Getting "fatal: repository not found" when "git clone" a URL

Question

This must be something basic but I can't figure it out. I have a github URL that has the form "https://.../tree/main". How do I download the contents? In response to

git clone https://.../tree/main

I get

fatal: repository 'https://.../tree/main/' not found

When I do

git clone https://.../

I get the contents of the root folder but not the subfolder that I want.

What am I doing wrong?

Answer 1

A GitHub HTTPS URL is of the form https://github.com/OWNER/NAME, with an optional .git on the end. The URL you're using, with /tree/main, is designed to render the main branch in the user interface, but that isn't a valid repository and you can't use it with Git.

If it's the case that you want to look at the main branch in the repository, then clone it using a proper URL, and then run git checkout -b main origin/main, which will create a main branch that's a copy of the remote branch. If you're already on the branch main, then that won't work, and you can just look at the repository once it's cloned.

Answer 2

You can see this error also when the repository is empty.

shell> git clone https://repos.example.org/test_001.git
Cloning into 'test_001'...
fatal: repository 'https://repos.example.org/test_001.git/' not found

This means it is not enough to create on the server an empty repository for testing

admin# pwd
/usr/local/www/repos.example.org

admin# git init --bare test_001.git

---

Note: For example, a virtual host in Apache

admin# cat /usr/local/etc/apache24/extra/repos.example.org.conf 
     <VirtualHost *:80>
     ServerName repos.example.org
     DocumentRoot /usr/local/www/repos.example.org
     </VirtualHost>

     <VirtualHost *:443>
     ServerName repos.example.org
     DocumentRoot /usr/local/www/repos.example.org
     SSLCertificateFile /usr/local/etc/ssl/certs/repos.example.org.crt
     SSLCertificateKeyFile /usr/local/etc/ssl/private/repos.example.org.pem
     </VirtualHost>

admin# cat /usr/local/etc/apache24/Includes/usr-local-repos-example.conf 
<Directory /usr/local/www/repos.example.org>
  Options Indexes FollowSymLinks
  DirectoryIndex index.html
  AllowOverride All
  Require all granted
</Directory>

admin# ll /usr/local/www/repos.example.org/test_001.git/
total 40
drwxr-x---  7 admin  www  512 Jul 22 12:36 ./
drwxr-x---  3 admin  www  512 Jul 22 12:36 ../
-rw-r-----  1 admin  www   23 Jul 22 12:36 HEAD
drwxr-x---  2 admin  www  512 Jul 22 12:36 branches/
-rw-r-----  1 admin  www   66 Jul 22 12:36 config
-rw-r-----  1 admin  www   73 Jul 22 12:36 description
drwxr-x---  2 admin  www  512 Jul 22 12:38 hooks/
drwxr-x---  2 admin  www  512 Jul 22 13:00 info/
drwxr-x---  7 admin  www  512 Jul 22 13:00 objects/
drwxr-x---  4 admin  www  512 Jul 22 12:36 refs/

Test the URL https://repos.example.org/test_001.git/ in a browser. You should see the content of the repo

Index of /test_001.git
Parent Directory
HEAD
branches/
config
description
hooks/
info/
objects/
refs/

---

Clone the repo locally on the server and commit something. This will also create the branch master. Then, you'll be able to clone the repo over https

shell> git clone https://repos.example.org/test_001.git
Cloning into 'test_001'...

Answer 3

Alternatively, the solution could be something simpler than the answers above:

1. Go to the GitHub repo you want to clone

2. 'Fork' the repo to your GitHub account

3. Now, from your machine's CLI, do the "usual thing":

   $ git clone https://github.com/<your account>/<your fork> 
   

I found this to be necessary when recently trying to clone some repos at the raspberry pi GitHub site.