A13cui

Used to be @alecui

21 / RO

Public Key

npub1gykyxa0f5e96pelee3spdvxhet9r0aeky05dfl9apjal4wym346q9rd8w8

NIP-05 Address

a13cui_at_emacs.ch@mostr.pub

Profile Code

nprofile1qqsyztzrwh56vjaquluuccqkkrtu4j3h7umz86x5lj7sewl6hzdc6aqpz4mhxue69uhhyetvv9ujumt0wd68ytnsw43qd7y0lf

Publishing to

relay.mostr.pub

Author Public Key

npub1gykyxa0f5e96pelee3spdvxhet9r0aeky05dfl9apjal4wym346q9rd8w8

Show more details

Last Notes

2023-09-19 23:25:28

This is definitely what we needed, Chess.com...
https://www.chess.com/news/view/opera-launches-new-custom-chess-browser
Guess raising the price of premium wasn't enough, our money is getting spent properly.
We’re excited to bring the world of chess into this new version of #Opera’s #AI-powered browser to help make the beautiful game that we all love even more accessible, convenient, and fun.
— Kuhnert, VP of Business Development at Chess.com
Steer away from them and just use #Lichess. It's free and libre, ad-free and you actually get unlimited puzzles and board analyses. Well, stay away from Opera too, but that's for another time.
#Chess
https://media.emacs.ch/media_attachments/files/111/094/384/589/707/372/original/aa17bf61610a06ea.png

2023-09-18 15:19:29

- reply

@npub12jq…hq6m wait, people take JRuby seriously? That's a first

2023-09-15 22:33:15

- reply

@npub1fzx…t6h3 right now no, it's for fun, I would've phrased it differently if it was for a job
But at least I want to see what I'd be getting into if I accepted an ABAP offer

2023-09-15 22:06:12

A devil on my shoulder keeps nagging me to go full corpo mode and use ABAP, should I listen to it, kill that stupid imp or double down?

2023-09-13 22:01:09

- reply

@npub1fzx…t6h3 no, I live at the extremes
Either only parens or no parens
And no corrosion

2023-09-07 23:38:00

https://media.emacs.ch/media_attachments/files/111/026/489/799/363/686/original/bd61210d3af68120.png

2023-09-04 22:22:50

- reply

@npub18yv…xxyn moving CharAt$ to the end didn't work, it still says it's a syntax error, but I can't figure out why
it seems to be specifically the fault of DrawCrate since it otherwise works perfectly fine, even just a sole DrawCrate (1, 1, 1, _RGB32(1,1,1)) in the DrawStacks sub triggers it
Renaming the sub doesn't help either
I am actually baffled and nothing in the DrawCrate is syntactically incorrect as far as I can see

2023-09-03 00:17:09

- reply

@npub18yv…xxyn is QB64's parser one pass or two pass? what's the reasoning behind that?

2023-08-25 05:51:43

- reply

@npub1skl…jhe5 multiple times with a friend, I somehow had the gift of hitting in the same spot and it gave him a headache afterwards

2023-08-23 07:37:45

- reply

@npub12jq…hq6m I'll be honest, saying "Chrome is more secure than Firefox" sure is a hot take. How are you a minimalist and yet you use that memory hog?

2023-08-18 22:25:22

- reply

@npub1j8m…v5r7 can you please add this under a CW? thank you

2023-08-17 18:22:52

- reply

@npub18yv…xxyn let's not forget how RH gave up on improving Windows virtualization because they don't care about the graphics driver anymore which is the sole reason why Windows works at like 10-20 FPS in QEMU, otherwise the other drivers are really capable

2023-08-16 02:09:41

- reply

@npub17es…clcm
"Wanna know the ancient secret of not having concurrency issues? It's called not having shared mutable data in the first place. And for a library that is supposed to provide pure functions, none seems required whatsoever."
you've skipped about a couple of lemmas in your argument. who said they're providing pure functions? I again did something that you can't do (along with engaging with what I said): looking at the source code and how the library is used. while looking through it, I noticed that it looks like only C++ needed to implement sync since Java and JS seem to handle it just fine (not like you'd care about actually looking at the code since you're set on disagreeing with me). it seems perfectly content with either sync or async contexts (although async would definitely be preferable for, you know, the case when you're in a fucking Android dialer and need to format phone numbers in real time). also that data is there just for convenience, you can... just... regenerate it if you really need it. I would argue that they should've used protobufs by now especially since they were around before Google bought Android (this library was made specifically with it in mind), it would've been more efficient, but I'm not the one making the rules.
https://github.com/google/libphonenumber/blob/master/FALSEHOODS.md btw, yet another piece on why programmers like you oversimplify phone numbers. I bet you check emails using regex considering you oversimplify and ridicule this problem, huh? fuck you, I'm not engaging with you since you're clearly bad faith and spent essentially no thought on looking for a couple of minutes on what this library actually offers. don't @ me ever again, you're as bad faith as I perceived you to be in your wayland discussions, go sugma.
EOD.

2023-08-15 12:56:25

- reply

@npub1463…00qn you're putting out wallpapers like it's nothing, I can't get enough of them

2023-08-14 17:01:53

- reply

@npub17es…clcm because... Um... People want to also use it in async contexts without having concurrency issues? The reason why everyone wants to use async libraries in the first place? Also I love how you didn't engage with the rest of the arguments I laid out because it obviously makes you look dumb

2023-08-13 23:13:06

- reply

@npub17es…clcm Linux 0.0.1 being barely functional, what can I say. You couldn't even get yourself to bring actual OSs like Research Unix VN (where N is {1..7}). V1 was at 4k lines of asm and that was a fully functional OS, unlike Linux 0.0.1. that's a really nice false equivalence you're making right there. I went ahead and cloned the source code to see what it's doing in those "50k" lines and what are the biggest files. these are the top 20:
225 ./geocoding/phonenumber_offline_geocoder.cc
239 ./regexp_adapter_icu.cc
240 ./utf/utf.h
242 ./asyoutypeformatter.h
263 ./base/basictypes.h
296 ./stringutil.cc
358 ./utf/rune.c
406 ./shortnumberinfo.cc
450 ./base/memory/scoped_ptr.h
466 ./utf/unicodetext.h
524 ./utf/unicodetext.cc
766 ./asyoutypeformatter.cc
909 ./phonenumbermatcher.cc
979 ./phonenumberutil.h
1056 ./test_metadata.cc
1608 ./alternate_format.cc
3273 ./phonenumberutil.cc
4030 ./short_metadata.cc
14119 ./lite_metadata.cc
15052 ./metadata.cc
./metadata.cc literally has only a static const unsigned char data[] that's for some reason made out of hex, same with the other metadata files. if we add all of those up, we get 34257 which leads us to 16318 lines. The metadata is generated by a Java file iirc from a XML file and packagers usually turn on the option, but can be dynamically regenerated if needed, so the metadata files are only for convenience. Here's that file: https://github.com/google/libphonenumber/blob/master/resources/PhoneNumberMetadata.xml (the biggest one, you can see the sheer complexity of phone numbers)
next utf/ is relatively heavy, but:
This project uses some third-party code:
src/phonenumbers/utf/ sources come from lib9 which is also used in Go.
So we end up at 14512 lines. Huh, we're getting close to Linux 0.0.1, aren't we?
You also need synchronization because of course you can also use this in an async context, c'est la vie. Nice making it sound worse than it actually is, eh? :blobcat3c:

2023-08-13 22:23:51

- reply

@npub17es…clcm but parsing the other dozens of phone number formats described in a wiki article that you conveniently skipped over requires those number of lines
not sure if 50K+ lines, but several thousands for sure

2023-08-13 22:21:03

- reply

@npub17es…clcm but it is a valid way to write a phone number nonetheless

2023-08-13 22:06:41

- reply

@npub17es…clcm 00 is a way to write + and I see it quite often actually. In my case, 0040 (or +40) is the country code for Romania, so I didn't pull that example out of my ass.
Also, what do you think those pieces of software use to validate phone numbers, huh? If you want a nice rabbit hole: https://en.wikipedia.org/wiki/National_conventions_for_writing_telephone_numbers
It ain't that simple, huh?

2023-08-13 21:57:59

- reply

@npub17es…clcm anything that's supposed to be international is going to be really damn complicated. ask people who have to deal with UTF-8 or time zones or localization or general i18n stuff. not everyone does your format: I write it all together, as on 0764123456... except that can be 0764 123 456... or +40764123456... or 0040764123456 or a myriad of other ways people format their phone numbers.
I have no clue what they've specifically done in those 50K lines, but it sure as shit isn't as easy as a regex, it has to be reliable 100% of the time

2023-08-12 16:53:39

- reply

@npub1skl…jhe5 @npub1c7d…s5u0
https://media.emacs.ch/media_attachments/files/110/877/685/532/447/690/original/1c93ec88888c4d7f.png

2023-08-11 11:33:14

- reply

@npub1463…00qn I definitely need to ramp up since I've been doing this for approximately the same amount of time in total, but I specialized in pixel art cars and we have a whole community out there dedicated to this + all of the momentum has eroded away since 8th grade. Who knows, maybe now is the time :p

2023-08-11 10:53:29

- reply

@npub1463…00qn P.S.: I love your pixel art style
How long have you been doing this for?

2023-08-11 10:34:55

- reply

@npub1463…00qn thank you, I now have a cool wallpaper:)

2023-08-10 17:09:33

- reply

@npub1c4s…aqv4 34 *MILLION*?

2023-08-10 12:55:46

- reply

@npub1fzx…t6h3 ironically, I am using Common Lisp at this very moment for my Tcl LSP :blobcat3c:

2023-08-10 12:52:02

- reply

@npub1fzx…t6h3 if you're willing to go into superior better syntax territory, there's also Nim