feat: Read the graph and install a core from the command line - #1
Merged
Conversation
The command a person types. It asks the agent what is indexed on this machine and what the indexer found in one repository, and prints a count by kind and by edge, which is what a terminal can say about a graph it cannot draw. Anything that would rather read the answer than look at it asks for JSON and gets the agent's own. It talks only to the agent, never past it to the indexer: the agent is the half that is always up and the half that knows where the indexer landed, so reaching around it would mean learning both. An agent that is not running is reported as what to do about it rather than as a socket error.
Adds ui, which points a browser at the view the agent serves. It asks the agent first, so an agent that is not running is a line saying so rather than a browser landing on an error page. --no-open prints the address for anyone who would rather open it themselves.
Adds install, which is what makes the agent runnable by somebody who is not already set up to run the core by hand. It puts a core on this machine and writes down which one, so the agent knows what to start. Two ways to have it, chosen at install time. A container, which is what exists today. Or a Python program, for when the core is published as a package: until then that path says so plainly and points at the one that works, rather than recording a runtime that would fail to start later. Both put the index in the same place, so it does not matter which one built it, and a container runs as whoever installed so what it writes there belongs to them.
A container that can only see the index reads no files, so an install now records a directory to make visible and the agent mounts it at the path it already has on this machine. Your home, which is where repositories nearly always are; one outside it is not readable this way.
The targets are the ones in the Makefile, so what runs here is what runs when somebody runs it themselves.
The ignore pattern for the built binary was unanchored, so it matched the command directory as well and the entry point was never committed. A fresh clone could not build.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The command a person types. It lists the repositories this machine covers, prints what the code graph holds for one of them, opens the view in a browser, and installs a core to answer all of it.
A core runs in a container, so the repositories it indexes have to be visible from inside one. Paths it is given are mounted rather than assumed.
Checks run for the first time here: formatting, vet, tests under the race detector, and a build. The targets are the ones already in the Makefile, so what runs here is what runs when somebody runs it themselves.
Companion to sourceant/sourceant#128, sourceant/design#1 and sourceant/agent#1, all merged.