| 1 | <!DOCTYPE html> |
|---|
| 2 | <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en"> |
|---|
| 3 | <head> |
|---|
| 4 | <title>Contributing: Armed Bear Common Lisp (ABCL) - Common Lisp on the JVM</title> |
|---|
| 5 | <meta charset="utf-8"/> |
|---|
| 6 | <link rel="stylesheet" type="text/css" href="style.css"/> |
|---|
| 7 | </head> |
|---|
| 8 | |
|---|
| 9 | <body> |
|---|
| 10 | <div class="header"> |
|---|
| 11 | <h1>Contributing: Armed Bear Common Lisp (ABCL) - Common Lisp on the JVM</h1> |
|---|
| 12 | </div> |
|---|
| 13 | |
|---|
| 14 | <!--#include virtual="left-menu" --> |
|---|
| 15 | |
|---|
| 16 | <div style="margin-left:auto;margin-right:auto;width:20cm"> |
|---|
| 17 | |
|---|
| 18 | <h2>A quick guide to producing patches for ABCL</h2> |
|---|
| 19 | |
|---|
| 20 | This guide describes how to contribute patches to ABCL, while making |
|---|
| 21 | sure that the patch doesn't introduce regressions. |
|---|
| 22 | |
|---|
| 23 | <ol> |
|---|
| 24 | <li> Build abcl and run the ansi-tests with the pristine tree before making patches |
|---|
| 25 | and store the test results. Invoke the following commands from the |
|---|
| 26 | abcl main directory: |
|---|
| 27 | <code> |
|---|
| 28 | <br/>ant abcl.clean |
|---|
| 29 | <br/>ant abcl.wrapper |
|---|
| 30 | <br/>ant test.ansi.interpreted |
|---|
| 31 | <br/>ant test.ansi.compiled |
|---|
| 32 | </code> |
|---|
| 33 | <br/>The test runs will report where their logs are written, keep those |
|---|
| 34 | files at hand for comparing them with later runs with modified code. |
|---|
| 35 | </li> |
|---|
| 36 | <li> |
|---|
| 37 | Develop your patch. |
|---|
| 38 | </li> |
|---|
| 39 | <li> |
|---|
| 40 | Build abcl and run the ansi-tests with your modification and store the test results. |
|---|
| 41 | The commands are as in the first step. |
|---|
| 42 | </li> |
|---|
| 43 | <li> |
|---|
| 44 | Compare the new test results with the pristine results, if there are no |
|---|
| 45 | additional failures, the patch should be ok. Example comparison for |
|---|
| 46 | linux would be |
|---|
| 47 | <code> |
|---|
| 48 | <br/>diff -u abcl-test-20093726-2037.log abcl-test-20094426-2044.log |
|---|
| 49 | </code> |
|---|
| 50 | <br/>Note that the file names are generated dynamically by the test runs. |
|---|
| 51 | The list of failed tests can be found at the end of the log, so that's |
|---|
| 52 | practically the only thing you're interested in. If the lists don't |
|---|
| 53 | differ, you've successfully created a patch with no ansi-test regressions. |
|---|
| 54 | </li> |
|---|
| 55 | </ol> |
|---|
| 56 | </div> |
|---|
| 57 | |
|---|
| 58 | <div class="footer"> |
|---|
| 59 | <hr /> |
|---|
| 60 | <p>Hosted as part of <a href="https://common-lisp.net/">common-lisp.net</a></p> |
|---|
| 61 | </div> |
|---|
| 62 | |
|---|
| 63 | </body> |
|---|
| 64 | </html> |
|---|
