime.
in your .svelte: element.className.match(/s-[\s\S]{12}/g)[0]
Only the primary (left) mouse button can be used with SwiftUI's built-in gestures (TapGesture, DragGesture, LongPressGesture, etc.). They do not support right or middle mouse input.
The answer seems to be a resounding "No, not in modern FF" - thanks to @nanigashi for clarifying.
The best way to ergonomically toggle/access a setting that I've found was to pin an about:config tab and/or find a short identifying substring of the setting name to enter in the about:config search (i.i in the case of full-screen-api.ignore-widgets)
This appears to be a warning message basically stating that there was a tag within the SVG file itself that was not handled by the Svg Loader. I've found you can either simply ignore the message or remove the tag from the SVG.
Just to understand the question. You are using these two input elements (default is that they are text fields) as buttons, and you need the value of "action" (the name of the inputs) so that you can see if it was "Enter" or "Cancel" that was clicked when submitting the form? And then you need a localized display text for each button?
def apply_twice(func, arg):
return (func(arg)) # argument is 10 ,that added to to value 5
def add_five(x): # func when called once add_five inside becomes 10+5=15
return x + 5
print(apply_twice(add_five, 10))
def apply_twice(func, arg):
return func(func(arg)) # argument is 10 ,that added to to value 5 now called twice
def add_five(x):# func when called twice add_five inside becomes 10+5+5=20
return x + 5
print(apply_twice(add_five, 10))
def apply_twice(func, arg):
return func(func(arg))+ func(func(arg))# argument is 10 ,that added to to value 5 now called twice
def add_five(x): # two time repeated func , when called twice add_five inside becomes 10+5+5=20 +20=40
return x + 5
print(apply_twice(add_five, 10))
For those having issues even with tabIndex removed and all other normal solutions exhausted this is something to consider as it was an intermittent issue and hard to diagnose.
I use Yii2 and Kartik Select2 widget in a bootstrap modal which makes it hard to override the dedropdownParent setting. The widget worked perfectly in most situations anyways so I think he sets it correctly.
If the modal content container does not have a max-height and overflow-y:scroll style settings the select2 was not allowing search functionality when the modal content grew beyond the original modal size. After doing many tests and then some research on the symptoms its clear that there is a scroll detector function in the select2 initialization that breaks if it "thinks" the content should have a scroll even if it doesn't.
To fix this issue I set the modal-content max-height to 75vh and overflow-y:scroll. This way the select2 scroll detection will actually work and the search function will initialize.
Thank you for posting the result of your investigation, this was useful!
I'm coming from https://github.com/ocornut/imgui/pull/9099 (I am maintainer of that software) and things have been rather confusing but your latest comment helped connect some pieces of the puzzle.
I'd be ideally trying to solve this, aka allow MBCS-compiled apps using RegisterClassA(), DefWindowProcA() etc to work on all setups. I've been hoping I can avoid calling ImmGetCompositionString() (suggested by https://github.com/ocornut/imgui/pull/3653/files and implied by WM_IME_COMPOSITION ) in order to avoid requiring user to link my library with imm32.
So I am looking at:
\> To let your program work regardless of how the “language for non-Unicode programs” is set, you should handle WM_IME_COMPOSITION message properly. The simplest way to handle the message is to pass it to the DefWindowProcW() function.
However if i do this:
case WM_IME_COMPOSITION:
{
// Handling WM_IME_COMPOSITION ensure that WM_IME_CHAR value is correct even for MBCS apps.
// (see #9099, #3653 and https://stackoverflow.com/questions/77450354 topics)
IMGUI_DEBUG_LOG("WM_IME_COMPOSITION wParam %08X, lParam %08X\n", wParam, lParam);
return ::DefWindowProcW(hwnd, msg, wParam, lParam);
}
When you get a divison by zero error/ exception it boils down to a rounding issue and how numbers are technically handled (machine representation of numbers). It depends how you calculate the number e (i.e. a by an approximation function) and it depends how you exponentiate. For exponentiation and calculating roots (nth root) there exists approximation functions that contain divisions. So rounding is the main culprit here.
That's the difference between theory and concrete implementation.
I was trying to solve the wrong problem. Just in case anyone else has the same misconception as me: body-parser is not needed. Inversify handles the parsing.
Yeah, the adapter adds a json body parser by default. It can be disabled by passing useJson option flag to false.
My issue ended up being a typical frontend issue when the form was submitted (I was trying to submit JSON using a plain HTML form).
If you were sending a application/x-www-form-urlencodedContent-Type request, yeah, that's tricky, for express silently ignores the body unless you set a parser for it. If you are interested in supporting urlencoded bodies in the future, useUrlEncoded flag can be handy.
Not really. You can, but you'll end up secretly Church-encoding lists or something. You are asking how to equip the type of rose trees with the eliminator for lists; conceptually, the "only" way to do this is to factor through lists.
Just moving the partials SCSS files into a subfolder worked. This was solved thanks to this answer. Check it for more information. This could be marked as duplicated.
The bit in Bold is the interesting bit, in that "Wwise<space>"wildcard.wildcard allows me to copy only the files starting with Wwise<space> to the location desktop. I am sure there are more ways of messing wit this, but I stopped after getting it to do what I wanted.
If you're talking about getting syntax highlighting in your markdown file, this has to do with how you're rendering it. The renderer will interpret that language, and some renderer's have more languages implemented than others.
Your thought process and explanation seem logical to me. What exactly is the issue or concern you have here? What kind of answer or suggestion or advice are you looking for... specifically?
has no non-static data members of type non-standard-layout class (or array of such types) or reference,
has no virtual functions and no virtual base classes,
has the same access control for all non-static data members,
has no non-standard-layout base classes, only one class in the hierarchy has non-static data members, and
So depending on how std::vector is implemented, all of those may very well be true. But it has nothing to do really with being PoD.
Different combinations of other traits may be used to check for PoD-like classes, lik the is_trivially_xyz family (ie. is_trivially_copyable).
Yet, from my work on an ASM-compiler, that needed to interopt with ASM/classes generated by MSVC, I think I remember there is no true trait to give you a true "pod" definition of a type, in a way that the compiler sees it - my example being the definition of "return values" in the x64 calling convention - here, a combination of is_trivially will not necessary align with what the compiler sees as pod, and I did not found a reliable way to detect it (even after asking stackoverflow).
Regardless, this might be a very specific example. In most cases, checks for trivially may be more than enough. For a lot of cases, it will be enough (and more correct/expressive) to just check for a specific property, instead of making your own "trivial" combined trait (for example, in template code that does copy of a value, just checking for trivially_copyable would be the most appropriate trait.
Reasons:
Blacklisted phrase (1): stackoverflow
RegEx Blacklisted phrase (2.5): Can someone explain
I'm facing the same issue at the moment and i have done the same as you did, but still Google Play rejecting my app. If you came up with a solution for this matter would u please share it with me !
Thanks in advance
Chrome recently introduced a restriction (Chrome 104+) that caps the maximum cookie lifetime to 400 days.
So even if you set a longer expiry (like 3 years or 2038), Chrome will silently shorten it.
This limit applies to both expires and max-age directives.
Older cookies created before v104 may still keep their original dates, but any new cookies are forced under the 400-day rule.
I've just ended up here with a similar issue. Not sure if it's the same problem that the OP had or a new one but I thought it might be helpful to record it here anyway. My issue was related to this:
https://github.com/rstudio/rstudio/issues/16454
Running Sys.unsetenv("PKG_CPPFLAGS") and then installing rbibutils worked fine. Ubuntu 24.04.3 & RStudio 2025.09.2
It depends on the stack used for a product. A frontend can be rendered in the browser using React/Angular/Vue.js/etc or server-side using frameworks like Wicket/Velocity etc.
Depending on the frameworks used it can be quite useful to know a little about back-end tech. I've done projects where the front-end developers had no back-end experience which made in depth discussions very difficult.
IMO a developer should at least have minimal knowledge of tech used in a project no matter what their work area is. This makes discussions easier and gives everyone the ability to chip in when needed.
@Valnik You can remove the UNIQUE (col1, col2) constraint. Just the index is enough; it handles all combinations. The only thing to check is how to handle NULLs, should they be unique or not?
I am writing a scraper in Python and ran into a CAPTCHA on the login form. The site sometimes returns a regular reCAPTCHA v2 and sometimes hCaptcha. I tried to bypass this through Selenium, but CAPTCHA keeps blocking the automation.
My purpose is to automatically obtain the CAPTCHA solution token and insert it into the form so that the script can continue running without manual involvement.
What I have already tried: headless Chrome, different user agents; waiting for the iframe to appear; tried recognizing the CAPTCHA image through third-party libraries, but reCAPTCHA/hCaptcha do not work this way; found mentions of external services, but I do not understand how exactly to integrate them with Python via API.
What I need:
A step-by-step recommendation on how to technically organize CAPTCHA solving in Python: which API to use; how to create a task, obtain the token; how to pass it back into Selenium or the request; whether a universal method exists for different CAPTCHA types.
I would appreciate an example of working code or a proven architecture.
As already mentioned by @VLAZ,
That is actually not a runtime error. that is a transpile time error.
you declared:
type TParser = (str: string, errStr: string) => number;
which is actually wrong. since there is a default parameter, we need to make errStr optional.
type TParser = (str: string, errStr?: string) => number;
notice the added ? before : string,that is used to indicate the parameter is optional, you can give a value to it, or you may not, typescript will allow you.
When setting cookies with strict mode, cookies are not set on redirecting. So, in order for the cookies to be set, the page needs to be viewed to the end-user in the browser allowing the cookies to be set.
An approach to this is to redirect using a form post instead of 302 redirect. By rendering an actual page at the end of the authentication process, you're breaking out of the redirect sequence allowing the browser to set the cookie:
while True:
user_input = input("Enter an integer: ")
try:
value = int(user_input)
break # input is valid → exit loop
except ValueError:
print("Invalid input. Please enter a valid integer.")
Here, this is what I use on a daily basis when I try to avoid online services such as Github, Gitlab... Of course when I don't need the CI/CD and the bells and whistles such as actions, accounts, etc.
This answer is for future reference regarding how to achieve a full Git workflow without relying on a remote Git hosting service (like GitHub, GitLab, etc.), by using a local bare repository as your "remote."
💻 Full Git Experience with a Local Bare Repository
You can replicate the standard Git workflow (clone, commit, push, pull, etc.) entirely on your local machine by setting up a bare repository to serve as your central, shared history.
1️⃣ Setting Up the "Remote" (Bare Repository)
First, create a bare Git repository in a designated, central location on your machine. This directory will not contain a working tree (i.e., no files, just the Git history).
Pick a directory for your bare repo (e.g., in your home directory).
Bash
mkdir ~/my-project-bare.git
cd ~/my-project-bare.git
git init --bare
2️⃣ Setting Up the Working Copy (Local Clone)
Now, treat that bare repository path as the "remote URL" and clone it into your actual working directory.
Move to a different location where you want your project files to reside.
Bash
cd /path/to/my/development/area
Clone the local bare repository using its file path.
Bash
git clone ~/my-project-bare.git my-project-working
cd my-project-working
3️⃣ Standard Workflow
From this point on, all Git operations are exactly the same as if you were using an external remote.
Configure your identity (optional, but good practice):
# Make changes to files...
git add .
git commit -m "Initial commit of the project setup"
git push origin master # or main
The ~/my-project-bare.git directory now acts as your central source of truth. If you want to simulate multiple developers or different machines, you can simply clone the bare repo again into a different folder (e.g., my-project-testing), and use git pull and git push between the working clones and the bare repo.
@abaci Thanks...I think the Chiclet Slicer is a glorified slicer, isn't it? Couldn't find any 'HTML Content Custom Visual' but I did look at 'get more visuals' for HTML-related visuals. Not sure why HTML would be useful here?
Am I wrong to imagine that a Power BI report would be a useful (better) substitute for SharePoint gallery view?
Reading documentation is the best way to learn a tool in my experience, however I often struggle to focus on reading, and when I do I find YouTube videos explaining the tools to be invaluable. Combining both, and doing a bit of experimentation on your own I think is the most productive way to learn things. AI is great for specific niche questions that you haven't been able to solve, but make sure not to get dragged into the vibe-coding pit.
Getting started on a project using a new technology can be very daunting. I find the best way to start is by watching a tutorial, following along initially, and once the tutorial is done, do it again without help. You'll learn what you don't understand, and then you can go back and read the documentation or rewatch the video to better understand the parts you don't get.
It's a slower process that AI, but it means you'll actually learn and understand what you're doing, which is far more important than speed.
You need to tell pytest-django to keep the data loaded in initial_db for the duration of the session, or ensure it's committed in a way that survives the test rollback.
Here is the correct way to implement Session-scoped Data Fixtures with pytest-django.
That's a very specific and challenging requirement for iOS automation!
Unfortunately, due to the tight security sandbox Apple enforces, there is no direct, public, or officially supported way for Appium/WebdriverIO (or any other standard automation framework) to programmatically toggle Airplane Mode or Mobile Data settings on a non-jailbroken physical iOS device or even a standard iOS Simulator.
The specific problem or situation you are trying to address. Knowing the context (e.g., applying for a loan, filing taxes, submitting a project proposal) will help determine if any documents are missing or if the current ones are sufficient.
I sure do love reading “Microservices: The Docs” or “How to Build an Extremely Niche Feature That You Can Barely Describe: The Wiki.”
All jokes aside, I prefer to look for trends and see what people actually use for what problems.
Docs can explain how to do things and sometimes why, but they rarely offer the kind of wisdom that matters in real problem solving, which what engineering is imo, for example, why people building low latency, high throughput services choose X language and Y framework.
I use ClickHouse because of a very specific requirement, and I didn’t discover it by reading its docs. I found it by reading the experiences of people who use it. Blogs and forums are good; rabbit holes are even better.
in your current terminal session, you do need to unset this environment variable when switching to Google AI Studio’s API (non-Vertex AI). Otherwise, the SDK will keep working in Vertex AI mode.
If you are using the latest Flutter version, then there is one major issue currently open related to resizeToAvoidBottomInset broken on Android. Keyboard breaks layout. So, you can track and get the workaround for the issue.
@jcalz Yeah, I hand wrote the code, so it's wrong, but you got the right point. And also the issue you posted is also what answers my question. Thanks!
You could use a label tag and show text in Spanish. When the text in the label tag is clicked, it refers to a hidden button whose value is in English. The submit button is hidden using CSS. Below is a simple example:
