Jenner compatibility tests for 1 script

jenner-check/.gitignore

@@ -0,0 +1,3 @@
+# runner scratch — never commit
+*_response.json
+response.json

jenner-check/README.md

@@ -0,0 +1,59 @@
+# Jenner compatibility bundles
+
+This directory was added by a pull request from
+[Jenner](https://jenneranalytics.com). It contains a SAS bundle generated
+from code already in this repository, so you can run `test/marks.sas`
+against the Jenner SAS engine and see the results for yourself.
+
+## What's in here
+
+```
+jenner-check/
+├── README.md                       # this file
+├── run_jenner.sh                   # runner (curl; Linux/macOS/WSL)
+├── run_jenner.bat                  # runner (Windows)
+├── run_jenner.sas                  # runner you can call from base SAS (PROC HTTP)
+└── t001_marks_elementary_tests/
+    ├── script.sas                  # test/marks.sas, self-contained
+    ├── autoexec.sas                # options set before the run
+    ├── input/report.dat            # the repo's own data
+    ├── expected.json               # pinned fields from a passing run
+    ├── expected/                   # human-readable snapshot (log, output, file URLs)
+    └── meta.json                   # provenance + notes on what was adapted
+```
+
+`script.sas` is your `test/marks.sas`. The only change is that the
+external `infile 'statclass1.dat'` was replaced by inline `datalines`
+carrying this repo's own `test/report.dat`, so the bundle is
+self-contained. Every PROC, option, and statement is exactly as written
+in the original.
+
+## How to run it
+
+From inside `jenner-check/`:
+
+```bash
+./run_jenner.sh --all                     # run every bundle
+./run_jenner.sh t001_marks_elementary_tests   # run just this one
+```
+
+On Windows use `run_jenner.bat`. From base SAS (9.4 M5+):
+
+```sas
+%include 'run_jenner.sas';
+%jenner_check_all();
+```
+
+Each run prints `status`, `exit_code`, and the captured log; the full
+listing and all generated graphics come back as downloadable URLs.
+
+## Optional: Jenner Compatible badge
+
+```markdown
+[![Jenner Compatible](https://jenneranalytics.com/badges/jenner-compatible.svg)](https://jenneranalytics.com)
+```
+
+## Don't want future PRs from us?
+
+Reply with `no-more-prs` anywhere in a comment, or open an issue titled
+`jenner-check: opt out`, and we'll stop automated PRs to this repo.

jenner-check/run_jenner.bat

@@ -0,0 +1,43 @@
+@echo off
+rem run_jenner.bat - Windows runner for Jenner compatibility checks.
+rem
+rem Usage:   run_jenner.bat <script.sas> [response.json]
+rem
+rem Submits a single .sas file to api.jenneranalytics.com. For
+rem bundle-aware mode (autoexec.sas + script.sas concatenation) on
+rem Windows, use WSL and invoke run_jenner.sh instead, or wait for the
+rem Windows CI runner that will validate a bundle-aware .bat.
+rem
+rem Output:  response.json contains the API response. Read it back in SAS:
+rem     filename resp 'response.json';
+rem     libname  resp JSON fileref=resp;
+rem     proc print data=resp.root; run;
+rem
+rem Requires: curl.exe (ships with Windows 10+ at C:\Windows\System32).
+
+setlocal
+
+if "%~1"=="" (
+  echo Usage: %~nx0 ^<script.sas^> [response.json]
+  exit /b 2
+)
+
+set SCRIPT=%~1
+set OUT=%~2
+if "%OUT%"=="" set OUT=response.json
+
+set HOST=api.jenneranalytics.com
+
+curl.exe -sS -X POST "https://%HOST%/v1/run" ^
+  -F "script=@%SCRIPT%;type=application/x-sas" ^
+  -F "deterministic=1" ^
+  -F "timeout=60" ^
+  -o "%OUT%"
+
+if errorlevel 1 (
+  echo curl failed with errorlevel %errorlevel%
+  exit /b 1
+)
+
+echo Response written to %OUT%
+exit /b 0