Problem
codedb_remote lets you find a file (tree, search) and locate symbols (outline, symbol), but it has no way to actually view a slice of file content from a remote repo. Today the workflow is:
action=search, query=foo → returns {path, line_num, line_text}
- … and now you're stuck. To see the surrounding 30 lines you have to clone the repo locally.
This defeats the "explore an external repo without cloning" use case the tool is pitched on.
Proposal
Add a read action:
GET /:owner/:repo/read?path=src/foo.zig&lines=10-60
Response shape:
{
\"repo\": \"owner/repo\",
\"path\": \"src/foo.zig\",
\"lines\": \"10-60\",
\"content\": \"...\",
\"total_lines\": 250
}
The codegraff snapshot already contains a CONTENT section (see `iterateContent` in `codedb-cloud/src/snapshot.ts`) — implementing this server-side is a ~30-line patch.
The wiki backend would need parallel work in its own repo.
Why
Closes the most-requested workflow gap. A search hit is only useful if you can read the surrounding code.
Problem
codedb_remotelets you find a file (tree,search) and locate symbols (outline,symbol), but it has no way to actually view a slice of file content from a remote repo. Today the workflow is:action=search, query=foo→ returns{path, line_num, line_text}This defeats the "explore an external repo without cloning" use case the tool is pitched on.
Proposal
Add a
readaction:Response shape:
{ \"repo\": \"owner/repo\", \"path\": \"src/foo.zig\", \"lines\": \"10-60\", \"content\": \"...\", \"total_lines\": 250 }The codegraff snapshot already contains a CONTENT section (see `iterateContent` in `codedb-cloud/src/snapshot.ts`) — implementing this server-side is a ~30-line patch.
The wiki backend would need parallel work in its own repo.
Why
Closes the most-requested workflow gap. A
searchhit is only useful if you can read the surrounding code.