are there any nice Git merge conflict resolution tools that work in the terminal? when i look at the text merge conflict I can never remember which code is which, and I really want to see them side by side
it feels like someone must have written some kind of nice tool in Rust
just used memray to find out where my python program was allocating memory and was super impressed by how easy it was to use https://github.com/bloomberg/memray
the steps were just:
1. `memray run python3 my_program.py` 2. copy and paste the snippet it prints out ("memray flamegraph memray-manage.py.37954.bin") to generate a flamegraph 3. open the html flamegraph in my browser
@nprofile1q... personally I sometimes see LLM use that the LLM user says is fine but that does not feel fine to me (like it's not something I would be comfortable doing)
and also sometimes I use an LLM in a way that I think is fine but then it turns out to not be as ok as I thought
I'm not shouting at anyone about it but I guess I understand the skepticism and I feel like we don't have a good shared understanding of what it means to "take care"
@nprofile1q... i don't have any interest in convincing anyone to use LLMs or convincing anyone that it's useful, I think it's cool that many people don't use LLMs and I'd much rather they didn't start if they don't want to
a) I feel like LLMs have helped me learn some programming things that were hard for me to learn previously b) I'm tired of LLMs being relentlessly pushed on us and do not want to participate in that by talking about my positive experiences c) there are also lots of hazards associated with using LLMs to generate code that are not that easy to avoid
sometimes i find it funny that I learned about how inheritance works in object oriented programming in university when really the practice I've settled on for myself is "never implement code sharing with inheritance" which would have been a lot shorter
(i feel like there's a whole story about the history of inheritance out there that could be very interesting to read though)
just noticed "how integers and floats work" came out 3 years ago today
originally this this was a zine about "how computers represent data in binary" (what's little endian? big endian? what's struct alignment? why do 32 bit integers cause so many problems? how do we represent strings?) but we ended up running out of space after we explained integers and floats so never got to struct alignment and null-terminated strings :)
i've been trying out jj and one thing I struggle with is that my workspace feels very... messy? In Git often I have 20 semi-abandoned branches that I might come back to later in 6 months if I want to.
but in jj I don't know how to mark a change as "I don't need you to show this to me in `jj log` but I also don't want to abandon it"
so far i've been using jj on a project for short bursts if I need to rebase a lot and then I go back to Git
(please only suggestions of what you personally do)
I've started using ** in Fish instead of find & xargs and I really love it, instead of writing `find . -name '*.py' | xargs wc -l` I can just write `wc -l **.py`
it hasn't 100% replaced `find` for me but I'm definitely using xargs way less than I used to