From 5bc53cfe53d0afe813b19f28d6151273e7b86499 Mon Sep 17 00:00:00 2001
From: sten <OrcusMaximus@gmail.com>
Date: Tue, 12 Sep 2023 17:01:45 +0100
Subject: [PATCH] Update PhutilCowsay.php to work for small cows

Summary:
Update PhutilCowsay.php to work for small cows.

In doing so, we also simplify the code to just use multiline regexps rather than trying to parse a line at a time.

cowsay(cow='small'){{{What about me?}}}

Test Plan:
Check cowsay works for the built in non-perl cow:
```
cowsay(cow='companion'){{{Built in}}}
```

Test all the perl cows:
```
cowsay(cow='bunny'){{{Testing bunny}}}
cowsay(cow='cower'){{{Testing cower}}}
cowsay(cow='daemon'){{{Testing daemon}}}
cowsay(cow='default'){{{Testing default}}}
cowsay(cow='dragon-and-cow'){{{Testing dragon-and-cow}}}
cowsay(cow='dragon'){{{Testing dragon}}}
cowsay(cow='elephant'){{{Testing elephant}}}
cowsay(cow='eyes'){{{Testing eyes}}}
cowsay(cow='flaming-sheep'){{{Testing flaming-sheep}}}
cowsay(cow='head-in'){{{Testing head-in}}}
cowsay(cow='kitty'){{{Testing kitty}}}
cowsay(cow='koala'){{{Testing koala}}}
cowsay(cow='meow'){{{Testing meow}}}
cowsay(cow='moofasa'){{{Testing moofasa}}}
cowsay(cow='moose'){{{Testing moose}}}
cowsay(cow='mutilated'){{{Testing mutilated}}}
cowsay(cow='satanic'){{{Testing satanic}}}
cowsay(cow='sheep'){{{Testing sheep}}}
cowsay(cow='skeleton'){{{Testing skeleton}}}
cowsay(cow='small'){{{Testing small}}}
cowsay(cow='squirrel'){{{Testing squirrel}}}
cowsay(cow='stegosaurus'){{{Testing stegosaurus}}}
cowsay(cow='supermilker'){{{Testing supermilker}}}
cowsay(cow='surgery'){{{Testing surgery}}}
cowsay(cow='turkey'){{{Testing turkey}}}
cowsay(cow='turtle'){{{Testing turtle}}}
cowsay(cow='tux'){{{Testing tux}}}
cowsay(cow='www'){{{Testing www}}}
```

Reviewers: O1 Blessed Committers, valerio.bozzolan

Reviewed By: O1 Blessed Committers, valerio.bozzolan

Subscribers: tobiaswiese, valerio.bozzolan, Matthew, Cigaryno

Differential Revision: https://we.phorge.it/D25436
---
 src/utils/PhutilCowsay.php                | 41 ++++++++++-------------
 src/utils/__tests__/cowsay/cube_perl.test |  3 +-
 src/utils/__tests__/cowsay/small.expect   |  7 ++++
 src/utils/__tests__/cowsay/small.test     | 12 +++++++
 4 files changed, 39 insertions(+), 24 deletions(-)
 create mode 100644 src/utils/__tests__/cowsay/small.expect
 create mode 100644 src/utils/__tests__/cowsay/small.test

diff --git a/src/utils/PhutilCowsay.php b/src/utils/PhutilCowsay.php
index ca0290b9..aeabf145 100644
--- a/src/utils/PhutilCowsay.php
+++ b/src/utils/PhutilCowsay.php
@@ -41,34 +41,29 @@ final class PhutilCowsay extends Phobject {
     $template = $this->template;
 
     // Real ".cow" files are Perl scripts which define a variable called
-    // "$the_cow". We aren't going to interpret Perl, so strip all this stuff
-    // (and any comments in the file) away.
-    $template = phutil_split_lines($template, true);
-    $keep = array();
-    $is_perl_cowfile = false;
-    foreach ($template as $key => $line) {
-      if (preg_match('/^#/', $line)) {
-        continue;
-      }
-      if (preg_match('/^\s*\\$the_cow/', $line)) {
-        $is_perl_cowfile = true;
-        continue;
-      }
-      if (preg_match('/^\s*EOC\s*$/', $line)) {
-        continue;
-      }
-      $keep[] = $line;
-    }
-    $template = implode('', $keep);
+    // "$the_cow". We aren't going to interpret Perl, so just get everything
+    // between the EOC (End Of Cow) tokens. The initial EOC might be in
+    // quotes, and might have a semicolon.
+    // We apply regexp modifiers
+    // * 's' to make . match newlines within the EOC ... EOC block
+    // * 'm' so we can use ^ to match start of line within the multiline string
+    $matches = null;
+    if (
+      preg_match('/\$the_cow/', $template) &&
+      preg_match('/EOC[\'"]?;?.*?^(.*?)^EOC/sm', $template, $matches)
+    ) {
+      $template = $matches[1];
 
-    // Original .cow files are perl scripts which contain escaped sequences.
-    // We attempt to unescape here by replacing any character preceded by a
-    // backslash/escape with just that character.
-    if ($is_perl_cowfile) {
+      // Original .cow files are perl scripts which contain escaped sequences.
+      // We attempt to unescape here by replacing any character preceded by a
+      // backslash/escape with just that character.
       $template = preg_replace(
         '/\\\\(.)/',
         '$1',
         $template);
+    } else {
+      // Text template. Just strip away comments.
+      $template = preg_replace('/^#.*$/', '', $template);
     }
 
     $token_patterns = array(
diff --git a/src/utils/__tests__/cowsay/cube_perl.test b/src/utils/__tests__/cowsay/cube_perl.test
index d9582d2b..a950e588 100644
--- a/src/utils/__tests__/cowsay/cube_perl.test
+++ b/src/utils/__tests__/cowsay/cube_perl.test
@@ -1,5 +1,5 @@
 # test case for original perl-script cowfile
-$the_cow =
+$the_cow = <<EOC
  $thoughts
   $thoughts
     /---\\__/---\\\\
@@ -9,6 +9,7 @@ $the_cow =
      / ..--.. \\\\
     | \\ .... / ||
     \\---/--\\---//
+EOC
 ~~~~~~~~~~
 {
   "text": "I made a friend!",
diff --git a/src/utils/__tests__/cowsay/small.expect b/src/utils/__tests__/cowsay/small.expect
new file mode 100644
index 00000000..9d112578
--- /dev/null
+++ b/src/utils/__tests__/cowsay/small.expect
@@ -0,0 +1,7 @@
+ __________________
+< How are my eyes? >
+ ------------------
+       \   ,__,
+        \  (oo)____
+           (__)    )\
+              ||--|| *
diff --git a/src/utils/__tests__/cowsay/small.test b/src/utils/__tests__/cowsay/small.test
new file mode 100644
index 00000000..bc3e08e4
--- /dev/null
+++ b/src/utils/__tests__/cowsay/small.test
@@ -0,0 +1,12 @@
+$eyes = ".." unless ($eyes);
+$the_cow = <<EOC;
+       $thoughts   ,__,
+        $thoughts  ($eyes)____
+           (__)    )\\
+            $tongue||--|| *
+EOC
+~~~~~~~~~~
+{
+  "text": "How are my eyes?",
+  "eyes": "oo"
+}