Thread 107003819

What are you maids working on?

Last one: >>106982750

>maidnigger just waits for the previous thread to die to immediately post his maidnigger shit as the OP
and you cultists wonder why no one likes you

>>107003906
The non-maid /dpt/ fell off the catalog hours ago due to lack of interest. Non-maid /dpt/ is only sustainable when glowies are paid to post in it and currently the US government is still shut down. This is why the thread has been 90% more civil and on topic.

Maid /dpt/ always hits the bump limit because maids see the OP and come to participate. Given that the anti-maid glowies are currently living the NEET life, this is a good time for maids to enjoy themselves by discussing advanced Mathematics and Computer Science research with less disruptions.

The subject of this maid /dpt/ is the book "A Study in String Processing Languages". Please find a copy from your favorite archive site and enjoy reading about String Processing.

>>107003819 (OP)
>>107004006
>The non-maid /dpt/ fell off the catalog hours ago due to lack of interest. Non-maid /dpt/ is only sustainable when glowies are paAAAAAAAAAAAAAAAAAAAAAAAAAAAAACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACK

>>107004085
>>The non-maid /dpt/ fell off the catalog hours ago due to lack of interest. Non-maid /dpt/ is only sustainable when glowies are paAAAAAAAAAAAAAAAAAAAAAAAAAAAAACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACK

>>107004006
>The non-maid /dpt/ fell off the catalog hours ago due to lack of interest. Non-maid /dpt/ is only sustainable when glowies are paAAAAAAAAAAAAAAAAAAAAAAAAAAAAACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACK

>>107004096
>The non-maid /dpt/ fell off the catalog hours ago due to lack of interest. Non-maid /dpt/ is only sustainable when glowies are paAAAAAAAAAAAAAAAAAAAAAAAAAAAAACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACK

>>107004134
>/maidpt/
>mental retardation central

>>107004183
>important maid computer research:
>posting aislop
i think ive got my fill of mental retardation for the moment

>>107004200
If maids were posting AI slop it would be cuter and more moe and maid themed. Not some guy being mean to robots. Androids should not look like anthropomorphic iPhones. They should look like cute maids with huge boobs.

>>107003906
>>107004085
>>107004091
>>107004096
>>107004105
>>>/b/
Go be a dumb swarthoid elsewhere.

>>107004260

>Go be a dumb swarthoid elsewhere

>>Go be a dumb swarthoid elsewhere
>>107004260

>>107004260
>Go be a dumb swarthoid elsewhere.

>>107004466
>>The non-maid /dpt/ fell off the catalog hours ago due to lack of interest. Non-maid /dpt/ is only sustainable when glowies are paAAAAAAAAAAAAAAAAAAAAAAAAAAAAACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACK

my downloader program is coming along nicely

>>107004466
>>The non-maid /dpt/ fell off the catalog hours ago due to lack of interest. Non-maid /dpt/ is only sustainable when glowies are paAAAAAAAAAAAAAAAAAAAAAAAAAAAAACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKcACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACKACK

Thank you janny for cleaning the thread. Here are more math and CS papers for maids. Please browse them at your leisure:

>https://webspace.ship.edu/msrenault/divisibility/StupidDivisibilityTricks.pdf
>https://www.robertgaskins.com/files/gaskins-gould-cal-snobol4-1972.pdf
>https://ftp.regressive.org/snobol/misc/ftp.cs.arizona.edu/gb.pdf

>>107004525
>janny sucking himself off

>>107004525
>>107004488
>>107004510
>>107004223
>>107004006
botted the thread again award

>>107003819 (OP)
How to get an Asuna wife?

>>107004532
bro would probably congratulate himself for deleting his own balls lol

I've decided to do some side project because I was bored and now I get why people do hobby projects. It is fucking blissful to be able to just code and learn without having to worry about delivering on time.

>>107004541

maybe her papa or uncle passed away and maid threads were transferred to her

>>107004510
>>107004624
What did you make this in?

>>107004134
CLEAN IT UP JANNY

>3 1/2 hours old
>not a single post about programming
OP fucking sucks at baking. Fuck you, OP.

>>107005668
Do you think the threads would be better if the OP contained a link to some code to discuss and play with for the thread? I am open to suggestions for improvements, but I can't actually compel people to participate.

>>107004200
https://www.youtube.com/shorts/IQHjhBq7cwE

haskell

>>107005963
so true

Been working on rewriting this thing to be cross platform and not rely on external Windows only tools from the late 90's in my spare time.

>>107003906
fpbp

>>107004006
>A Study in String Processing Languages
I am implementing some of the matchers listed in the second chapter of this book in Java.

I'm better than you. You all suck.

>>107006896
I now have a very simple handcoded backtracking pattern recognizer. It is based on the SNOBOL4 patternd described in the first couple chapters and is not capable of side-effects. I am going to try to implement the language described in the book.

how do i get a /dpt/ bf?

>>107005793
>Do you think the threads would be better if the OP contained a link to some code to discuss and play with for the thread?
Not really. It'd just be a pile of shit usually that everyone would take a big dump on; they can make other threads if that's the sort of humiliation ritual they want.
I guess with that I ought to contribute. Writing C macros to make iterating over a hash table trivial and applying them to a large codebase. Which is easy, except I keep getting distracted with non-programming stuff.

haskell for virgin chuds
lisp for sex havers

>>107007878
this, lispfags btfo

>>107005793
It'd be better if the OP contained a screenshot of code they wrote.

>>107008381
Screenshots of code aren't useful, and aren't as nice to look at as maids.

>>107007650
>backtracking regex engines are BAD
>backtracking pattern recognizers that work in the same manner are GOOD

>>107008381
fucking this

in cmake how are you expected to add source files / headers that are located in sub folders purely for the sake of code organization and not because they are libraries or because of any sort of structural reasons with the project. I basically want the contents of these subfolders to be the equivalent as if they were in the root of the project (they can include headers located "above" them or in an adjacent sub folder for example)

>>107003819 (OP)
Thoughts on the Elixir language?

This nigger is actually trolling me.

>>107008692
this is what you deserve

>>107008423
I don't actually have a strong opinion about backtracking. I just thought it was fun to implement. The book goes on to describe a language called Summer and I might try to implement it.

>>107008832
Well if you enjoy this, I recommend you look at perl/pcre backtracking regex engines afterwards. It's hardcore.
You start with simple alternations, then simple quantifiers implemented with a stack (not using recursion) is already a bit of a mindfuck. Then it gets more complex with "procedure" patterns/recursive patterns, backtracking control (via quantifier modfiers, lookarounds, independent patterns, (explicit) backtracking control verbs: (*PRUNE) (*COMMIT), etc...)).

>>107005883
Too bad that the multiplatform IDE only can run on Windows.

how do we save OOP?

>>107009968
Right click, Save.

>>107009977
you think this is a joke? OOP is fucking dying and you're making jokes?
https://www.youtube.com/watch?v=EKjUd7DAqCo

>>107009988
kunyetsa erwy pachemu Go rast ryyact i dlyatov

>>107009997
spazzibba nigga but I don't speak r*ssian

>>107010014
Me neither that's what your video title says.

C++ worth learning in 2025?

>>107004134
>invents bipedal robot
>hasn't invented replacement for squatting toilet

>>107009988
you vill be functional

The language should have simple modules/namespaces to group your functions, so you would use classes only when you need classes. No inheritance. Have abstract classes/interfaces.

>>107010225
>>107009968

>>107010048
if you're interested in areas where C++ is used (large, complex + performance reqs like robotics, video games, browsers, car software, etc), then yes
personally I would rather save my sanity and do something else

>>107009968
Ban inheritance.

All problems today are caused by
QuantumParticle inherits from Atom inherits from Molecule inherits from Cell inherits from Organism inherits from Human inherits from Life inherits from Earth inherits from Planet inherits from Moon inherits from Tire inherits from Rubber inherits from TyreRubber inherits from TyreRim inherits from Differential inherits from Axle inherits from Car inherits from Driver inherits from Galaxy inherits from Star inherits from Universe inherits from Multiverse inherits from SpaceTime

Maybe this belongs in /stupid questions/ but I decided to sign up at my local CC for programming classes for Spring, just to do something during the nighttime. The last time I programmed was with C++, in high school, in 2006 but I liked it (yes I am old). The three courses recommended these books but don't require them, are any of them worth keeping or not? Ignore the Linux one, the school store sent it by mistake.

>>107010048
C++ is a language where the footguns have footguns. Seriously many ways to fuck up when writing it, and not all of them involve getting crashes or the wrong results; for example, it's tremendously easy to write code that makes far too many copies of things, and not know that you've done it other than wondering "why is my code so slow?"
Learn it if you need to, but be aware that it's slow to build and you'll need to debug it a lot.

>>107010225
yay haskell

>>107010407
>The three courses recommended these books but don't require them, are any of them worth keeping or not?
Read them (at least partially) and form your own opinion.
The C++ one is worth keeping if you're going to keep on using that language on a regular basis, as that's written by the language creator.
You probably don't need two Python books.

>>107010407
Programming language books are useless, you can learn a new PL in 2h if you already know one. It's knowing how to program that's complicated. Start to programming something, anything, and look at the language docs/reference anytime you need it.

>>107010493
>you can learn a new PL in 2h if you already know one
you can learn a new PL very badly in 2h
knowing how to use a PL well takes more effort, as they all have quirks and non-standard features; those are often the bits where the special sauce is, why you're learning that language in the first place

>>107010515
sure, but I mean in 2h you can learn every primitive you need to implement any program: the operators, the basic types, control flow, function, the primitive data structures, etc..

>>107005793
"what are working on" OP is like 15 years old, there's no changing it
this is the programming blog thread, to discuss what you're working on
i am not currently working on anything

>>107009968
OOP is doing well and doesn't need saving

inheritance by itself is fine
just ban dumbasses who use inheritance for classification, like >>107010403

>>107005793
>Do you think the threads would be better if the OP contained a link
kys

Maybe I'm just too old for programming.

Woah maid posting is still alive?

>>107004510
You might like this. A high entropy ID generator that is pronounceable.

Proquint: https://arxiv.org/abs/0901.4016

Erlang is king.

I feel like I've been stretching myself too thin. Additionally to my day job as a programmer I am also finally learning proper touch typing, got a new keyboard and am learning a ton of shortcuts.
I'm also reading books about learning strategies, revisiting some math stuff and over the weekend I tried a tiling window manager and a bit of vim. I think I'll leave vim for next year, not sure about the tiling wm yet.

>>107012660
I'm working for 4 hours a day on programming in addition to working nights. You're working a job many people would kill for. Dude's earning like six figures and going "ack this... typing stuff... and 30 minutes of math a day... I'm stretched so thin!!!" Sack up, my dude.

Baremetal servers vs Linux servers

>>107012927
wtf they had built in shades

>>107010260
>>107010446
I have read about the shooting yourself in the foot problem. I'll at least learn the basics so I can feel better about saying I "know" it when looking at job options in the future. Seems there's always a handful asking for C/C++ to maintain some old system.

No one knows C++, they just carve out some usable subset that's like 10% of the actual language, write their own version of the standard library and call it a day.
Oh and every team has a different permutation of features they use so switching between projects is like learning a new language.

>>107013588
I'm just kicking around what else I should learn considering job prospects in my area...I see a lot of c/c++ or .net. boomer enterprise companies I guess.
Also, papa john's needs to fix their shit

>>107013588
>they just carve out some usable subset that's like 10% of the actual language
yes because you need 10% of the features of modern programming languages to achieve literally anything you want
the root of all evil comes from the turbo autismos who want to achieve "beautiful" solutions to every problem
you end up with shitfests like Typescript and it's turing complete type system
that's why I respect the Gofags, they stood their ground against the autismos but perhaps a bit too much

>>107014293
>you end up with shitfests like Typescript and it's turing complete type system
Saying this when C++ templates are Turing complete is wild.

[vk::push_constant] uint32_t image_handle;

T resource_from_handle<T>(uint32_t handle) {
return spirv_asm {
OpCapability BindlessTextureNV;
OpExtension "SPV_NV_bindless_texture";
OpSamplerImageAddressingModeNV 32;
result: $$T = OpConvertUToSampledImageNV $handle
};
}

[shader("compute")]
void main(uint2 id: SV_DispatchThreadID) {
let image = resource_from_handle<RWTexture2D<float4>>(image_handle);
let color = image.Load(id);
let grey = dot(float3(0.3, 0.6, 0.1), color.xyz);
image.Store(id, float4(grey, grey, grey, 1));
}


Damn Slang and Nvidia are based. Now I don't have to fuck around with descriptors at all.

whats the point of the rust if you end up having to clone everything once you get to large projects anyways

>>107015201
validating xour identity

>>107003819 (OP)
the C++ STL is shit, right? i should just implement my own data structures, you only have to do it once and can use it for any program, do you disagree with this?

>>107015201
Are you retarded?

>>107014293
>that's why I respect the Gofags, they stood their ground against the autismos but perhaps a bit too much
minimalist fags that insist on making the programmers suffer with a language too bare bones are idiots in their own right

>>107003819 (OP)
Anon, How to implement scrollbar from scratch for my file manager in TempleOS?

>>107016857
You can store a scroll offset representing how far down you're currently scrolled. Then when you calculate which lines to render, sum up the virtual height as you go. When you reach the scroll offset, you know that lines are now visible. When you reach past the scroll offset + height of one view, you can stop drawing. Scrolling just changes the scroll offset. That's one way to do it.

I am reading about the implementation of futures and asynchronous execution, and working on an async library for Teal (typed Lua) as I go.

https://rust-lang.github.io/async-book/02_execution/01_chapter.html

>>107017961
step 1: tokio

Back to work.

>>107016305
Stl isn't that bad and you can also get 3rd party libs. Only implement them if it's something you're interested in doing for its own sake.

This is a stupid question.
Let's say I am coding a c math library.
What is the preferred method?
static inline u8 IsOdd_u8(u8 n) { return n & 1; }
static inline u8 IsOdd_u16(u16 n) { return n & 1; }
static inline u8 IsOdd_u32(u32 n) { return n & 1; }
static inline u8 IsOdd_u64(u64 n) { return n & 1; }
static inline u8 IsEven_u8(u8 n) { return !(n & 1); }
static inline u8 IsEven_u16(u16 n) { return !(n & 1); }
static inline u8 IsEven_u32(u32 n) { return !(n & 1); }
static inline u8 IsEven_u64(u64 n) { return !(n & 1); }
static inline u8 IsOdd_i8(i8 n) { return n & 1; }
static inline u8 IsOdd_i16(i16 n) { return n & 1; }
static inline u8 IsOdd_i32(i32 n) { return n & 1; }
static inline u8 IsOdd_i64(i64 n) { return n & 1; }
static inline u8 IsEven_i8(i8 n) { return !(n & 1); }
static inline u8 IsEven_i16(i16 n) { return !(n & 1); }
static inline u8 IsEven_i32(i32 n) { return !(n & 1); }
static inline u8 IsEven_i64(i64 n) { return !(n & 1); }

or
#define IsOdd(n) n & 1
#define IsEven(n) !(n & 1)

>>107019842
#define IsOdd(n) ((n) & 1)
#define IsEven(n) (!((n) & 1)))

it is never possible to have too many parentheses in a macro

>>107019861
Oh, goodness, you're right. That was almost a close call.

>>107014355
Yo dawg, we heard you like languages so we put a language inside your language so you can...

Write unmaintainable code that requires a hermeneutic process for humans to parse your illegible scrawls to the machine gods?

>>107003819 (OP)
I discovered a new programming paradigm.

I don't know what to call it or whether to tell people about it or not, I never thought something like this was possible.

>>107019861
>not parenthesising 1
ngmi

>>107020664
>>107020666
Tell us more, Satan.

>>107020664
>>107020666
Don't you think we have enough of those?

>>107020666
Of course it exists, it's just that you don't know how it's called.

>>107019842
What's the point of these functions?

>>107003819 (OP)
Just finished setting up full encryption flow for my site and prepared nginx for domain expansion. Learning wireguard so I can use the a new domain (in honor of my dog) to reach self-host services at home. Maybe hook up a few friends and coworkers with some non-google tools.

I was working on a miniOS for my pico, but those Microcontrollers only connect to 2.4ghz and my phone's hotspot won't show, even with compatibility mode. What's the fucking point of a mini tool faster than android apps or websites to track workouts if I cannot connect to wifi? I started looking into making an app with Kotlin to middleman API calls for the pico through Bluetooth (because why would the pico be capable of bt wifi) but then what's the point of using a pico in the first place? Then the power button on the li-po shim I bought went out after 200 presses, and pimoroni changed their mind about shipping a new one last second because "muh tariffs" on a fucking $10 item. To be fair, I didn't purchase directly from them, so the replacement was in good gesture, but the reason for retraction felt fake as fuck.

I wrote menu classes, no-block popups, warnings, and saved WiFi connection func, which I'll save for later use. Maybe I'll make my own IoT devices and use a pico to quickly controll them all.

Bah

Beautiful and clean C++ code

>janny deleted most of the maidposts
>mfw

>>107022166
I want to fuck her so bad...

>>107003819 (OP)
>be me
>fire up .NET 10 preview
>compile my app AOT
>outputs a native binary faster than your Rust build finishes cargo-check
>zero reflection, zero DI container, zero nonsense
>just global using static and a dream

meanwhile DI "architects" be like:
"uhhh how do I register my factory interface in the IServiceCollection scoped to an IHttpContextAccessor??"
bro you’re writing XML with extra steps

C# today: modern syntax, structs, spans, deterministic memory, native AOT
C# yesterday: “managed meme language”

Low-level bros shaking: “b-b-but muh manual memory management!”
Sorry, champ, my trimmed self-contained single-file ELF starts faster than your handcrafted C++.

DI nerds coping: “but how will we mock the service locator??”
bro just pass the function

old .NET devs: worshiping transient scoped singleton lifetimes
new .NET devs: delete the container, embrace static purity, ship native

C# isn’t “managed” anymore, it’s just civilized C++ with a moral compass
stay mad

>>107022210
C# GODS *pauses for garbage collection* WON!

>>107021226
>Maybe I'll make my own IoT devices and use a pico to quickly controll them all.
If you're interested something like 9p makes networking with iot and i2c devices more straightforward (down to 300 lines of C), they just appear as regular files
https://youtu.be/DbVNGA4FA3Y?t=180

>>107019842
This is what _Generic is for

what is the proper way of handling web sockets in python?
I've a connection where I can receive 3 types of messages where each goes through their own own pathway of processing
what I thought of is; in my receive loop, just put all the raw messages into a async queue. have another task categorize them to 3 different async queues and finally have 3 more tasks that consume the categorized queues and do the necessary processing
it sounds retarded but I cant think of a more efficient way
(each of these message types end in database ops or api calls using the data etc)

>>107003819 (OP)
TIL programming with time is another way to summon cthulhu.

>>107023069
In sepples this is just

template<typename T>
bool isEven(T t) {
static_assert(std::is_integral_v<T>, "T must be integral type");
return (t & 1) == 0;
}

>>107023358
This is pretty standard architecture for a networked application

>>107021224
Checking parity of integers, which is pretty common.

Is this enough boilerplate?
static json_ptr *alloc_stack(size_t *);
static int alloc_string(struct json_string *, size_t *);
static int alloc_array(struct json_array *);
static int alloc_object(struct json_object *);
static int maybe_grow_stack(json_ptr **, size_t, size_t *);
static int maybe_grow_string(struct json_string *, size_t *, size_t);
static int maybe_grow_array(struct json_array *);
static int maybe_grow_object(struct json_object *);
static void trim_string(struct json_string *);
static void trim_array(struct json_array *);
static void trim_object(struct json_object *);
static int make_string(struct json_string *, char *, size_t);
static void free_string(struct json_string *);
static int free_value(json_ptr *, size_t, size_t);

>>107024581
>static
I was assuming these were internal functions, then I realized you don't have a state or context but just a bunch of functions. :(

Implementing type inference for a lambda calculus but suddenly started thinking I would like to have arbitrary rank types so now I have to spend a long time thinking about how that implementation will be different from my rank one type implementation.

>>107019842
It's irrelevant, any non int will be promoted to int by the compiler.

>>107024708
bidirectional type inference

>>107024811
I'll be honest, I don't get how unification and bidirectionality are any different

>>107024817
with a type scheme you have a bunch of (forall a ...) what you want is if you have a (forall a b c ...) and need a T that you can substitute a,b,c etc for it
vs the opposite if you need a scheme but have a T
i cant remember the version i implemented based off of but it used 3 types of type, tau types, rho types and sigma types
sigma := foralled (maybe with constraints) rho types
rho ::= sigma -> rho | tau
in other words (sigma ->) ... tau
tau = plain monomorphic types

>>107024899
This seems like a paper I just read by spj this morning about rank n types. It mostly confirmed the approach o was gonna take but he made a big deal about subsumption which I did not make a delineate in my implementation so I have to figure out how subsumption fits into my unification and when I should instantiate quantified types.

>>107024817
>>107024899
also when you implement it you have a type check and a type infer step, and can shift between the two when needed to gain more information
e.g. inferring (a :: S) -> checking a against S
checking x against S (x is some variable) -> infer the scheme of x to be S' (infer and generalise), check that it subsumes S [you can probably merge these steps]
inferring or checking (a b) -> infer a, it will have e.g. (S -> R), then check b against S
so you can see how all of that type scheme information is available

>>107024932
yes, it's very likely that exact paper, i thought it involved spj but I couldn't find it with a search

basically unification is symmetric but subsumption is not, its like subtyping (except that A subsumes B means A is a supertype of B)

>>107024950
"Practical type inference for arbitrary-rank types" from July 2007. Would post a link but I am at work. Thanks anon, I will continue to study this. So far I haven't implemented your inference yet, but I did do rank 1 kind inference for poly morphic minded types and I figured the general shape of type inference would be pretty similar. Though it's more clear to me now I am only doing unidirectional type inference so I'll have to see how to adapt mine for bidirectionality.

>>107024988
Also sorry for the typos, I'm phone posting at work.

>>107024988
if you think about it its only similar if you want the same thing at the kind level, i.e. polykinds and higher rank kinds
otherwise you might not need it
one thing you shouldnt forget is that if youve got (forall ...) then it should have kind *, i.e. you cant have (forall ... T) just be whatever kind T is

>>107025018
I mean it might be simpler to have the same structure at least, i.e. check vs infer kind. But to start you should ignore polykinds because the more stuff you get in the more book keeping there is (e.g. look at how many bugs GHC has)

>>107025018
>if you think about it its only similar if you want the same thing at the kind level
Yeah originally I was just gonna do rank 1 types and therefore type inference would be analogous to kind inference with the addition of patterns, but if I do rank-n types I will probably try to keep it similarly minded and try out rank-n kinds too.

>>107025033
I see, I will probably stick with at rank 1 kinds because I like the idea of having Proxy but really it's not something I have a great need for since I don't really use complex kinds in Haskell much either.

>>107024581
Now make them all AVX aligned and reject any other length.

>>107025033
>look at how many bugs GHC has
I thought Haskell was LE well-made and reliable language? Are functards bad programmers despite their arrogance?

>>107025776
GHC is a huge project started decades ago with a limited amount of funding and a ton of features

>>107025776
Imagine thinking that a compiler having open bugs is some sort of own. GCC has at least 10000 open bugs on their tracker right now.

>>107025815
>with a limited amount of funding
Enough funding to produce hundreds of theoretical papers about it though. It's not like they had to rush it like you imply.

>>107025855
>Imagine thinking that a compiler having open bugs is some sort of own.
But it is.
>GCC has at least 10000 open bugs on their tracker right now.
I know and I mocked them the way they deserve in the past.
Also I reject this false dichotomy.

>>107025855
papers being funded != GHC being funded

>>107025976
coalesced brain time

>SAR
x86 ciscers...

>>107016401
A bare bones language can always grow features, slowly and carefully, as it becomes clear that they're important and not just syntactic sugar that makes programs look different for no reason or provide 'power' with no real vlaue.

It's much harder to remove things once you've added them. Plus, having features makes people think you have to use them, even if that's not true. Every weird feature you add is a trap that makes bad programmers think they should use to look like good programmers, making worse programs, while the bad programmers stay bad. And clueless.

I'd rather a minimalist language grow slowly. It did take an annoyingly long time for Go to get generics though, I'll give you that.

>>107026209
indeed, white men call it SRA

>>107003819 (OP)
>what are you working on

A C remake of a python script that generates usable usernames based on a list of full names.

I'm relatively new to programming. I've only really spent the last 2 months learning. I can do moderately complex C and python. Basic terminal applications and such.

>>107026351
scheme is dead thoughbeit

>>107026351
Feature can mean anything, including something very well integrated to the language, highly composable etc.. Including an FP type thing, like type classes. It's does not necessarily mean an awkard non composable thing like you'd find in PHP or Perl.

>It's much harder to remove things once you've added them.
Yes but simplicity is hard to achieve. It's not because you make a simple language that it will automatically be extendable cleanly. IMO you have to think about advanced features right from the start. There are a lot of features that can conflict with each other in more or less reconciable ways. although I think you can have it almost all with low level primites and a complex type system and semantic checking.

GC is a constraint/weight on almost everyting: code generation and JIT (gc roots in registers, safe points), types (either the heap objects must have a type field or there must be code generated to walk over the data structures during collection), exceptions, concurrency. Probably other things.

Fancy FP types require that the original type system must be a clean Hindley-Milner type system, not be retarded.

Macros, compile time execution and modules must be designed at the same time, and with the structure and phases of the compiler. Macros also require a clean AST and lexical scoping.

Exceptions and linear types conflict with each other if they are independent but with some semantic checking and a some constraints it should be possible to have both. Exceptions can also conflict with defer. Gotos can also softly conflict with linear types, indirect gotos definitely need to have constraints on where the can go to.

Pointers can conflict with a lot of things, must have constraints.

>>107027265
>low level primites
*low level primitives

>>107023628
Too verbose.
#include <concepts>
bool isEven(std::integral auto t) {
return (t & 1) == 0;
}

>>107027810
>cmov
too new, potentially slow on slightly dated architectures
[cpde]
~t & 1

I wish I could write code in multiple columns like in the bible.

Linux is bound to fail because it's trying to be a Tower of Babel.

>>107025855
Probably easier to do a paper over trying to understand the 100k (out of 200k) lines of Haskell code that make the typechecker.

After all of this time, I just realised that "firm"ware is called that, because firm is between hard and soft.

() parenthesis
[] brackets (<> angle brackets)
{} braces (or curly braces)

stop calling them by the wrong names

>>107029811
() brackets
[] square brackets
<> angle brackets
{} curly brackets

Stupid burgers.

>>107027265
>Exceptions and linear types conflict
just use par

>>107030799
par?

>>107030845
imagine a thing you can "call" that takes 2 exit continuations

>>107030990
>2 exit continuations
Yes, the FP equivalent of a procedure that can both return and throw an exception, but I don't see how it's relevant.

IDK if this is the right thread, but what could be going wrong?
> vs code
> remote connection to linux box via ssh
> python project with uv and ruff
> able to edit and run code
> intellisense works just fine
> but the type checker and linter/autoformatter do not work
I've seemingly turned on everything in the settings and all of the necessary extensions are installed

> check out rust gui that doesn't seem barebones and has charting widgets https://github.com/longbridge/gpui-component
> its 700 dependencies
> 6gb
> rust-analyzer needs a solid minute+5gb ram to look up definitions
> +2gb when trying to run the application in debug mode
what the fuck is wrong with them. At least it compiled in 5mins but shit man its so fucking memory hungry it lags.

>>107031212
If anybody ever has this issue, this was the fix:
> specify ruff parameters in pyproject.toml
> make sure pylance is pointing at the right nodejs executable (mine wasn't for some reason)

>>107029827
This

>>107031029
>I don't see how it's relevant.
that's literally how exceptions work

>>107032360
dude, nothing you've said make sense

C with references

>>107032911
C needs to be reforged entirely

i asked this in /twg/ but none of them actually write code in their jobs (my mistake), so i'll ask here too

Which book would (You) recommend for learning how to make an intepreter

https://craftinginterpreters.com
or
https://interpreterbook.com

leaning towards the latter because its short, its in Go and it has a sequel where you actually make a compiler for the Monkey language as well

>>107033553
Don't know, never read any. But if you want to just start, learn recursive descent and how to parse expressions with it, optionally the Shunting-Yard algorithm. Once you have simple arithmetic expression containing only integers (no variables), write a recursive function to evaluate it, then evaluate it with an explicit stack. At that point you know how to compile expressions into bytecode/IR for a stack machine. Learn how to compile conditionals and loops (need to emit jump and conditional jumps), then finally local variables, functions, function calls. With a stack VM, you need 2 stacks for interpreting functions, the expression stack that you were using until now, and the function call stack. For compiling code with local variables, you'll need to do a pass on the AST and do symbol resolution (ie associate each identifier AST node with a local variable slot in the current function's stack frame).

Then expand from there: add strings, arrays, hashtables. There is nothing to it except that know function bytecode alone won't be enough and you'll static data for storing any literal larger than a number, aka the string in print("hello word"), the literal array in call_some_func([1, 2, 3, 4]), etc..
Look at type checking, closures, macros, GCs, register-based VMs, assembly and native code generation, coroutines, concurrency, etc.. anything you want.

>>107031248
>gpui-component
>6gb
How is that even possible? Are the assets like 8k bitmaps or something?

>>107032560
think of return as "set this value and jump to whatever called me"
think of throw as "set this exception and jump to the last catch"

>>107034109
thanks for the detailed and informative answer. I hope I can regain my passion for programming and reach your level
>But if you want to just start, learn recursive descent and how to parse expressions with it,
So after asking Claude and giving it your post, it seems like I'm going to be going with Crafting Interpreters because it LITERALLY starts the same way you recommended me to start. Thanks again anon

>>107034398
I know very well. The original question was how to have both exceptions and linear types. The problem you can have a linearly typed data structure that is supposed to be consumed in a certain function's scope, but with exceptions there is the possibility that a callee will throw and jump back to the caller of the function that must do the consume call.

>>107034559
par handles that just fine, its a tensor in linear logic

>>107034754
>par
is this what you mean?
https://github.com/faiface/par-lang

AI is wasteful but useful.
Functional Programming is wasteful and useless.

>>107036035
not the name of the language, but ironically, i dont actually see par here. par requires codata with multiple continuations
its also called multiplicative disjunction and represented with an upside down ampersand

>>107036035
>>107036409
https://github.com/duo-lang/duo-lang/blob/main/std/Codata/Par.duo
though i don't know if duo-lang has linear types and it seems to be abandoned anyway

I'm new to Java programming and I'm struggling to compile a test program
The command prompt keeps giving me "Error: Could not find or load main class" and the tutorials I'm using aren't helping

>>107036733
For the simplest test program, you need a public class to contain a public static void method called main that takes an array of strings as its only argument. From there, you can do whatever you like.
Or, you build a test scheme with a framework like JUnit. (See its documentation for how to mark a class as supplying tests.)
Or, you use the new Java 25 way of doing a main class. In that case, you need to use the right runtime or it won't know what to do.
Which one were you trying to do?

>>107036733
unless you getting paid to write java just stop yourself from pain

>>107036733
If you are using Maven or Gradle you need to specify a main class somewhere in the configuration.
If you are making a JAR yourself with javac and jar you need to add a manifest.mf in the root of the jar specifying the main class.
If you're doing just a file, java HelloWorld.java is enough.

>>107037068
>Or, you use the new Java 25 way of doing a main class. In that case, you need to use the right runtime or it won't know what to do.
How do I get the right runtime?
>Which one were you trying to do?
Trying to run a simple Hello World program with a file

>>107037605
You can do
java HelloWorld.java
to run a file since Java 11.
You can download other jdks from: https://adoptium.net/es/temurin/releases

>>107036409
Riiight. I have to admit that my knowledge of linear types doesn't get past the 101 of ressource introduction, consumption and the related borrowing, but I understand where you were getting at. Simply put, the language you envision is functional, has advanced linear types and exceptions don't exist but are emulated with continuations, and because continuations are FP, they still carry linear types.

Concerning exceptions and (basic) linear types, there really are 2 kinds of exceptions: try/catch/throw that maintain a stack discipline and setjmp/longjmp + the jmp_buf value that random access. It's the 2nd type that is the hardest to check, but I think the checking is manageable if the jmp_buf-s can only be passed via function arguments and returns (but can be deeply nested in data strcutures arguments, they don't have to be passed and returned directly), not by global variables. Free variables may be fine, not sure.

For each function, an analysis must be done (transitively looking at the callees) to determine whether each function can throw and which jmp_buf argument is used. Then, for each function definition, check wether each call site is ok. This requires checking, for each call site, which jmp_buf argument are ok to be used in a longjmp and which are not ok.

Any of you faggots ever done Vulkan? I have pretty high CPU load (10%) on a completely static, black scene - no command buffer re-records, just vkAcquireNextImageKHR, vkQueueSubmit2 for timeline semaphore support, and finally vkQueuePresentKHR. vkQueuePresentKHR always waits for full vsync even if I use VK_PRESENT_MODE_IMMEDIATE_KHR, but again, that's not really the problem - the problem is that the CPU load goes up for nothing, while VLC is running at an idle 1.25% in paused mode.

My log shows that no other function takes up that much time:
[404]: vkQueuePresentKHR done
[404]: Loop begin
[404]: vkAcquireNextImageKHR done
[404]: vkQueueSubmit2 done
[420]: vkQueuePresentKHR done
[420]: Loop begin
[420]: vkAcquireNextImageKHR done
[420]: vkQueueSubmit2 done
[437]: vkQueuePresentKHR done
[437]: Loop begin
[437]: vkAcquireNextImageKHR done
[437]: vkQueueSubmit2 done
[454]: vkQueuePresentKHR done
[454]: Loop begin


Each frame loop happens approx. 16-17 ms later. Sleeping for 10-14 ms before the call to vkQueuePresentKHR doesn't do anything to affect the load. How likely is it that the presentation engine has to convert the swapchain image to whatever DWM is internally expecting, and that's where the load comes from, and that it's not just a busy-wait screwup? How do I even check this? Image format is VK_FORMAT_B8G8R8A8_SRGB, space is VK_COLOR_SPACE_SRGB_NONLINEAR_KHR, usage is VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT, sharing mode is VK_SHARING_MODE_EXCLUSIVE.

How the heck do I create a new repo from github mobile app on android? do I need to use the browser ver first and then update from the app? it makes no damn sense

>>107003819 (OP)
I look like this irl

>>107037703
>exceptions don't exist but are emulated with continuations
its not "emulated", continuations are goto, an exception is a different return path, the only difference is how often you "catch" (i.e. set that return path) but the same thing applies with regular return - if you don't need to run extra code after it you don't need to change the return address.
all of what you said is unnecessary because it's a language based on the sequent calculus (look up mu mu tilde, though that isn't automatically about being linear) type checked to be linear. so it is linear. could you add primitives that fuck it up? yes... don't
all your interfaces and shit need to be properly typed with the various combinators and custom linear [co]datatypes, e.g. how multiplicative (use both) and additive (use 1) pairs differ

>>107039988
probably less than you and no clue other than your assumptions seem reasonable that either its busy waiting or converting
have you tried profiling beyond just a logger?

I wish I was better at math. Instead I sometimes just randomly change the algorithm until it works.

i still dont know what this means

>>107041082
In assembly this is just jmp address

>>107039988
Are you polling with nanosleep 0 on your threads? I get about the same CPU usage when I do that.

>>107040782
>sometimes just randomly change the algorithm until it works.
Welcome to engineering

>>107041082
Task<T> is a generic structure over some unconstrained type T(ie it has one generic argument that can accept any type).
The structure consist only of one, unnamed element(1-tuple/newtype/wrapper) which is an Option(Maybe/can be present or be None) of a pinned Box(heap allocated RAII value that cannot be moved/has no move constructor in C++) and that Box points to some dynamically dispatched(ie has concrete type discarded and is accessible through vtable) value that implements Stream(it's probably an async interator, ie iterator that you need to .await for) which yields Action<T> elements, Send(it can be send to another thread) and 'static lifetime meaning it does not hold references to any non-static values.
So in practice, it's just a fat pointer which can be null or it can point to any value on heap(and the vtable corresponding to its concrete type) that is an async iterator producing some Action<T>s and can be safety moved to different threads and is not limited by any lifetime.

>>107040674
>all of what you said is unnecessary because it's a language based on the sequent calculus
your language but mine, mine is imperative. It's not unnecessary complexity, it's strategically placed complexity. You can't avoid software complexity but you can choose where to put it in the hope that it pays of.

In a FP language with linear types, sure the semantic checking is a lot simpler than doing the equivalent checking for an imperative language and you can compile FP control flow constructs to efficient control flow machine code, but how much complexity is there so that the FP language allocates and handles data?

The FP language relies on immutable so it requires a very complex GC to not run dog slow, and it will still run slow because of the GC and because of excessive memory usage and poor locality, compared to a low level manually memory managed imperative language. The FP language user must jumps to if wants to *approach* the speed of the low level imperative language, there is a Haskell presentation about what a guy did to make it run fast, it was not pretty.

>>107039988
Load up vkcube in your preferred API analyzer (RenderDoc, Nsight, PIX) and see what they're doing to get ~1% usage

>>107039988
Use a profiler and see where you program is wasting time.

>>107037605
>newmaid wants to learn Java
>mfw
Post your code and how you are trying to compile it.

>>107034303
memory leak maybe?
anon should open a issue on github.

I was hired to do Static Code Analysis. To get my feet wet, they want me to write them some articles regarded SCA. Give me a subject.

>>107044207
this isn't the GPT prompt, you need to go to chatgpt if you want examples.

>>107044300
I'm generating a book on the subject atm. I just wanted some human touch.

My boss (I'm the the static analysis anon) is going to buy me a phone. Recommend a cheap phone that has a QWERTY physical keyboard, and is safe. Thanks,

>>107043008
Immutability simplifies the gc, there is no escape analysis because children can't outlast their parents. The boxing and copying is something else.

>>107043008
ok but you should be basing it on sequent calculus still

>>107044734
Immutability could be achieved via SSA, though. Keep in mind that SSA is isomorphic with CPS and ANF (both functional ILs).

>>107036733
use an IDE
don't listen to tryhards compiling from terminal, that's just adding unnecessary problems so they can feel validated for explaining busywork

>>107045976
Sounds like a good idea

When I'm implementing an algorithm sometimes it doesn't work and I have no idea so I rewrite the whole thing or block I think is broken. How can I read the code and build an accurate model of what it does so I can put into words exactly what is wrong with the code

"I have no respect for my code"

>>107046367
>respecting code
it's literally just text files to be processed by the compiler jew

>>107046334
Start by skimming the code to identify its inputs, outputs, and purpose. Write down what the code is supposed to do. Pick one example input and mentally step through the code line by line. Track how key variables change and note where results differ from expectations. Add small, focused print statements or assertions to see what’s really happening. When you find a mismatch, describe it clearly: what you expected, what you saw, and why it’s wrong. Fix only the part that breaks the logic instead of rewriting everything. Capture the failing example as a test so it never breaks again. Keep an eye on stupid bullshit like off-by-one errors, mutation bugs, and bad boundary checks. Finally, summarize every bug as “intent observed cause fix”.

>>107046334
Use debugger

>>107046334
In c.
And I should say how to learn to build a more accurate model of what the code does
I tried structured programming, but for many algorithms control flow isn't the pain point; usually it's more "are these statements returning what I think they are returning"
Like I can learn simple stuff like bools being cast to int when you use some unary operators on them

>>107046456
>>107046497
I generally know which line or black is wrong, the problem is figuring out what is wrong with my mental model of the code. Why did I write that line, what did I think it did, why does it not do that. not what does the line actually do

>>107046596
Write comments

>>107045976
>IDE

>>107045976
>>107036733
Use C instead of Java

>>107046845
>>107046334
>>107046596
>How can I read the code and build an accurate model of what it does
You need to write programs in a manner that it makes the logical structure of how it works easily apparent and recognizable. For that you need to recognize patterns in the algorithm. Each algorithm is not a special snowflake, they all have resemblances with other algorithms and use similar patterns with related algorithms, you need to recognize those. Then you figure a good way to implement the algorithmic patterns that makes them run well and make the data structure and control flow easily recognizable when read by a human.

>>107046845
sorry, you shouldn't have been quoted in >>107046985

Why learn to program at all when society believes that 6 trillion shekels is enough to make a bot that can program better than all humans put together? Isn't it our duty as programmers to TRUST other people, how would your programs even work of the documentation was wrong? If every function spec or contract FAILED. Face it trust is the foundation of our society, thus we must all get behind AI, else society has already failed.

fun charge_time kO uF =
let val r1 = kO * 1000.0
val r2 = 10.0 * 1000.0
val c1 = uF / 1000000.0
in 1.05 * c1 * (r1 + r2) end

map (fn k => charge_time k 1000.0) [0.0, 100.0, 200.0, 500.0, 1000.0]
SML is so comfy

>>107047810
haskell is even comfier

>>107047698
Programmers deal in logic, take your meds.

>>107047698
>society has already failed
this but unironically

whats a good use case for *** in C?
like pointer to pointer to pointer?
I've seen ** use cases but not ***

>website is rate limited after a few megs
>http library takes callback
>throw in callback when limit reached
>check if the error is my error
>make a new http client object
>set range field
feels dirty but it just works

>>107048544
unflattened tensors in C, so you can do m[i][j][k].
an rgb image is a stack of 3 matrices.

>>107048701
I found one

https://docs.gtk.org/gtk3/func.parse_args.html

seems it needs access to argv's first pointer
so pointer-to-array-to-pointer

this way it can straight up replace argv

>>107045976
using netbeans with java 8 was great. but everything changed when java 9 attacked.

>>107048544
** lets you insert on the head of a list without changing its address, so ig *** would be the same with a list of lists

>>107048772
yeah thanks for the input

so it lets you change a ** by getting the first member of the array and with this you can create your own array and straight up replace it if you got the reference to it.

atleast thats how I think gtk's ***argv works

>>107048753
>I found one
yeah I just told you

>>107048759
I have Netbeans 6 with Java 6, Netbeans 11 with Java 8, and Netbeans 21 with Java 21 at my job.

>>107048861
with netbeans and java 8 it was great to just right-click and select build an image (or something like that), and it would create a self-contained deployable project. no need to install jre.
now you need retarded stuff like jpackage or something.

>>107048884
I install tomcat and a jre for my victims. They like fiddling with the logs and the settings, makes their managers feel important instead of hostages.

>>107003819 (OP)
Maid bros, why didnt you share this here before?
https://www.youtube.com/watch?v=yrFYVd8BiL8

>>107048952
>meanwhile at the science foundation

I've began throwing shit at the wall to see if it sticks. I got a job now so I don't particularly have to grind anymore. But the job is in static analysis so I need to learn as much as I can.

So I've began ruminating a JavaScript interpreter. I call it "QuisheJS". Give me a better name ffs.

I generated a 54-chapter book for it with Sonnet 4.5. Of course I'm not going to use a single line of code form it. But these 'dossiers' are like a compass, they 'orient' me as of what should my next step be.

https://chubak.neocities.org/quishe-dossier

I just texted my boss about this travesty:


void foo(FooObj &&) {}

void main() {
FooObj fo;
foo(std::move(fo));
fo.bar = 1; // <-- wat?
}



A good static analyzer should pick up on this. RValue refs are retarded anyways.

making text out of gifs for shitposting

>typescript doesn't allow multiple constructors for a class
I've been happy with every language design choice until this. What the fuck? How in the world is
class shitfuck{
constructor (param1, param2?) {
if (param2){
...
}
}
}


better than

class shitfuck{
constructor (param1) { }
constructor (param1, param2) { }
}

>>107050984
the issue isn't with typescript, javascript doesn't have method overloading in the usual sense. although this seems like something typescript should be able to interpret and translate at compile time with the type system.

>>107048544
setting a ** from a function

void parse_args(char *str, int *argc, char ***argv);

int argc;
char **argv;
parse_args("foo bar", &argc, &argv);

U8 *str=MStrPrint("Hello Terry!\n");
"%s", str;

Free(str);

this thread is inane

>>107051141
in lua this is just
print("Hello Terry!")

>>107050984
Just use static methods as constructors. Overloaded constructors are dumb anyways.

>>107051016
>although this seems like something typescript should be able to interpret and translate at compile time with the type system.
class Foo {
constructor(num: number) {}
constructor(str: string) {}
}
class Bar {
constructor(num: number) {}
constructor(str: string) {}
}

let Kek: any = Math.random() ? Foo : Bar;
let kek = new Kek(123);

What should this compile into?

>>107052760
>Overloaded functions are dumb anyways.
FTFY

Working on my proglang. Implementing generators. It kinda bends my brain to think about stacks and switching between them. Sadly it must be done.

>>107052844
can chatgpt generate code in it yet?

>>107052875
More or less. It's a lisp so it already knows the basics. It can't get the particulars right unless instructed.

It's been awesome for discussing implementation ideas tho. Just shooting ideas back and forth until something solid forms in my mind. Basically a help vampirism sponge.

>>107052921
Good luck anon. I managed to land a job in PLT after 2 years of grokking it non-stop. Maybe you'll be able to get lucky as well.

Just one thing, if you wanna chase PLT in a professional manner, learn to implement standards, not ass-pull languages out of thin air. Nobody wants to learn your damn syntax, and it'll end up a toy that even you yourself won't use. I've generated a dossier on Scheme using Sonner 4.5 by feeding it loads of books and papers, maybe you can manager to implement Scheme based on it:

https://chubak.neocities.org/scheme-dossier

That is, if you're interested in PLT as a profession. If you just wanna have fun, go ahead. "Make a Lisp". One of the rites of passage for any programmer is to MAL, and MAL hard.

https://github.com/kanaka/mal

>>107052988
That's fucking awesome anon. Congratulations.

I'm in a hobbyist situation. Have a job outside tech. I quit tech at an early age because I wanted to work on OSes and languages, not be miserable making dumb apps which is what actually pays. Wouldn't mind pivoting if I could get an actual job along these lines.

I think I'm good with standards. I'm writing my own implementation of ELF as I go along, even hacked in my own extensions for my use cases, everything within the confines of the standard. Eventually I'll have my own dynamic loader. I'm also conforming to Linux ABIs. Haven't dealt with anything obnoxiously huge like the C++ ABIsc.. Yet.

As for using my language, I do want to use it... I have a static site generator that I plan to rewrite in my language, hiccup style. That should be a good exercise for it, right?

>>107053167
Thanks man. This means a lot to me.

I know where you're coming from. i want to work on cool stuff too. But I decided to persevere instead of giving up. I considered launching my own business. But my efforts finally paid off. Knock on wood.

I'm working on a documentation language for my own use as well. I just can't lock on the damn syntax. This paper has made me wary of 'taxonomy' of documentation languages:

https://dl.acm.org/doi/pdf/10.1145/3632865

and I wanna make a (Article, TemplateProgram) language similar to Typst. I watched a Dylan Beattie talk called "The Web We Almost Got" and he said, HTML shoulda looked like this:

~h1{header}


Because tilde is the least used letter in ASCII. And I wanna base my language (Dukette) on that. But I am confused as of where to begin.

>>107053250
>But I decided to persevere instead of giving up.

Takes courage to do that. Based

Paper you cited is amazing. I was just reading about document languages a week ago. The woes of markdown, ambiguous syntaxes, how the alternatives approached the problems. The scribble case study is particularly relevant for my case... I'm gonna read it carefully

>>107053358
Can you link to the article about documentation languages you were reading? Thanks.

>>107053383
Sorry I just can't seem to find it anymore. It wasn't a peer reviewed academic article or anything, just a blog post from one of the markdown alternatives showing some examples of syntax ambiguities and how they avoid it. Led to some annoying syntax and I forgot about the whole thing.

https://www.youtube.com/watch?v=SZOr0m-K5PQ
https://github.com/InsomniacGames/ig-cachesim
Neat

am i stupid or is the "programming in lua" book is too advanced for total beginners. at least the fourth edition is, maybe the online version/first edition is easier?

i say this because everyone points at it as the best place to learn for total beginners but reading it it shows you how the 8 queens problem would be solved in lua and the exercises expect you to tweak the example. all it's been shown so far in the book is simple stuff like global variables and naming conventions.

is there an easier way to get into lua for programming beginners? or should i just skip this sort of stuff in the book and then come back at it later?

>>107055553
most programming books about a specific language are just about learning the language, not learning the programming basics as a beginner, there's other books for that

>>107055553
if you really can't figure it out then yeah skip it
or post it here, someone will probably solve it

>>107055553
N queens filtered me in uni (using scheme), I wonder if I could solve it now.

>>107055553
I have never read a programming book

is state machine actually a meme in UI development?

retard

>>107055913
State machines are very useful for certain tasks, which includes some UI flows like multi step forms.

>>107056128
I've look at it and its bloats up your app big time, makes everything hard to read

I'm using QML and see that signals and slots are already pretty good enough

>>107056162
>I've look at it and its bloats up your app big time, makes everything hard to read
This entirely depends on your implementation.

>>107056221
will it help for example say I have a table that has data in it, from a model that represents some data from the user's own settings like plugins installed

In the table theres an remove plugin button, which then needs to be reflected in the settings and accross the app for example without having to hard restart the app. This for me become sagetti code in the backend with a lot of crisscross

>>107056354
A table with data is not a state machine.

>>107003819 (OP)
I'm going to masturbate to Asuna.

>>107056354
>In the table theres an remove plugin button, which then needs to be reflected in the settings and accross the app for example without having to hard restart the app. This for me become sagetti code in the backend with a lot of crisscross
You can easily deal with kind of thing by having a "global" internal application state. When the user modify a setting, it modifies the application internal state, and the rendering part and logical/model part of the application read that state to know what to do/what to render.
You need to map each settings to the application state. Sometimes it's a direct 1 to 1 correspondence, sometimes a single setting can affect multiple internal state fields, sometime the value a single state field is dependent of the values of multiples settings, etc...
This can be seen as a state machine transitions but it can also be seen as logic rules. They are equivalent. Essentially, and for a simple logic model like this one, you'd have internal_state_parameter = logic expression that is a function on 1 or more setting but you can also do it the other way around which is, when a setting parameter is modified, recompute the value of every internal state that depends on it.

It took so much effort to draw these little depth and stencil buffer views.

>>107056676
My settings file has a signal onSettingsChanged, where I can slot this in anywhere. Except that its not fine grained, and may just lag the whole application this way, doing it per setting is tedious and feels not right

>>107056775
>My settings file has a signal onSettingsChanged, where I can slot this in anywhere
>file has a signal onSettingsChanged
>I can slot this in anywhere
what does it mean?

Changing settings is a rare event, so why would the application lag? what are you doing that would be costly?

>having trouble thinking about a simple recursive function to walk through directories

I should just end myself at this point.
At least I'm not running to chatgpt yet.

>>107052670
>in lua this is just

HolyC
I64 StrBin2Int(U8 *str, I64 r=0){
while (*str) {
U64 ch=*str++;
if(ch=='1'||ch=='0'){
r<<=1;
r+=ch=='1';
}
else throw('CharERR');
}
return r;
}

>>107056695
vulkan or GL

>>107057265
func walk_directory(path file)
{
if (is_file(file)) {
// do action
}
else if (is_directory(file)) {
for file in list_files(file) {
walk_directory(file);
}
}
else {
}
}

>>107057294
>links
>filesystem changes mid iteration

>>107057294
>func walk_directory(path file)
originally wanted to "string file" and then wanted to change it to "string path" but fucked up

>>107057289
Vulkan.

>>107057315
nice

>>107057312
>>links
yep, good catch
>>filesystem changes mid iteration
I don't know to handle this honnestly

>>107057324
make it so the filesystem changes don't happen, you lock the files

>>107057294
Thanks, anon.

>>107057312
>>107057336
This might be overkill for my toy project (static site generator that I probably won't even finish)

>>107057324
NTA but different OSs might have a variety of atomic operations maybe? or you could just clone it which on some OSs won't immediately necessitate any work?

if the recursive directory is for posix C you can use ftw()

>>107057294
pub fn walk_dir(directory: &PathBuf) {
if let Ok(entries) = fs::read_dir(&directory) {
for entry in entries {
// println!("{:?}", entry.unwrap());
if let Ok(direntry) = entry {
// println!("{:?}", direntry.metadata().unwrap());
if let Ok(metadata) = direntry.metadata() {
if metadata.is_file() {
println!("{:?} is a file", direntry.file_name());
}
}
}
}
}
}


The fuck is this language, man

>>107057630
pub fn walk_dir(directory: impl AsRef<Path>) {
fs::read_dir(&directory)
.into_iter()
.flatten()
.flatten()
.filter(|entry| entry.metadata().as_ref().is_ok_and(Metadata::is_file))
.for_each(|entry| println!("{:?} is a file", entry.file_name()))
}

>>107057765
>>107057630
what the fuck

Vulkan tutorials make you write 1k lines of unreadable OOP garbage for a triangle. You can do it in under 300 with procedural code.

is microcommitting a sin?
I commit every little change not to forget what I was thinking.

>>107058351
>You can do it in under 300 with procedural code.
proof? I gave up on vulkan because of how tedious it was.

>>107058421
I stopped using git. I'm a lone programmer, what the fuck do I need git for?

>>107058431
I use it so I can time travel to my old code if I break anything

>>107058429
It's what I did on my first vulkan renderer. Packed structs, macro error handler, procedural style.

in C++ how are you suppose to convey derived type information through an interface that only knows about base types?

Like if I have some central 'object management' system that stores all of my objects as the base type and various subsystems that latch onto it and access those objects to do specific things with various derived types. One subsystem may be a serializer/deserializer and needs to cast the base objects to their true types to serialize them.

store the type on the object?
hash table mapping the related code to a typeid().hash_code()? in the subsystem to ID the object?
something else?

function fibmem(n, memo = {}) {
if (n in memo) return memo[n];
if (n <= 1) return n;
memo[n] = fib(n - 1, memo) + fib(n - 2, memo);
return memo[n];
}

console.log(fibmem(60))


i tried making a fibonachi with memorization built in but it doesnt work in the terminal

>>107058421
just squash before you push

>>107058565
but then I will not be able to discriminate every change I did and why

>>107058555
>function fibmem
> memo[n] = fib(n - 1, memo) + fib(n - 2, memo);
$ node -e 'function fibmem(n, memo = {}) { if (n in memo) return memo[n]; if (n <= 1) return n; memo[n] = fibmem(n - 1, memo) + fibmem(n - 2, memo); return memo[n]; } for (i=0; i <= 10; i++) { console.log(i, fibmem(i)) }'
0 0
1 1
2 1
3 2
4 3
5 5
6 8
7 13
8 21
9 34
10 55

what is a nice subset of c++ to learn? I can barely do anything beyond c style stuff. I avoid derived classes like the plague.

>>107058601
>what is a nice subset of c++ to learn?
All of it.
Except reflection/contracts/profiles obviously, that shit glows

>VK_PRESENT_MODE_SHARED_CONTINUOUS_REFRESH_KHR specifies that the presentation engine and application have concurrent access to a single image, which is referred to as a shared presentable image. The presentation engine periodically updates the current image on its regular refresh cycle. The application is only required to make one initial presentation request, after which the presentation engine must update the current image without any need for further presentation requests. The application can indicate the image contents have been updated by making a presentation request, but this does not guarantee the timing of when it will be updated. This mode may result in visible tearing if rendering to the image is not timed correctly.

Damn that sounds kinda convenient. Sucks that it's Linux/Android only.

>>107057265
>>107057294
>>107057630
>>107057765
good morning sirs

public class FileCollector {
private List<File> files;
private boolean includeDirectories;

public FileCollector includeDirectories(boolean value) {
this.includeDirectories = value;
return this;
}

public List<File> collectFiles(File baseDir) {
files = new ArrayList<>();
collect(baseDir);
return files;
}

private void collect(File dir) {
if (includeDirectories || !dir.isDirectory()) files.add(dir);
if (dir.isDirectory()) {
for (File file : dir.listFiles()) {
collect(file);
}
}
}
}

// usage:
// List<File> files = new FileCollector().includeDirectories(false).collectFiles(baseDir);

>>107003819 (OP)
Corporate businesses should be permabanned from social media if they use AI to repost the same unoriginal slop content. I'm sick of it. I am disengaged. I am unhappy. I am getting unhealthy from it and you're wasting my time.

The mitigation for this is an additional content slop tax for making low effort rubbish for attention.

>>107058549
the serializer is easy because it's just a virtual function.
the deserializer needs a factory.
generally just using an enum does the job for the factory of less than 100 types. More than that (well it depends, the only concern would be build time, and accidentally adding/removing an enum and corrupting a save file due to numbers changing), you probably want to go with a different approach which will probably use UUID's or strings + manually register to the factory instead.
And for downcasting (it could be avoided, but sometimes that's not possible) dynamic_cast works, it's a bit slow but it works (if you don't mind doing UB, and if you don't inherit multiple times, I think you could just cast it raw, the dynamic_cast overhead is similar to calling 30~ warm function calls on msvc, which is like 100 nanoseconds I think?, it's like 3x cheaper on other OS's).
If your situation does not need more than void*, std::any is an option, if you were paranoid about performance, I think it wouldn't be that hard to write your own wrapper / macros so that you can switch between std::any and void* to do benchmarking (since it also has a similar dynamic_cast overhead, I think? on linux std::variant is only 2x faster than std::any, and std::variant is a pain due to header dependencies).

>private data, private mutex, function for any type of access / operation you want to do on it
OR
>public data and public mutex
?

>>107059226
second one is the "whore" paradigm

>>107058822
>if (includeDirectories || !dir.isDirectory()) files.add(dir);
> if (dir.isDirectory()) {
good morning sir

I want to make some sort of machine learning/predictive analytics to predict which a unit of the chemicals facility I work at will have unusual operating conditions within N hours based on either a current list of operating condition deviations or a list of process measurements
Anyone know what kind of ML I can use to do this? I have tons of data to throw at the problem

>>107059226
public Mutex<Data>

new
>>107060072
>>107060072
>>107060072