URL Parsing Failure Issue

[linrongbin16]

There are many people had the internal url parsing exception, which is because of missing use cases.

Please add your use cases here.

Three things are needed:

1. The output of git remote get-url origin.
2. The expected git host url example.
3. The router config you will use to make it work.

[linrongbin16]

http/https protocol

case-1 https://github.com/linrongbin16/gitlinker.nvim

1. git remote url config output: https://github.com/linrongbin16/gitlinker.nvim.git
2. expected git host url: https://github.com/linrongbin16/gitlinker.nvim/blob/c5a4bdd78f3dab3c6017a3b56932b334b94f55c6/.github/ISSUE_TEMPLATE/bug.md?plain=1#L10
3. router configuration: default config

case-2 https://git.samba.org/samba.git and https://git.samba.org/bbaumbach/samba.git

relate ticket: #146

1. git remote url output: https://git.samba.org/samba.git and https://git.samba.org/bbaumbach/samba.git
2. expected git host url: https://git.samba.org/?p=samba.git;a=blob;f=wscript;hb=83e8971c0f1c1db8c3574f83107190ac1ac23db0#l7 and https://git.samba.org/?p=bbaumbach/samba.git;a=blob;f=wscript;hb=8de348e9d025d336a7985a9025fe08b7096c0394#l7
3. router configuration: default config

git@/ssh protocol

case-2 [email protected]:linrongbin16/gitlinker.nvim.git

1. git remote url config output: [email protected]:linrongbin16/gitlinker.nvim.git
2. expected git host url: https://github.com/linrongbin16/gitlinker.nvim/blob/c5a4bdd78f3dab3c6017a3b56932b334b94f55c6/.github/ISSUE_TEMPLATE/bug.md?plain=1#L10
3. router configuration: default config

case-3 ssh resolved host name

related ticket: #87

1. git remote url config output: [email protected]:linrongbin16/gitlinker.nvim.git
2. expected git host url: https://github.com/linrongbin16/gitlinker.nvim/blob/c5a4bdd78f3dab3c6017a3b56932b334b94f55c6/.github/ISSUE_TEMPLATE/bug.md?plain=1#L10
3. router configuration: default config

user have configured host name mapping in ~/.ssh/config, for example:

Host github-personal.com
    HostName github.com
    User git
    IdentityFile ~/.ssh/id_rsa_personal
    Compression no

Then ssh -ttG will output something contains below line:

hostname github.com

Thus resolved host name is github.com (instead of github-personal.com).

case-4 ssh://[email protected]/org/repo.git

relate ticket: #114

1. git remote url config output: (not sure I guess it's) [email protected]/org/repo.git or ssh://[email protected]/org/repo.git (hi @skoch13 could you help confirm?)
2. expected git host url: https://xyz.abc/org/repo/repositories/c5a4bdd78f3dab3c6017a3b56932b334b94f55c6/files?line=1
3. router configuration: [git%.xyz%.abc] = 'https://xyz.abc/{_A.USER}/{_A.REPO}/repositories/{_A.REV}/{_A.FILE}?line={_A.LSTART}'

case-5 ssh://[email protected]/org/repo.git

relate ticket: #187

1. git remote url config output: (not sure I guess it's) [email protected]/org/repo.git or ssh://[email protected]/org/repo.git
2. expected git host url: (I guess it's) https://[email protected]/org/repo/c5a4bdd78f3dab3c6017a3b56932b334b94f55c6/files#L14
3. router configuration: (I guess it's) [username%.github%.com] = 'https://[email protected]/{_A.USER}/{_A.REPO}/{_A.REV}/{_A.FILE}#{_A.LSTART}'

case-6 ssh://[email protected]/repo.git

see: https://git-scm.com/book/en/v2/Git-on-the-Server-The-Protocols#:~:text=on%20your%20system.-,The%20SSH%20Protocol,to%20set%20up%20and%20use.

1. git remote url config output: ?
2. expected git host url: ?

[linrongbin16]

I found several things:

1. https://www.gnu.org/software/emacs/manual/html_node/url/Parsed-URIs.html
2. https://github.com/IonicaBizau/git-url-parse
3. https://stackoverflow.com/questions/31801271/what-are-the-supported-git-url-formats
4. https://git-scm.com/book/en/v2/Git-on-the-Server-The-Protocols

The 3nd link contains many use cases, I will refactor the url parser, and try to cover them.

---

Git url parsing design

Syntax:

1. use {} contains the component keywords of git url (they're also the same thing in the README.md - Fully Customize Urls section).
2. use [] contains optional (0 or 1) components.
3. use []* contains 0 or more components.
4. use []+ contains 1 or more components.

Url patterns:

1. {protocol}://[{username}[:{password}]@]{host}[/{org}]*/{repo}
2. [{username}[:{password}]@]{host}:[{org}/]*{repo}
3. [file://]/[{org}/]*{repo}

This pattern could cover below cases:

• {protocol}://{host}/[{org}/]{repo}, example:
  • https://github.com/org/repo.git
  • ssh://github.com/org/repo.git
  • https://git.samba.com/samba.git
• {protocol}://{username}[:{password}@{host}/[{org}/]{repo}
  • ssh://username:[email protected]/org/repo.git
  • https://[email protected]/org/repo.git
• [{username}[:{password}@]{host}:{org}/{repo}, example:
  • [email protected]:org/repo.git
  • gituser:[email protected]:repo.git
  • git.xyz.com:org/repo.git
• [file://]/[{org}/]*{repo}, example:
  • file:///path/to/your/repo
  • /org/repo

Note: current component user will be deprecated (but not removed to keep backward compatible), renamed to org.

[skoch13]

@linrongbin16 correct, mine is set to ssh://[email protected]/%PROJECT_KEY%/%REPO%.git

[linrongbin16]

@linrongbin16 correct, mine is set to ssh://[email protected]/%PROJECT_KEY%/%REPO%.git

thanks!