Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Write yourself a Git (2018)

wyag.thb.lt

Post reply on HN
Print view View original

31–40 of 111 posts

Re: Write yourself a Git (2018)

#31

Post by AnIdiotOnTheNet »

cabalamat wrote: post #29

I have an admission to make: I don't understand git. By this I mean I have a few simple commands I use (status/add/commit/push/pull) and if I try to do anything more complicated it always ends up with lots of complex error messages that I don't understand and me nuking the repository and starting again. So I think: there must be a better way. I have often thought about implementing a VCS. The idea behind one doesn't…

Yeah, I too don't really understand git. It seems that it was developed without any concern for affording a good mental model of its operation to its users, and thus it is just a complex black box you chant arcane rituals at and hope it doesn't decide to burn your world down. I know I could build a mental model of it if I put enough time into it, but who wants to do that when there's actually useful things to do? So instead when I have to use it to contribute to open source projects I have a sheet of notes with incantations to cover the specific things I've had to do with it in the past.

Re: Write yourself a Git (2018)

#32

Post by cabalamat »

cmroanirgo wrote: post #30
cabalamat wrote: post #29

I have an admission to make: I don't understand git. By this I mean I have a few simple commands I use (status/add/commit/push/pull) and if I try to do anything more complicated it always ends up with lots of complex error messages that I don't understand and me nuking the repository and starting again. So I think: there must be a better way. I have often thought about implementing a VCS. The idea behind one doesn't…

I'm like you. I use SourceTree to get a 'visual grasp' on what I find to be the noise of git commands. However, if you're into command line, you can try fossil: it's got lots going for it. Your idea of a "user's mental model" might land you into trouble though, because all of us come from different backgrounds (subversion, SSafe, git, HG...) and they all maddeningly redefine terms in different ways (eg branch, forks,…

> Your idea of a "user's mental model" might land you into trouble though

If I do do this, I will explicitly lay out the user's metal model in the documentation at the start. Then it will be the user's fault if they can't be bothered to read it.

Re: Write yourself a Git (2018)

#33

Post by neals »

cabalamat wrote: post #29

I have an admission to make: I don't understand git. By this I mean I have a few simple commands I use (status/add/commit/push/pull) and if I try to do anything more complicated it always ends up with lots of complex error messages that I don't understand and me nuking the repository and starting again. So I think: there must be a better way. I have often thought about implementing a VCS. The idea behind one doesn't…

I think this is a large majority of Git users. The last company I worked at had 30 good developers, 1 of which I think really deeply understood git.

So I guess 97% of users don't really get git.

Re: Write yourself a Git (2018)

#34

Post by cabalamat »

AnIdiotOnTheNet wrote: post #31
cabalamat wrote: post #29

I have an admission to make: I don't understand git. By this I mean I have a few simple commands I use (status/add/commit/push/pull) and if I try to do anything more complicated it always ends up with lots of complex error messages that I don't understand and me nuking the repository and starting again. So I think: there must be a better way. I have often thought about implementing a VCS. The idea behind one doesn't…

Yeah, I too don't really understand git. It seems that it was developed without any concern for affording a good mental model of its operation to its users, and thus it is just a complex black box you chant arcane rituals at and hope it doesn't decide to burn your world down. I know I could build a mental model of it if I put enough time into it, but who wants to do that when there's actually useful things to do? So…

> It seems that it was developed without any concern for affording a good mental model of its operation to its users

I think Linus' design goal was something that runs as quickly/efficiently as possible on large repositories.

> it is just a complex black box you chant arcane rituals at and hope it doesn't decide to burn your world down

My feeling too!

Re: Write yourself a Git (2018)

#35

Post by derekp7 »

cabalamat wrote: post #29

I have an admission to make: I don't understand git. By this I mean I have a few simple commands I use (status/add/commit/push/pull) and if I try to do anything more complicated it always ends up with lots of complex error messages that I don't understand and me nuking the repository and starting again. So I think: there must be a better way. I have often thought about implementing a VCS. The idea behind one doesn't…

Look for a video on YouTube called Git Happens. I've found it fairly effective with my coworkers. It doesn't go over the command syntax, but instead dives into a logical overview of the underlying data structures.

Re: Write yourself a Git (2018)

#36

Post by driusan »

Having written my own git client, I can tell you that "the most complicated part will be the command-line arguments parsing logic" doesn't go away. I wouldn't be surprised to wake up one day and find someone published a proof that NP != P, and the proof involved trying to parse the git command line.

Re: Write yourself a Git (2018)

#37

Post by Nr7 »

cmroanirgo wrote: post #30
cabalamat wrote: post #29

I have an admission to make: I don't understand git. By this I mean I have a few simple commands I use (status/add/commit/push/pull) and if I try to do anything more complicated it always ends up with lots of complex error messages that I don't understand and me nuking the repository and starting again. So I think: there must be a better way. I have often thought about implementing a VCS. The idea behind one doesn't…

I'm like you. I use SourceTree to get a 'visual grasp' on what I find to be the noise of git commands. However, if you're into command line, you can try fossil: it's got lots going for it. Your idea of a "user's mental model" might land you into trouble though, because all of us come from different backgrounds (subversion, SSafe, git, HG...) and they all maddeningly redefine terms in different ways (eg branch, forks,…

Using a git client with a graphical user interface has definitely helped me "grok" git as well. I use Git Extensions myself.

Re: Write yourself a Git (2018)

#38

Post by rob74 »

lapinot wrote: post #11
ssivark wrote: post #9

I have nothing to directly comment on the tutorial. Just a tangential mention regarding the tedious argument parsing boilerplate in Python, I have found Python Fire to be much more convenient: https://github.com/google/python-fire It would have shaved off another 15-20 lines from the 503 line example ;-)

Yeah, i couldn't help but notice the first piece of code is an ugly "switch case". There's a python idiom for this, it's putting your functions in a dictionary and doing something like `cmds_dict.get(args.command, default)(args)`. I guess we all have our religious habits for argument parsing (more of a docopt-er myself).

I guess the author wanted the code to also look friendly to people who are not that familiar with Python - everybody understands a switch, while what you are describing would probably puzzle some people...

Re: Write yourself a Git (2018)

#40

Post by Jupe »

driusan wrote: post #36

Having written my own git client, I can tell you that "the most complicated part will be the command-line arguments parsing logic" doesn't go away. I wouldn't be surprised to wake up one day and find someone published a proof that NP != P, and the proof involved trying to parse the git command line.

Case in point:

https://git-man-page-generator.lokaltog.net/

:)

Post reply on HN