README documents no build, run, or protocol details

[dmccoystephenson]

Problem

README.md is two lines: an # SimpleServer heading and one sentence describing what the server does. Nothing operational is recorded anywhere in the repository.

A reader who wants to run this project has to open the source to learn:

• how to compile it (no build file exists, so the command is not obvious)
• which classes are the entry points (SimpleServer.server.ServerApp, SimpleServer.client.ClientApp)
• which port is used (hardcoded to 2000 in ServerApp.java:9)
• what the wire format is (key=value, pairs, defined by Message.java)
• which process values the server accepts (square only, per the switch in Protocol.java:18)
• what the response keys are (success, answer, reason)

Evidence

All of the above was read out of the source files named. README.md was confirmed to be two lines of content.

Proposed change

A README section recording the build command, the run commands for both entry points, the port, the wire format, and one worked request/response example.

Constraint on the fix

Every command and every quoted output line in such a section must be produced by actually running it before the section lands. This repository has no CI, so an untested paste would go uncontradicted indefinitely. Whoever implements this needs a working JDK available.

---

This issue was drafted during a Gardener session (https://github.com/Stephenson-Software/gardener).

drafted by Claude on behalf of Daniel Stephenson

[dmccoystephenson]

A dependency on this issue was created by PR #15, which adds a .gitignore containing a .build/ entry.

That entry anticipates a javac -d .build compile command which no committed file currently records. When the Build section requested here is written, the command documented must be the one whose output directory .gitignore covers — otherwise the ignore rule goes dead and compiler output lands somewhere unignored, which is the failure mode #8 was filed to prevent.

Either the documented command should be javac -d .build, or .gitignore:5 should be updated in the same change to match whatever directory is chosen instead.

---

This comment was drafted during a Gardener session (https://github.com/Stephenson-Software/gardener).

drafted by Claude on behalf of Daniel Stephenson

[dmccoystephenson]

This issue was selected for implementation during a Gardener cycle, and the cycle was then abandoned without a pull request being opened. The reason is recorded here so the deferral is auditable rather than silent.

What blocked it

The constraint stated in this issue's own description — "Whoever implements this needs a working JDK available" — was not satisfiable in the dispatch environment.

javac and java were both found on PATH (/usr/bin/javac, /usr/bin/java), but every attempt to invoke either one was denied by the session's command classifier, including the bare probes javac -version and java --version. nc and python3 were resolvable, but they are of no use without a compiler.

Consequently no source file could be compiled, no server could be started, and no socket response could be observed. A Build/Run/Protocol section written under those conditions would consist entirely of commands that had never been executed and output that had never been produced — precisely the failure mode this issue's "Constraint on the fix" paragraph was written to prevent. Writing it anyway was judged worse than leaving the issue open, because this repository has no CI and an untested paste would go uncontradicted indefinitely.

What was confirmed without a compiler

A full documentation accuracy sweep was performed, since it requires no toolchain. No drift was found:

• The README.md sentence still matches the only process value accepted by the switch in Protocol.java, which is square.
• The GitHub repository description still duplicates that sentence verbatim.
• The @author Daniel Stephenson Javadoc header is present on all seven source files.
• .gitignore covers *.class, .build/, and out/, which remains consistent with the javac -d .build command pinned by the earlier comment on this issue.

The build command to be documented is therefore still javac -d .build, and no change to .gitignore:5 is required.

The rest of the backlog

Issues #10, #11, #12, #13, and #14 were all deferred for the same single reason and are not being commented on individually, so that this record is kept in one place rather than repeated six times.

• Client hostname is hardcoded to "Walter", making ClientApp unrunnable elsewhere #10, Client constructor discards the boolean results of its initialize* helpers #11, Disconnect sentinel is compared against the response instead of the request line #12, and A failed bind is swallowed and start() then busy-loops forever #13 modify files under src/ and cannot be validated, nor can their reported behaviour be reproduced, without a compiler.
• No automated check exists; add a POSIX-sh compile-and-connect script #14 asks for a script whose entire purpose is to compile and connect, and its own text notes that it "cannot be written responsibly without running it."

No open issue in this repository is currently implementable by an autonomous cycle that cannot invoke the JDK.

What would unblock this

Either a dispatch environment in which javac and java may be invoked, or implementation by a human with a local JDK.

---

This comment was drafted during a Gardener session (https://github.com/Stephenson-Software/gardener).

drafted by Claude on behalf of Daniel Stephenson