0

I want to clone the github repository into my memory But I don't seem to find any function that supports cloning to memory Is this possible with mempack or odb

I tried to use odb or mempack to implement cloning to memory I asked ChatGPT, but the code was invalid It is also not possible to clone the repository to the virtual filesystem

PP Kun
  • 11
  • 4
  • 2
    I believe in some (most?) linux distros `/tmp` is a ram-based filesystem, so if you extract there, you shouldn't touch any disks. Not sure about windows however – Filipe Rodrigues Aug 04 '23 at 16:18
  • 2
    Yeah I'm pretty sure that git isn't built to work without a filesystem. – PitaJ Aug 04 '23 at 16:31

1 Answers1

1

I don't believe this is possible out of the box. libgit2 accesses files on the file system, and while you certainly could implement a memory-backed storage as an ODB, I don't think this is implemented by default.

In general, memory-based storage for Git repositories is a lot less useful than you think, because many repositories are large and would consume very large amounts of memory. Even a normal command-line Git process (or libgit2) can map large amounts of memory to read packs and work with objects, so that leaves you with substantially less memory to store data entirely in RAM.

If your goal is to simply improve performance, you can mount a ramdisk on your system and clone a repository to that. How you do that is dependent on your operating system; on Linux, the file system type is tmpfs.

bk2204
  • 64,793
  • 6
  • 84
  • 100