fix(parsers): implement parseDockerVolumeString function to handle various Docker volume formats and modes, including environment variables and Windows paths. Add unit tests for comprehensive coverage.

bootstrap/helpers/parsers.php

@@ -16,6 +16,209 @@ use Spatie\Url\Url;
 use Symfony\Component\Yaml\Yaml;
 use Visus\Cuid2\Cuid2;
 
+function parseDockerVolumeString(string $volumeString): array
+{
+    $volumeString = trim($volumeString);
+    $source = null;
+    $target = null;
+    $mode = null;
+
+    // First, check if the source contains an environment variable with default value
+    // This needs to be done before counting colons because ${VAR:-value} contains a colon
+    $envVarPattern = '/^\$\{[^}]+:-[^}]*\}/';
+    $hasEnvVarWithDefault = false;
+    $envVarEndPos = 0;
+
+    if (preg_match($envVarPattern, $volumeString, $matches)) {
+        $hasEnvVarWithDefault = true;
+        $envVarEndPos = strlen($matches[0]);
+    }
+
+    // Count colons, but exclude those inside environment variables
+    $effectiveVolumeString = $volumeString;
+    if ($hasEnvVarWithDefault) {
+        // Temporarily replace the env var to count colons correctly
+        $effectiveVolumeString = substr($volumeString, $envVarEndPos);
+        $colonCount = substr_count($effectiveVolumeString, ':');
+    } else {
+        $colonCount = substr_count($volumeString, ':');
+    }
+
+    if ($colonCount === 0) {
+        // Named volume without target (unusual but valid)
+        // Example: "myvolume"
+        $source = $volumeString;
+        $target = $volumeString;
+    } elseif ($colonCount === 1) {
+        // Simple volume mapping
+        // Examples: "gitea:/data" or "./data:/app/data" or "${VAR:-default}:/data"
+        if ($hasEnvVarWithDefault) {
+            $source = substr($volumeString, 0, $envVarEndPos);
+            $remaining = substr($volumeString, $envVarEndPos);
+            if (strlen($remaining) > 0 && $remaining[0] === ':') {
+                $target = substr($remaining, 1);
+            } else {
+                $target = $remaining;
+            }
+        } else {
+            $parts = explode(':', $volumeString);
+            $source = $parts[0];
+            $target = $parts[1];
+        }
+    } elseif ($colonCount === 2) {
+        // Volume with mode OR Windows path OR env var with mode
+        // Handle env var with mode first
+        if ($hasEnvVarWithDefault) {
+            // ${VAR:-default}:/path:mode
+            $source = substr($volumeString, 0, $envVarEndPos);
+            $remaining = substr($volumeString, $envVarEndPos);
+
+            if (strlen($remaining) > 0 && $remaining[0] === ':') {
+                $remaining = substr($remaining, 1);
+                $lastColon = strrpos($remaining, ':');
+
+                if ($lastColon !== false) {
+                    $possibleMode = substr($remaining, $lastColon + 1);
+                    $validModes = ['ro', 'rw', 'z', 'Z', 'rslave', 'rprivate', 'rshared', 'slave', 'private', 'shared', 'cached', 'delegated', 'consistent'];
+
+                    if (in_array($possibleMode, $validModes)) {
+                        $mode = $possibleMode;
+                        $target = substr($remaining, 0, $lastColon);
+                    } else {
+                        $target = $remaining;
+                    }
+                } else {
+                    $target = $remaining;
+                }
+            }
+        } elseif (preg_match('/^[A-Za-z]:/', $volumeString)) {
+            // Windows path as source (C:/, D:/, etc.)
+            // Find the second colon which is the real separator
+            $secondColon = strpos($volumeString, ':', 2);
+            if ($secondColon !== false) {
+                $source = substr($volumeString, 0, $secondColon);
+                $target = substr($volumeString, $secondColon + 1);
+            } else {
+                // Malformed, treat as is
+                $source = $volumeString;
+                $target = $volumeString;
+            }
+        } else {
+            // Not a Windows path, check for mode
+            $lastColon = strrpos($volumeString, ':');
+            $possibleMode = substr($volumeString, $lastColon + 1);
+
+            // Check if the last part is a valid Docker volume mode
+            $validModes = ['ro', 'rw', 'z', 'Z', 'rslave', 'rprivate', 'rshared', 'slave', 'private', 'shared', 'cached', 'delegated', 'consistent'];
+
+            if (in_array($possibleMode, $validModes)) {
+                // It's a mode
+                // Examples: "gitea:/data:ro" or "./data:/app/data:rw"
+                $mode = $possibleMode;
+                $volumeWithoutMode = substr($volumeString, 0, $lastColon);
+                $colonPos = strpos($volumeWithoutMode, ':');
+
+                if ($colonPos !== false) {
+                    $source = substr($volumeWithoutMode, 0, $colonPos);
+                    $target = substr($volumeWithoutMode, $colonPos + 1);
+                } else {
+                    // Shouldn't happen for valid volume strings
+                    $source = $volumeWithoutMode;
+                    $target = $volumeWithoutMode;
+                }
+            } else {
+                // The last colon is part of the path
+                // For now, treat the first occurrence of : as the separator
+                $firstColon = strpos($volumeString, ':');
+                $source = substr($volumeString, 0, $firstColon);
+                $target = substr($volumeString, $firstColon + 1);
+            }
+        }
+    } else {
+        // More than 2 colons - likely Windows paths or complex cases
+        // Use a heuristic: find the most likely separator colon
+        // Look for patterns like "C:" at the beginning (Windows drive)
+        if (preg_match('/^[A-Za-z]:/', $volumeString)) {
+            // Windows path as source
+            // Find the next colon after the drive letter
+            $secondColon = strpos($volumeString, ':', 2);
+            if ($secondColon !== false) {
+                $source = substr($volumeString, 0, $secondColon);
+                $remaining = substr($volumeString, $secondColon + 1);
+
+                // Check if there's a mode at the end
+                $lastColon = strrpos($remaining, ':');
+                if ($lastColon !== false) {
+                    $possibleMode = substr($remaining, $lastColon + 1);
+                    $validModes = ['ro', 'rw', 'z', 'Z', 'rslave', 'rprivate', 'rshared', 'slave', 'private', 'shared', 'cached', 'delegated', 'consistent'];
+
+                    if (in_array($possibleMode, $validModes)) {
+                        $mode = $possibleMode;
+                        $target = substr($remaining, 0, $lastColon);
+                    } else {
+                        $target = $remaining;
+                    }
+                } else {
+                    $target = $remaining;
+                }
+            } else {
+                // Malformed, treat as is
+                $source = $volumeString;
+                $target = $volumeString;
+            }
+        } else {
+            // Try to parse normally, treating first : as separator
+            $firstColon = strpos($volumeString, ':');
+            $source = substr($volumeString, 0, $firstColon);
+            $remaining = substr($volumeString, $firstColon + 1);
+
+            // Check for mode at the end
+            $lastColon = strrpos($remaining, ':');
+            if ($lastColon !== false) {
+                $possibleMode = substr($remaining, $lastColon + 1);
+                $validModes = ['ro', 'rw', 'z', 'Z', 'rslave', 'rprivate', 'rshared', 'slave', 'private', 'shared', 'cached', 'delegated', 'consistent'];
+
+                if (in_array($possibleMode, $validModes)) {
+                    $mode = $possibleMode;
+                    $target = substr($remaining, 0, $lastColon);
+                } else {
+                    $target = $remaining;
+                }
+            } else {
+                $target = $remaining;
+            }
+        }
+    }
+
+    // Handle environment variable expansion in source
+    // Example: ${VOLUME_DB_PATH:-db} should extract default value if present
+    if ($source && preg_match('/^\$\{([^}]+)\}$/', $source, $matches)) {
+        $varContent = $matches[1];
+
+        // Check if there's a default value with :-
+        if (strpos($varContent, ':-') !== false) {
+            $parts = explode(':-', $varContent, 2);
+            $varName = $parts[0];
+            $defaultValue = isset($parts[1]) ? $parts[1] : '';
+
+            // If there's a non-empty default value, use it for source
+            if ($defaultValue !== '') {
+                $source = $defaultValue;
+            } else {
+                // Empty default value, keep the variable reference for env resolution
+                $source = '${'.$varName.'}';
+            }
+        }
+        // Otherwise keep the variable as-is for later expansion (no default value)
+    }
+
+    return [
+        'source' => $source !== null ? str($source) : null,
+        'target' => $target !== null ? str($target) : null,
+        'mode' => $mode !== null ? str($mode) : null,
+    ];
+}
+
 function applicationParser(Application $resource, int $pull_request_id = 0, ?int $preview_id = null): Collection
 {
     $uuid = data_get($resource, 'uuid');
@@ -304,8 +507,10 @@ function applicationParser(Application $resource, int $pull_request_id = 0, ?int
                 $content = null;
                 $isDirectory = false;
                 if (is_string($volume)) {
-                    $source = str($volume)->beforeLast(':');
-                    $target = str($volume)->afterLast(':');
+                    $parsed = parseDockerVolumeString($volume);
+                    $source = $parsed['source'];
+                    $target = $parsed['target'];
+                    // Mode is available in $parsed['mode'] if needed
                     $foundConfig = $fileStorages->whereMountPath($target)->first();
                     if (sourceIsLocal($source)) {
                         $type = str('bind');
@@ -399,8 +604,9 @@ function applicationParser(Application $resource, int $pull_request_id = 0, ?int
                         $name = "{$name}-pr-$pullRequestId";
                     }
                     if (is_string($volume)) {
-                        $source = str($volume)->beforeLast(':');
-                        $target = str($volume)->afterLast(':');
+                        $parsed = parseDockerVolumeString($volume);
+                        $source = $parsed['source'];
+                        $target = $parsed['target'];
                         $source = $name;
                         $volume = "$source:$target";
                     } elseif (is_array($volume)) {
@@ -1312,8 +1518,10 @@ function serviceParser(Service $resource): Collection
                 $content = null;
                 $isDirectory = false;
                 if (is_string($volume)) {
-                    $source = str($volume)->beforeLast(':');
-                    $target = str($volume)->afterLast(':');
+                    $parsed = parseDockerVolumeString($volume);
+                    $source = $parsed['source'];
+                    $target = $parsed['target'];
+                    // Mode is available in $parsed['mode'] if needed
                     $foundConfig = $fileStorages->whereMountPath($target)->first();
                     if (sourceIsLocal($source)) {
                         $type = str('bind');
@@ -1401,8 +1609,9 @@ function serviceParser(Service $resource): Collection
                     $name = "{$uuid}_{$slugWithoutUuid}";
 
                     if (is_string($volume)) {
-                        $source = str($volume)->beforeLast(':');
-                        $target = str($volume)->afterLast(':');
+                        $parsed = parseDockerVolumeString($volume);
+                        $source = $parsed['source'];
+                        $target = $parsed['target'];
                         $source = $name;
                         $volume = "$source:$target";
                     } elseif (is_array($volume)) {

tests/Unit/DockerImageParserTest.php

@@ -1,6 +1,6 @@
 <?php
 
-namespace Tests\Unit\Services;
+namespace Tests\Unit;
 
 use App\Services\DockerImageParser;
 use PHPUnit\Framework\Attributes\Test;

tests/Unit/ParseDockerVolumeStringTest.php

@@ -0,0 +1,193 @@
+<?php
+
+test('parses simple volume mappings', function () {
+    // Simple named volume
+    $result = parseDockerVolumeString('gitea:/data');
+    expect($result['source']->value())->toBe('gitea');
+    expect($result['target']->value())->toBe('/data');
+    expect($result['mode'])->toBeNull();
+
+    // Simple bind mount
+    $result = parseDockerVolumeString('./data:/app/data');
+    expect($result['source']->value())->toBe('./data');
+    expect($result['target']->value())->toBe('/app/data');
+    expect($result['mode'])->toBeNull();
+
+    // Absolute path bind mount
+    $result = parseDockerVolumeString('/var/lib/data:/data');
+    expect($result['source']->value())->toBe('/var/lib/data');
+    expect($result['target']->value())->toBe('/data');
+    expect($result['mode'])->toBeNull();
+});
+
+test('parses volumes with read-only mode', function () {
+    // Named volume with ro mode
+    $result = parseDockerVolumeString('gitea-localtime:/etc/localtime:ro');
+    expect($result['source']->value())->toBe('gitea-localtime');
+    expect($result['target']->value())->toBe('/etc/localtime');
+    expect($result['mode']->value())->toBe('ro');
+
+    // Bind mount with ro mode
+    $result = parseDockerVolumeString('/etc/localtime:/etc/localtime:ro');
+    expect($result['source']->value())->toBe('/etc/localtime');
+    expect($result['target']->value())->toBe('/etc/localtime');
+    expect($result['mode']->value())->toBe('ro');
+});
+
+test('parses volumes with other modes', function () {
+    // Read-write mode
+    $result = parseDockerVolumeString('data:/var/data:rw');
+    expect($result['source']->value())->toBe('data');
+    expect($result['target']->value())->toBe('/var/data');
+    expect($result['mode']->value())->toBe('rw');
+
+    // Z mode (SELinux)
+    $result = parseDockerVolumeString('config:/etc/config:z');
+    expect($result['source']->value())->toBe('config');
+    expect($result['target']->value())->toBe('/etc/config');
+    expect($result['mode']->value())->toBe('z');
+
+    // Cached mode (macOS)
+    $result = parseDockerVolumeString('./src:/app/src:cached');
+    expect($result['source']->value())->toBe('./src');
+    expect($result['target']->value())->toBe('/app/src');
+    expect($result['mode']->value())->toBe('cached');
+
+    // Delegated mode (macOS)
+    $result = parseDockerVolumeString('./node_modules:/app/node_modules:delegated');
+    expect($result['source']->value())->toBe('./node_modules');
+    expect($result['target']->value())->toBe('/app/node_modules');
+    expect($result['mode']->value())->toBe('delegated');
+});
+
+test('parses volumes with environment variables', function () {
+    // Variable with default value
+    $result = parseDockerVolumeString('${VOLUME_DB_PATH:-db}:/data/db');
+    expect($result['source']->value())->toBe('db');
+    expect($result['target']->value())->toBe('/data/db');
+    expect($result['mode'])->toBeNull();
+
+    // Variable without default value
+    $result = parseDockerVolumeString('${VOLUME_PATH}:/data');
+    expect($result['source']->value())->toBe('${VOLUME_PATH}');
+    expect($result['target']->value())->toBe('/data');
+    expect($result['mode'])->toBeNull();
+
+    // Variable with empty default - keeps variable reference for env resolution
+    $result = parseDockerVolumeString('${VOLUME_PATH:-}:/data');
+    expect($result['source']->value())->toBe('${VOLUME_PATH}');
+    expect($result['target']->value())->toBe('/data');
+    expect($result['mode'])->toBeNull();
+
+    // Variable with mode
+    $result = parseDockerVolumeString('${DATA_PATH:-./data}:/app/data:ro');
+    expect($result['source']->value())->toBe('./data');
+    expect($result['target']->value())->toBe('/app/data');
+    expect($result['mode']->value())->toBe('ro');
+});
+
+test('parses Windows paths', function () {
+    // Windows absolute path
+    $result = parseDockerVolumeString('C:/Users/data:/data');
+    expect($result['source']->value())->toBe('C:/Users/data');
+    expect($result['target']->value())->toBe('/data');
+    expect($result['mode'])->toBeNull();
+
+    // Windows path with mode
+    $result = parseDockerVolumeString('D:/projects/app:/app:rw');
+    expect($result['source']->value())->toBe('D:/projects/app');
+    expect($result['target']->value())->toBe('/app');
+    expect($result['mode']->value())->toBe('rw');
+
+    // Windows path with spaces (should be quoted in real use)
+    $result = parseDockerVolumeString('C:/Program Files/data:/data');
+    expect($result['source']->value())->toBe('C:/Program Files/data');
+    expect($result['target']->value())->toBe('/data');
+    expect($result['mode'])->toBeNull();
+});
+
+test('parses edge cases', function () {
+    // Volume name only (unusual but valid)
+    $result = parseDockerVolumeString('myvolume');
+    expect($result['source']->value())->toBe('myvolume');
+    expect($result['target']->value())->toBe('myvolume');
+    expect($result['mode'])->toBeNull();
+
+    // Path with colon in target (not a mode)
+    $result = parseDockerVolumeString('source:/path:8080');
+    expect($result['source']->value())->toBe('source');
+    expect($result['target']->value())->toBe('/path:8080');
+    expect($result['mode'])->toBeNull();
+
+    // Multiple colons in path (not Windows)
+    $result = parseDockerVolumeString('data:/var/lib/docker:data:backup');
+    expect($result['source']->value())->toBe('data');
+    expect($result['target']->value())->toBe('/var/lib/docker:data:backup');
+    expect($result['mode'])->toBeNull();
+});
+
+test('parses tmpfs and other special cases', function () {
+    // Docker socket binding
+    $result = parseDockerVolumeString('/var/run/docker.sock:/var/run/docker.sock');
+    expect($result['source']->value())->toBe('/var/run/docker.sock');
+    expect($result['target']->value())->toBe('/var/run/docker.sock');
+    expect($result['mode'])->toBeNull();
+
+    // Docker socket with mode
+    $result = parseDockerVolumeString('/var/run/docker.sock:/var/run/docker.sock:ro');
+    expect($result['source']->value())->toBe('/var/run/docker.sock');
+    expect($result['target']->value())->toBe('/var/run/docker.sock');
+    expect($result['mode']->value())->toBe('ro');
+
+    // Tmp mount
+    $result = parseDockerVolumeString('/tmp:/tmp');
+    expect($result['source']->value())->toBe('/tmp');
+    expect($result['target']->value())->toBe('/tmp');
+    expect($result['mode'])->toBeNull();
+});
+
+test('handles whitespace correctly', function () {
+    // Leading/trailing whitespace
+    $result = parseDockerVolumeString('  data:/app/data  ');
+    expect($result['source']->value())->toBe('data');
+    expect($result['target']->value())->toBe('/app/data');
+    expect($result['mode'])->toBeNull();
+
+    // Whitespace with mode
+    $result = parseDockerVolumeString('  ./config:/etc/config:ro  ');
+    expect($result['source']->value())->toBe('./config');
+    expect($result['target']->value())->toBe('/etc/config');
+    expect($result['mode']->value())->toBe('ro');
+});
+
+test('parses all valid Docker volume modes', function () {
+    $validModes = ['ro', 'rw', 'z', 'Z', 'rslave', 'rprivate', 'rshared',
+        'slave', 'private', 'shared', 'cached', 'delegated', 'consistent'];
+
+    foreach ($validModes as $mode) {
+        $result = parseDockerVolumeString("volume:/data:$mode");
+        expect($result['source']->value())->toBe('volume');
+        expect($result['target']->value())->toBe('/data');
+        expect($result['mode']->value())->toBe($mode);
+    }
+});
+
+test('parses complex real-world examples', function () {
+    // MongoDB volume with environment variable
+    $result = parseDockerVolumeString('${VOLUME_DB_PATH:-./data/db}:/data/db');
+    expect($result['source']->value())->toBe('./data/db');
+    expect($result['target']->value())->toBe('/data/db');
+    expect($result['mode'])->toBeNull();
+
+    // Config file mount with read-only
+    $result = parseDockerVolumeString('/home/user/app/config.yml:/app/config.yml:ro');
+    expect($result['source']->value())->toBe('/home/user/app/config.yml');
+    expect($result['target']->value())->toBe('/app/config.yml');
+    expect($result['mode']->value())->toBe('ro');
+
+    // Named volume with hyphens and underscores
+    $result = parseDockerVolumeString('my-app_data_v2:/var/lib/app-data');
+    expect($result['source']->value())->toBe('my-app_data_v2');
+    expect($result['target']->value())->toBe('/var/lib/app-data');
+    expect($result['mode'])->toBeNull();
+});