Former | fully customizable Swift library | REST library

My question is very similar to this question that has already been asked and answered but is not 100% up-to-date.

We used the solution from Chris Gaskill for quite some time and it suited us perfectly because we wanted to redirect requests that contain more than one path segment (i.e. /foo/bar)

From Spring Boot 2.4 on, Boot uses the PathPatternParser instead of the AntPathMatcher, wherein the former does not support ** at the start of a pattern anymore (see docs).

Is there some other solution to get the same behavior? What do you use to redirect all requests, that did not match anything else, to the index.html of the Angular app?

This is the code of the controller that forwards the requests.

I have a data frame containing a shape-file that I want to merge with another data-set that contains years. I'm interested in adding a variable with years to the former while all other variables remain the same for each year. I'm not sure how to do this.

As an example, say I have the following data-set:

There are a number of different Q/A's regarding this topic on SO, but none that I have been able to find that fit my use-case. I am also very surprised that RStudio / the Shiny developers themselves have not come out with some documentation on how to do this. Regardless, take this example application:

Consider the following code, running on an ARM Cortex-A72 processor (optimization guide here). I have included what I expect are resource pressures for each execution port:

Although uzp2 could run on either the F0 or F1 ports, I chose to attribute it entirely to F1 due to high pressure on F0 and zero pressure on F1 other than this instruction.

There are no dependencies between loop iterations, other than the loop counter and array pointers; and these should be resolved very quickly, compared to the time taken for the rest of the loop body.

Thus, my intuition is that this code should be throughput limited, and considering the worst pressure is on F0, run in 8 cycles per iteration (unless it hits a decoding bottleneck or cache misses). The latter is unlikely given the streaming access pattern, and the fact that arrays comfortably fit in L1 cache. As for the former, considering the constraints listed on section 4.1 of the optimization manual, I project that the loop body is decodable in only 8 cycles.

Yet microbenchmarking indicates that each iteration of the loop body takes 12.5 cycles on average. If no other plausible explanation exists, I may edit the question including further details about how I benchmarked this code, but I'm fairly certain the difference can't be attributed to benchmarking artifacts alone. Also, I have tried to increase the number of iterations to see if performance improved towards an asymptotic limit due to startup/cool-down effects, but it appears to have done so already for the selected value of 128 iterations displayed above.

Manually unrolling the loop to include two calculations per iteration decreased performance to 13 cycles; however, note that this would also duplicate the number of load and store instructions. Interestingly, if the doubled loads and stores are instead replaced by single LD1/ST1 instructions (two-register format) (e.g. ld1 { v3.4s, v4.4s }, [x1], #32) then performance improves to 11.75 cycles per iteration. Further unrolling the loop to four calculations per iteration, while using the four-register format of LD1/ST1, improves performance to 11.25 cycles per iteration.

In spite of the improvements, the performance is still far away from the 8 cycles per iteration that I expected from looking at resource pressures alone. Even if the CPU made a bad scheduling call and issued uzp2 to F0, revising the resource pressure table would indicate 9 cycles per iteration, still far from actual measurements. So, what's causing this code to run so much slower than expected? What kind of effects am I missing in my analysis?

EDIT: As promised, some more benchmarking details. I run the loop 3 times for warmup, 10 times for say n = 512, and then 10 times for n = 256. I take the minimum cycle count for the n = 512 runs and subtract from the minimum for n = 256. The difference should give me how many cycles it takes to run for n = 256, while canceling out the fixed setup cost (code not shown). In addition, this should ensure all data is in the L1 I and D cache. Measurements are taken by reading the cycle counter (pmccntr_el0) directly. Any overhead should be canceled out by the measurement strategy above.

I'm wondering if it is legal/safe to modify intent(in) dummy arguments through pointers associated with corresponding actual arguments. Specifically when the pointer association was established before the subroutine call.

For example, is the following OK (it seems to work fine with gfortran)?

I'm attempting to solve a very simple problem - find strings in an array which only contain certain letters. However, I've run up against something in the behavior of regular expressions and/or grep that I don't get.

I have been testing my iOS app on my physical iPhone XS for quite some time now and all of a sudden it threw an error when I tried to run it saying:

"APP NAME" is no longer available.

So I went to re-build via Xcode into my device again to reinstall it. But I was getting an error message saying "Untrusted Developer".

So I tried to go to the Settings > General > VPN and Device Management screen to see if I need to 'Trust' my own developer profile again. But it's not appearing on that screen or any other new developer profile for that matter.

Did a bit of research and tried to troubleshoot. One of the solutions I've found and worked was to actually create a new developer account and use that to build and install the app on my device. I does work but I don't want to use another developer account to to build this app as the bundle identifier that I wanted to use was already tied up to that old account. And I have a bunch of other apps tied up to that old account as well.

The other reason why the above solution also wouldn't be ideal as I already have months worth of data in my device that was saved in that app. If created a new build using a new account it wouldn't be able to access the data of the old app built using the other account.

Just to note, that old account was formerly a paid account but now it expired. But I didn't have that problem until recently. So I am guessing it has something to do with the recent iOS 15 beta builds.

I will be filing a bug report for this. But does anyone have any idea or workaround to get rid of this issue? Any information that could potentially point me towards the right direction would be appreciated. Thanks!

I've filed this as a bug on AppleFeedback Assitant. They've reached out to me to ask for logs and the certificate I used to re-build the project. I'm guessing this this a bug and they are trying to fix it.

I have a list of strings that I want to sort by two custom key functions in Python 3.6. Comparing the multi-sort approach (sorting by the lesser key then by the major key) to the multi-key approach (taking the key as the tuple (major_key, lesser_key)), I could see the latter being more than 2x slower than the former, which was a surprise as I thought they are equivalent. I would like to understand why this is so.

In C++20, some ranges have both const and non-const begin()/end(), while others only have non-const begin()/end().

In order to enable the range adapters that wraps the former to be able to use begin()/end() when it is const qualified, some range adapters such as elements_view, reverse_view and common_view all provide constrained const-qualified begin()/end() functions, for example: