How to switch between Light, Dark, and Tinted appearances in the Xcode Simulator (iOS 18+)
If you're looking to test your app icon or UI under different system appearances in the Xcode Simulator, here's how to toggle between modes. This is especially useful when checking how your app icons or assets behave in Light, Dark, or the new Tinted mode (iOS 18 only):
Steps:
Long press on an empty space on the Home Screen or on any App Icon.
Tap the Edit button in the top-left corner (occasionally appears top-right depending on Simulator version).
Tap Customize.
Choose between Light, Dark, Automatic, or Tinted.
š Note: Tinted mode is only available starting in iOS 18.
Collect 3 or more thread dumps from the driver and any active executors, spacing them about 5 seconds apart. This should help identify where the Spark job is getting stuck.
Well, I'll be damned... I used to send it to myself using WhatsApp web, then I just open the same chat on my phone, download the file and click it again to execute.
I tried a different method (thanks, @CommonsWare), just plugging the phone and using the file explorer and it works.
My best guess is that my PC does something to the file when sending it through WhatsApp web. Something that my work PC does not do for some reason.
I had a similar issue, and I think I managed to solve it by increasing the value ofnet.netfilter.nf_conntrack_udp_timeout_stream using sysctl. It defaults to 120s, so this may be the setting that causes the timeout error in spark.
If you wound up here due to having perfectly good JSON but you still get the error above, I have a question for you:
Did you transfer a file from a Windows to a WSL context?
Because if you did, that file is going to be shredded moving from the Windows Ansi /r/n context to the linux Unicode /n standard. If you are using VScode or something, copying the contents of the file over should be enough to get it to convert properly.
The solution for this highly specific problem is to not install the "minidriver" part of the SafeNet Authentication Client package . If already installed, uninstall the entire package, then run the installer again, choose "Custom" install and make sure the Minidriver feature set is set to "don't install" (The SafeNet installer doesn't offer a "Modify" option).
This is apparently because the "minidrivers" are the ones that allow the badly designed Windows Smart Card logon system to talk to the SafeNet USB smart cards. With the minidrivers removed, all access is through the SafeNet extensions to the CryptoAPI 2 subsystem that is used by signing tools (including old tools based on the classic CryptoAPI 1).
Some one found a solution to this problem? I access a windows machine via remote desktop and the problem only show on Delphi editor. The object inspect work fine.
When VS Code created the mcp.json file, it was created as a user config file on the Windows side by default: /C:/Users/spencer/AppData/Roaming/Code/User/mcp.json
The file needs to be in your workspace: /{path to project}/.vscode/mcp.json
The parameter set file .SSV of the SSP standard (a sibling standard to FMI) is intended for this (https://ssp-standard.org/)
If your importing tool supports the SSP standard, you can put the FMU together with a n .SSV parameter file in and SSP file.
The FMI project works on a Layered Standard FMI-LS-REF (https://github.com/modelica/fmi-ls-ref) you will also be able to put one or severalĀ .SSV files in an FMU at a defined subfolder the /extra directory of an FMU.
Yes, lv_timer_handler() is definitely being called. But thanks to your information, I discovered that my LVGL initialization only called lv_init() but did not set a tick source.
I now also call lv_tick_set_cb([]() -> uint32_t { return (uint32_t) millis(); }) during initialization.
As a result, my image/JPEG is now drawn correctly even without calling lv_refr_now(NULL).
In the vendor demo, I couldn't find where the tick is set, but I suspect it's done via lv_conf.h.
I also had another issue: I had initialized the display with the wrong frequency (18MHz), which I had taken from another GitHub (espidf) example.
But in the official vendor demo, I saw that 14MHz is used instead.
Now, with the correct frequency, my display and slideshow run smoothly and without flickering. Great. Thanks!
UserName = input("Inany da Silva Serra: ")
passw = input("Search the password: ")
check(Inany, passw)
return UserName, passw
def check(Inany, password):
pword = {}
for line in open('unames_passwords.txt','r'):
user, password = line.split()
pword\[user\] = password
if user == UserName and password == passw:
return True
print("Thank you for logging in.")
else:
print("Username or password is incorrect")
def main():
login()
main()
Maybe you are using the cart block on your cart page while customizing and overriding the classic cart template. If you want to customize your cart template, you'll need to replace the cart block with the cart shortcode.
In this extended scenario, the sub-sampling is no longer over contiguous memory blocks, so the data streaming approach will also require script-level looping and will probably not be particularly efficient.
On my setup, looping over Slice2 calls is still ~20% faster than the the single call to ExprSize, but that may be because my computer has an old processor. I did notice, however, that the timing results varied a great deal and seemed to be connected to foreground activity in support of the Display floating window. For optimum and consistent timing results, I think it is important to either delay calls to ShowImage until the very end of the script, or to make sure the Display floating window is closed. Of course, these issues are side-stepped entirely if one runs the script on the foreground thread, instead of in the background, as you currently seem to do.
The other answers have adequately addressed your issue, however I would like to share a novel approach towards the goal of "cleanly acquiring a collection of mutexes". Due to the nature of the synchronized block in Java, it's not feasible to acquire several mutexes in turn (the loop would essentially need to be unrolled).
Object[] mutexes = new Object[4];
for (int i=0; i < 4; i++) mutexes[i] = new Object();
synchronized (mutexes[0]) {
synchronized (mutexes[1]) {
synchronized (mutexes[2]) {
synchronized (mutexes[3]) {
// all acquired
}
}
}
}
However, if you look at the resultant bytecode, you'll see that each synchronized block is opened with a MONITORENTER instruction and explicitly closed with a MONITOREXIT instruction. If we had direct access to these operations, we could iterate once to enter each monitor and then iterate again to exit each monitor. Is it possible to compile valid Java code that does this? Sort of.
public final class MultiLock {
public static void run(Object[] mutexes, Runnable task) {
monitorEnter(mutexes);
try {
task.run();
} finally {
monitorExit(mutexes);
}
}
private static native void monitorEnter(Object[] arr);
private static native void monitorExit(Object[] arr);
}
#include "MultiLock.h" // Header generated by javac
#include <stdlib.h>
#include <stdbool.h>
static inline bool is_valid_monitor(JNIEnv *env, jobject object) {
return object != NULL;
}
JNIEXPORT void JNICALL Java_MultiLock_monitorEnter(JNIEnv *env, jclass ignored, jobjectArray arr) {
jsize len = (*env)->GetArrayLength(env, arr);
jobject next;
for (jsize i = 0; i < len; i++) {
next = (*env)->GetObjectArrayElement(env, arr, i);
if (!is_valid_monitor(env, next)) continue;
(*env)->MonitorEnter(env, next);
}
}
JNIEXPORT void JNICALL Java_MultiLock_monitorExit(JNIEnv *env, jclass ignored, jobjectArray arr) {
jsize len = (*env)->GetArrayLength(env, arr);
jobject next;
if (len == 0) return;
for (jsize i = len - 1; i >= 0; i--) {
next = (*env)->GetObjectArrayElement(env, arr, i);
if (!is_valid_monitor(env, next)) continue;
(*env)->MonitorExit(env, next);
}
}
And use it like so:
// Load the natives somehow
Object[] mutexes = new Object[4];
for (int i=0; i < 4; i++) mutexes[i] = new Object();
MultiLock.run(mutexes, () -> {
// all acquired
});
See step 2. Create a passwordless connection. As the tutorial mentions, I used the Azure Portal to automatically create the connection for a system-assigned managed identity.
in case this happens to anyone else, strangely it was the variable names. changed PUBLIC_SUPABASE_URL and PUBLIC_SUPABASE_ANON_KEY to VITE_PUBLIC_SUPABASE_URL and VITE_PUBLIC_SUPABASE_ANON_KEY and it works. Weird bc I believe the svelte docs say that PUBLIC_... should work.
It is quite interesting actually the Google and Facebook uses it for faster string manipulation purposes the performance of fbstring is better so using of normal std:: string given by c++ is replaced for better performance.we create a file string and in that we include <folly/FBString.h> so when we call std:: string the backend fbstring will work it helps to boost the performance.it simple terms it is like overriding( remember don't question me not having same name and parameters) just explaining in conceptual way for std:: string we are writing our own backend code where it helps to boost our performance but be careful to placing the folder and tell the cmake to include before system include.your aim is to compiler need to give priority to your string file
This was user error. In the MAC I had to set the channel number correctly in the Sniffer dialog. Once I did that then it worked fine. Also in the Wireshark->Protocol->IEEE802.1 edit decryption keys, the password along with the SSID had to be entered as follows <password>:<ssid>
I think the requestMatchers have a higher priority than the method annotations.
I also encountered the same problem. I wanted to validate all resources under the /api/** path, but adding @PermitAll at the /api/abc path was ineffective, and even changing to @PreAuthorize("permitAll()") also didn't work.
Yes, absolutely. Gioui is specifically designed for cross-platform mobile development and works well with Go mobile for building Android apps. You can write your UI in Gioui and use gomobile bind or gomobile build to compile it into an Android APK. This is a fully supported and documented approach.
This is a very elegant solution but I do have one problem: although the final print generated to either the printer or to preview renders the headers/footers correctly the view of the pages in the print preview dialog shows the headers/footers upside-down.
If you run the container once with a different password and then restart it with another password, it wonāt update the existing database because itās already initialized. Try removing the old volumes or using a different volume.
So we need to remove values from table authorization_role column gws_websites and gws_store_groups based on the deleted stores.
Hi Abbas, I do not have gws_websites and gws_store_groups columns in my autorization_role table. I am on Magento 2.4.7-p6. Would you know where this info would be in newer version of Magento. I am using Community-edition.
With some regex engines {-1,} is a possible "lazy" quantifier, another option sometimes is to provide a flag for the pattern, signifying that it should default to "lazy" matching, instead of "greedy".
Yes, you can absolutely create a compelling hierarchy tree from a nested <ul> structure using pure CSS! It's a common and fun challenge. Your current approach is already very close; the key is to fine-tune the positioning and dimensions of the pseudo-elements (::before and ::after) to draw those connecting lines accurately.
Here's a refined CSS solution that should give you the desired command-line tree look, along with an explanation of the adjustments.
HTML
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>CSS Hierarchy Tree</title>
<style>
body {
font-family: monospace; /* For that terminal-like feel */
color: #333;
padding: 20px;
}
ul {
list-style: none;
margin: 0;
padding-left: 20px; /* Adjust as needed for initial indent */
position: relative;
}
li {
position: relative;
padding-left: 20px; /* Space for the horizontal line and text */
margin-bottom: 5px; /* Small space between items */
line-height: 1.5; /* Vertical spacing for text */
}
/* Horizontal line for each item */
li::before {
content: '';
position: absolute;
top: 0.75em; /* Adjust to align with the text baseline */
left: -5px; /* Pull the line slightly to the left to connect with parent's vertical line */
width: 15px; /* Length of the horizontal line */
height: 0;
border-top: 1px solid #999;
}
/* Vertical line for connecting child branches */
li::after {
content: '';
position: absolute;
top: 0.75em; /* Start from the same height as the horizontal line */
left: -5px; /* Align with the horizontal line start */
height: calc(100% + 5px); /* Extend beyond the current li to connect to siblings/next parent */
border-left: 1px solid #999;
}
/* Hide vertical line for the last child in a branch */
li:last-child::after {
height: 0.75em; /* Only extend to the horizontal line of the current li */
}
/* Remove the initial horizontal line for the very first item (if desired) */
ul:first-child > li:first-child::before {
border-top: none;
}
/* Remove the initial vertical line for the very first item (if desired) */
ul:first-child > li:first-child::after {
border-left: none;
}
/* Style for the nested ULs to create proper alignment */
li > ul {
padding-left: 20px; /* Indent for nested lists */
margin-top: 5px; /* Adjust spacing between parent and child UL */
}
/* Special handling for the vertical line coming *down* from a parent */
li:not(:last-child) > ul::before {
content: '';
position: absolute;
top: -5px; /* Align with the bottom of the parent's vertical line */
left: 0;
height: 10px; /* Length of the connecting vertical line */
border-left: 1px solid #999;
}
</style>
</head>
<body>
<ul>
<li>1830 Palmyra
<ul>
<li>1837 Kirtland
<ul>
<li>1840 Nauvoo
<ul>
<li>1841 Liverpool
<ul>
<li>1849 Liverpool
<ul>
<li>1854 Liverpool
<ul>
<li>1871 SaltLakeCity
<ul>
<li>1877 SaltLakeCity</li>
</ul>
</li>
</ul>
</li>
</ul>
</li>
</ul>
</li>
<li>1842 Nauvoo
<ul>
<li>1858 NewYork
<ul>
<li>1874 Iowa
<ul>
<li>2013 SaltLakeCity</li>
</ul>
</li>
</ul>
</li>
</ul>
</li>
</ul>
</li>
</ul>
</li>
</ul>
</li>
</ul>
</body>
</html>
What's Changed and Why?
Let's break down the key adjustments and the thought process behind them:
ul Padding:
padding-left: 20px; on the ul is crucial. This creates the initial indentation for each level and gives space for the vertical connecting lines.
li Positioning and Padding:
position: relative; on the li is still essential for positioning its pseudo-elements.
padding-left: 20px; on the li provides the necessary spacing for the ::before (horizontal line) and the visual "indent" for the list item's text.
margin-bottom: 5px; and line-height: 1.5; are for better visual spacing between list items.
li::before (Horizontal Line):
top: 0.75em;: This value is critical for vertically aligning the horizontal line with the text of the list item. 0.75em often works well to place it roughly in the middle of a typical line of text.
left: -5px;: This slight negative left value pulls the horizontal line a bit to the left. This allows it to overlap and connect perfectly with the vertical line coming down from the parent li (or the vertical line of a sibling li's ::after pseudo-element).
width: 15px;: This defines the length of the horizontal line extending from the vertical connector to the list item's text. You can adjust this to control how far the "āāāā" part extends.
li::after (Vertical Line):
top: 0.75em;: Similar to ::before, this ensures the vertical line starts at the same vertical position as the horizontal line.
left: -5px;: Aligns it precisely with the horizontal line's starting point.
height: calc(100% + 5px);: This is a significant change!
100% makes the line extend to the bottom of the current li.
Adding + 5px (or a similar small value) makes it slightly overshoot the current li. This overshoot is vital for connecting neatly with the next sibling li's horizontal line. Without it, you'd often see a small gap.
li:last-child::after: For the last child in a branch, we don't want the vertical line to extend indefinitely. Setting its height to 0.75em (or top's value) makes it just reach the horizontal line, effectively stopping the branch.
These rules are optional but highly recommended. They remove the horizontal and vertical lines from the very first list item (1830 Palmyra in your example). This makes the tree start cleanly without extraneous lines.
li > ul Indentation:
padding-left: 20px; on li > ul ensures that nested ul elements are further indented, creating the visual hierarchy.
margin-top: 5px; adds a little breathing room between a parent item and its nested child list.
li:not(:last-child) > ul::before (Connecting Vertical Line for Nested ULs):
This is a crucial addition to correctly render the vertical line between a parent item and its child ul when the parent is not the last child itself.
It creates a short vertical line (height: 10px;) that effectively bridges the gap from the parent's li::after down to the ul's content, maintaining the continuous vertical branch.
top: -5px; adjusts its position to connect seamlessly.
Tips for Customization
Line Thickness & Color: Easily change border-top and border-left values for 1px solid #999 to 2px dashed #007bff or whatever suits your design.
Spacing: Adjust padding-left on ul and li, margin-bottom on li, and margin-top on li > ul to control the horizontal and vertical density of your tree.
Em vs. Px: Using em for top values and px for width and height gives you a good balance. em values scale with font size, which is good for vertical alignment with text, while px gives precise control over line lengths.
Accessibility: While this is a visual representation, remember that the underlying HTML <ul> structure is inherently accessible and semantically correct for lists. This CSS merely enhances the presentation.
With these adjustments, you should achieve a much cleaner and more accurate visual representation of your timeline/tree structure using only HTML and CSS. Give it a try! You'll love the results.
You can find various search parameter listed in YouTube API v2.0 ā API Query Parameters like license, restriction, paid_content that can help filter videos that are restricted for such specific reason. Also, if you can use YouTube API v3.0 there is one more option videoSyndicated that will restrict a search to only videos that can be played outside youtube.com.
It needs to be fixed on the Rest API where the double quotes are incorrectly added.
The owner(s) of the project need to change it. Contact them about it so that they are aware and can fix it.
If they cannot/don't fix the issue in a timely manner, and you are not forced to use this Rest API, try find a more stable library for Rest instead.
Its a bad idea to try cater for buggy data by modifying it with more code later on. It just adds unnecessary code, and opens the door to unexpected behavior and accidental data modification.
You cannot enforce a range on content length. The way you limit the file size is by allowing the client to specify the desired length when requesting the presigned URL, and if the desired length is unacceptable, you just don't give them the presigned URL but error out instead. If acceptable, then you create the presigned URL with "ContentLength": desired_length as a parameter instead.
This was user error. I had to set the LWIP_TCP_KEEPALIVE in the compile options like -DLWIP_TCP_KEEPALIVE. Once I did this then I did not get any errors when setting the options.
A lot of years are gone but very likely the issue could led to how an https request is done using Winhttp or in general, HttpSendRequest.
After the Certificate exchange and encrypted handshake message Windows will try to verify if what has been received is valid.
To do this, it first check certificate in "Trusted Root Certification Authorities" and in case of failure will start to "retrieve Third-Party Root Certificate from Network".
So a called to DNS and external address is performed. The problem is that in some environment maybe the calls are dropped and so your https request get stacked until a timeout.
The timeout should be 15 seconds and then the request is unlocked.
This behaviors it's completely independent from the options you can set on the HttpSendRequest about ignoring certificates because this action will be execute only later in time.
Knowing the request workflow there can be multiple way to fix it.
basically set at HKEY_LOCAL_MACHINE\SOFTWARE\Policies\Microsoft\SystemCertificates\AuthRoot:
"EnableDisallowedCertAutoUpdate"=dword:00000000
"DisableRootAutoUpdate"=dword:00000001
Another way is to fix the certificate, maybe a self signed and add it correctly onto the windows certificate storage at "Trusted Root Certification Authorities" at machine level certificates.
The 15 seconds in real are a default value that can be override from local group policy:
You've described a classic case of āsame code, different behaviorā due to platform changeāand that can be tricky to pin down. Letās break it down and explore possibilities:
š§ What Could Be Going Wrong?
š“ RHEL vs. Alpine: Base Image Implications
Alpine Linux is lightweight and uses musl libc, while RHEL 9 uses glibc and has more extensive background services.
The .NET Core runtime may behave differently due to dependencies compiled against glibc. Memory allocation strategies and garbage collection patterns can vary.
RHEL ships more diagnostics and background processes that can subtly increase memory footprint.
š¦ HttpContext / Redirect Behavior
Your use of HttpContext and Redirect() looks correct and typically shouldn't hold memory.
But if the redirect endpoint is being hit very frequently, and responses are cached or buffered internally, memory can creep.
Possible culprit: server-side buffering, e.g., response streams not being disposed properlyāespecially if middlewares interact with HttpContext.
š¬ Diagnostics You Can Perform
Here are ways to isolate the issue:
ToolUse Casedotnet-countersTrack memory, GC, threadpool, HTTP counters in real-timedotnet-gcdumpSnapshot GC heap and analyze retained objectsdotnet-traceCapture traces to explore whatās allocating/proc/<pid>/smapsCheck actual memory usage per process, native vs managedK8s metrics (Prometheus + Grafana)Trend analysis over time per pod
š Code-Level Things to Watch For
Async Pitfalls: Though your endpoint is marked async, you donāt await anything. Consider dropping async unless neededāit may affect thread use.
Middleware or Filters: Look at upstream middlewares or filters that might buffer HttpContext.Response.
Logging: Excessive logging on redirect calls can gradually consume memory if not batched/flushed.
Try rolling back to Alpine with .NET 8.0 and compare memory diagnostics side by side with RHEL.
Consider building a minimal service that replicates your redirect pattern. Run identical traffic against both container bases and capture GC/memory snapshots.
Tune GC using environment variablesāe.g., set DOTNET_GCHeapHardLimit, DOTNET_GCHeapLimitPercent.
This issue isnāt likely caused by one line of codeāitās the interaction of the runtime with the new OS environment. Want help analyzing a memory dump or building test scaffolds to narrow it down? Iād be glad to collaborate.
A comprehensive, cross-platform React Native wrapper for secure key-value storage using native security features of Android and iOS. It supports biometric authentication, hardware-backed encryption, and deep platform integrations such as Android StrongBox, EncryptedSharedPreferences, and iOS Secure Enclave via the Keychain.
There are the lib: rn-secure-keystore
Same for me, quasar is not adding the styles with safe insets, fixed it temporarily by just adding the iPhone inset utility class to the body, until a proper fix is out.
The included codes from files are macro facilities of the C language that allow both single and multiple inclusions, as can be seen in the following example. It does not necessarily to have a specific file name extension such as ".h" unless we want to indicate a code with header content, etc.
Because using Property Let, Get and Set, we can have the same name for the read and the write method. Usually in a programming language your function identifier must be non-ambiguous. You can't use the same name in one specific scope. Using Property, the "visible" name is the same and the "context" of the code will "decide" what function is called. I believe that under the hood VBA will name the functions with different names.
In CSS, elements inside a flex container can indeed shrink in a specific order using the flex-shrink property in combination with the natural flow of the layout. However, there's no direct property like "shrink-order"
flex-shrink sets the relative rate at which an element shrinks when there's not enough space in the flex container. A higher value means the element shrinks faster or more compared to others.
also it will only work when the container has display:flex;
You can do client side, on the condition that you can fetch the Icecast stream.
To make client-side playback and ICY metadata extraction work via fetch() in the browser, CORS (Cross-Origin Resource Sharing) requirements must be properly met by the radio stream server.
I wrote the @music-metadata/icy module for this occasion. Credits to Brad who encourage me to contribute to StackOverflow, while others gave me much reason to run away.
Tools such as Valgrind and Clang memory sanitizer (clang++ -fsanitize=memory) can check for reads of uninitalised memory (valgrind warns you by default). Valgrind is a runtime detector and clang is a static analyser, so you'll probably want the latter. The GCC toolchain does not offer an equivalent tool as far as I know.
I did something similar in the past. When you have to manage a CAN device which has custom interface via SDO (some special SDO for example), it's better to manage it formatting your message directly in PLC code. This can be done using the object CAN interface which gives you a process image where you can place your custom messages.
are kinda mutually exclusive. See io.micrometer.core.instrument.distribution.DistributionStatisticConfig#getHistogramBuckets
Specifying 'slo buckets' will actually create them as requested, BUT these will be burried under lots of default ones, which are created upon enabling precentiles-histogram. There is list of 275 default buckets pre-created and we select subset of them based on expected minimum and maximum duration. By default (io.micrometer.core.instrument.AbstractTimerBuilder#AbstractTimerBuilder) these are 1 millisecond and 30s respectively. Which you can override using org.springframework.boot.actuate.autoconfigure.metrics.MetricsProperties.Distribution#minimumExpectedValue.
I don't understand this sufficiently, and this precision might be needed for some usecase. But if you need just if something is slower than some threshold (and mostly if smth is slower than 1s, it's bad regardless of how much), it might be safer just to specify slo thresholds.
If I'm still missing something or am wrong altogeher, please let me know!
Reasons:
RegEx Blacklisted phrase (2.5): please let me know
I faced the same issue where the first connection attempt or a connection after a longer period would immediately time out. From my observations, the first connection attempt always takes significantly more time. Increasing the timeout, pool size and pooling in the connection string helped in my case. I added these parameters: Timeout=60;Command Timeout=60;Pooling=true;MinPoolSize=1;MaxPoolSize=200
The process of migrating TRX tokens is described in the official documentation on GitHub:
1. Access Your Wallet: Check your TRX balance in the list of Ethereum tokens.
2. Send Tokens: Send your old ERC20 TRX tokens to migration smart contract.
3. Confirmation: After the transaction is confirmed and broadcasted to the blockchain, the migration process will start, and your wallet will be credited with new TRX tokens. Depending on network usage, it may take 5ā30 minutes to process the migration.
1-Create HLS chunks where the first chunk is 2 seconds, the second chunk is 3 seconds, the third is 5 seconds, and all subsequent chunks are 5 seconds each.
2-Cache the videos using fs, also delete them when seen.
I am coming at this from a math/physics angle rather than programming, so forgive me if I am focusing on the wrong thing, but I need some clarification on what exactly you are trying to transform here. Are you trying to preserve the ratio of the scale length to the point-scale distance?
By point-scale distance, I mean the length of the perpendicular line/arc from the point to the closest point on the scale. My understanding is that you are trying to satisfy:
It's known from differential geometry (see Theorema Egregium) that you cannot project from a sphere to a plane while preserving both shapes (or angles) and areas, which I suspect is very likely to be the root cause of your problems. I am not really sure if what you are trying to achieve can be done by only preserving one or the other or if you're trying to do something impossible, but it's probably worthwhile to actually carry out the math in 3D rather than a 2D projection. The (two ends of the) scale and the point together form a triangle on the Earth's surface, so you're really trying to transform (rotate, scale, translate) a spherical triangle, which I am not sure would work. Spherical trigonometry might help you here.
Likely why your 2D approximation does not work
The transformations you're composing are:
translation (cyan -> magenta)
scaling (ends up in green point)
rotation (takes the point in question from green point to pink point)
Using regular 2D/3D cartesian coordinates, these operations do not commute. Chiefly, the reason for this is that translation is not a linear transformation in Cartesian coordinates (Tiny proof sketch: The 0 vector does not map back to 0 under a translation). In other words, you'll get a different result if you change the order. In general, you'll apparently need to use homogenous coordinates, under which translation is linear, to avoid this problem; however, you might end up working with points/lines/areas off the surface of the Earth if you directly convert Cartesian coordinates to homogenous coordinates in this case. I cannot guess offhand if the approximation would work better with homogenous coordinates or not.
An immediate solution, if possible for your use case, is to uninstall react-native-ui-lib. This is the solution suggested in these react-native issue threads:
Otherwise you can try to patch react-native-ui-lib using patch-package, and there's a little in the above threads about how to do it, or wait for react-native-ui-lib to push an update that will fix the issue.
I'll add this here, since it's one of the top results from Google for the question:
Yes, the keyword is $Prompt$ in any of the parameter fields of the Run Configs, like VM Option, program options and so on. With $Prompt:Type your birthday$ one can even add a label to the prompt.
I didn't find a solution for a multiple choice / dropdown solution though.
It doesn't seem like opt can be piloted to produce the output you would like, so most likely you will want to post-process its output -- the heavier alternative would be to write a C++ LLVM module.
There may be several ways to go about this. Graphviz files can reference one another (e.g. this is how OTAWA handles it), so you can have the callgraph file referencing each of the CFG files.
If this isn't satisfactory and you must have a single visual, you will want to merge the files indeed. You will need to ensure some nodes have different names, and some others' names at the contrary should match. This StackOverflow question may be a useful read: Merging graphs in Graphviz
It seems that panda requires the .vert and .frag file extensions, and won't accept .glsl extensions.
It didn't seem to give me any output to warn me of this or ask me to do this, but when I duplicated my files and changed the extensions, it no longer spits out this version error, and I can now apply the shader to the scene and camera!
