- 
        Couldn't load subscription status. 
- Fork 3
Rebase the mcp/registry dependency to v1.3.3 #297
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Conversation
| 🔧 MCP Server Tool List UpdatesThe tool lists for modified MCP server specs have been automatically updated using  Summary
 This comment is automatically generated and will be updated as the workflow progresses. | 
1cf745f    to
    03298de      
    Compare
  
            
          
                go.mod
              
                Outdated
          
        
      | module github.com/stacklok/toolhive-registry | ||
|  | ||
| go 1.24.5 | ||
| go 1.25 | 
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Does ToolHive import this at all? If so, we're gonna have some trouble. We're on 1.24 in toolhive because that's what UBI currently supports.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
yeah, I was unsure about this too but it comes from registry which is unfortunately on 1.25, would it be possible to bump UBI?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It is not. The version wil bump in a month or so
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ah, that's a bummer. Let's figure out something then
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
made it to work with 1.24.6 👍
| pulls: 8029 | ||
| last_updated: "2025-10-21T02:31:33Z" | ||
| repository_url: https://github.com/StacklokLabs/ocireg-mcp | ||
| target_port: 8080 | 
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
the oci registry MCP supports the MCP_PORT environment variable, so the target port is not needed. The same is true for most other Stacklok MCP servers.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This comes from the upstream format where we have to set an explicit value if the server is using an HTTP-based transport type. I guess we can still overwrite it in the runtime by passing the MCP_PORT env var, right?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we could. But that's not how toolhive works. If you dont specify a port, toolhive will assign one. If you do specify it, toolhive will take the one that's been specified. So, ultimately this kind of setup where containers have the same port will result in conflicts and not being able to run MCP servers of HTTP or SSE transports if they have the same port.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
wait, I might have gotten this wrong. My understanding is target_port is the port that the server is implemented to listen on (inside the container)? From there on if that port is free on the host, toolhive may use it, otherwise it will choose a random one, right?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
toolhive will use it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
And what happens if that port is not available?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I agree we should always have a clear specification of the default target port.
In case an env variable can override the default, shouldn't we identify it with a specific attribute?
  - name: MCP_PORT
    overrideMcpPort: true
    description: HTTP server port (HTTP mode only)
    required: false
    default: "8080"
BTW: I can't see this MCP_PORT variable in the oci-registry server (only in the GH repo)
Signed-off-by: Radoslav Dimitrov <[email protected]>
Signed-off-by: Radoslav Dimitrov <[email protected]>
Signed-off-by: Radoslav Dimitrov <[email protected]>
Signed-off-by: Radoslav Dimitrov <[email protected]>
…:\n- sqlite\n\nAutomatically updated using 'thv mcp list' command.\n\nCo-authored-by: rdimitrov <[email protected]>
Signed-off-by: Radoslav Dimitrov <[email protected]>
Signed-off-by: Radoslav Dimitrov <[email protected]>
ceb8d9c    to
    7059b7f      
    Compare
  
    Signed-off-by: Radoslav Dimitrov <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
shouldn't we have separate source files for upstream-to-toolhive and viceversa, to make the scope clearer?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks! I'll split it right away 👍
The following PR: