4 ms·

It was amazing that within ~24hrs of Linus releasing `.git` there were corresponding implementations of the backing data store within ~1 week for all the other

It was amazing that within ~24hrs of Linus releasing `.git` there were corresponding implementations of the backing data store within ~1 week for all the other "major" competing version control systems (eg: hg, svn, darcs) and written in just about every possible language. The core of git is brutally simple and efficient, and maps incredibly well to "a linked list (and hash) of files serialized to disk". Someone wrote somewhere "I used to wonder if Linus was a genius but Git proves it to be true. Anybody could have (re-)written a *NIX implementation, but Linus invented Git from whole cloth." https://perl.plover.com/classes/git/samples/slide005.html https://perl.plover.com/classes/git/samples/slide005.html https://perl.plover.com/classes/git/samples/slide009.html https://perl.plover.com/classes/git/samples/slide009.html

ncmncm 5y ago
But, he didn't. Linus copied the design from Monotone, by Grayson Hoare (who also made the original pre-pre-1.0 Rust).
- ncmncm 5y ago
  Graydon. (Damn auto-correct!)
- homarp 5y ago
  https://lkml.org/lkml/2005/4/6/121 https://lkml.org/lkml/2005/4/6/121 "PS. Don't bother telling me about subversion. If you must, start reading up on "monotone". That seems to be the most viable alternative, but don't pester the developers so much that they don't get any work done. They are already aware of my problems ;)" time for monotone to be rewritten in Rust
  - ncmncm 5y ago
    Time for Git using SQLite for back-end storage. Git has corrupted its own store too many times for me ever to trust it anymore. Almost ready to switch to Fossil.
    - ramses0 5y ago
      Actually it's very likely that Git is exposing a failing or unreliable underlying storage medium (or memory?). I can't find the reference for it, but b/c git is so aggressive about actually checking hashes, it sometimes exposes unreliability that gets attributed to git, but is actually the hardware. Check for "memtester", "smartctl", and "badblocks" ... understand what they do and see if they turn up anything.
      - ncmncm 5y ago
        Possibly, but audits have also shown Git not correctly handling file system error reports. We also have file systems not correctly handling disk error reports. Last I saw, volume managers were not passing media error reports through to file systems. And, Intel has done its part to eliminate ECC from memory and buses. I guess what I am saying is that there seems to be more than enough blame to tar everybody.

ncmncm 5y ago

But, he didn't. Linus copied the design from Monotone, by Grayson Hoare (who also made the original pre-pre-1.0 Rust).

ncmncm 5y ago
Graydon. (Damn auto-correct!)
homarp 5y ago
https://lkml.org/lkml/2005/4/6/121 https://lkml.org/lkml/2005/4/6/121 "PS. Don't bother telling me about subversion. If you must, start reading up on "monotone". That seems to be the most viable alternative, but don't pester the developers so much that they don't get any work done. They are already aware of my problems ;)" time for monotone to be rewritten in Rust
- ncmncm 5y ago
  Time for Git using SQLite for back-end storage. Git has corrupted its own store too many times for me ever to trust it anymore. Almost ready to switch to Fossil.
  - ramses0 5y ago
    Actually it's very likely that Git is exposing a failing or unreliable underlying storage medium (or memory?). I can't find the reference for it, but b/c git is so aggressive about actually checking hashes, it sometimes exposes unreliability that gets attributed to git, but is actually the hardware. Check for "memtester", "smartctl", and "badblocks" ... understand what they do and see if they turn up anything.
    - ncmncm 5y ago
      Possibly, but audits have also shown Git not correctly handling file system error reports. We also have file systems not correctly handling disk error reports. Last I saw, volume managers were not passing media error reports through to file systems. And, Intel has done its part to eliminate ECC from memory and buses. I guess what I am saying is that there seems to be more than enough blame to tar everybody.