Enhance Gitea support by improving base64 decoding and UTF-8 encoding handling

includes/api-client.php

@@ -380,13 +380,15 @@ function oribi_sync_fetch_file( string $api_base, string $branch, string $file_p
 
     $body = wp_remote_retrieve_body( $response );
     
-    // For Gitea, the /contents/ endpoint returns base64-encoded content in JSON
+    // For Gitea, the /contents/ endpoint returns base64-encoded content in JSON.
+    // Gitea (like GitHub) inserts \n every 60 chars in the base64 — strip them before decoding.
     if ( $provider === 'gitea' ) {
         $decoded = json_decode( $body, true );
         if ( isset( $decoded['content'] ) && is_string( $decoded['content'] ) ) {
-            $body = base64_decode( $decoded['content'], true );
+            $clean = str_replace( [ "\r", "\n", " " ], '', $decoded['content'] );
+            $body  = base64_decode( $clean, true );
             if ( $body === false ) {
-                return new WP_Error( 'oribi_sync_decode_error', 'Failed to decode base64 content from Gitea' );
+                return new WP_Error( 'oribi_sync_decode_error', 'Failed to decode base64 content from Gitea.' );
             }
         }
     }

includes/post-sync.php

@@ -858,8 +858,9 @@ function oribi_sync_trash_removed_posts( array $current_slugs ): array {
 function oribi_sync_generate_post_markdown( WP_Post $post ): string {
     $fm = "---\n";
 
-    // Title (escape newlines)
+    // Title — decode HTML entities (WP stores & etc. in DB) so the YAML
-    $fm .= 'title: ' . str_replace( [ "\r", "\n" ], ' ', $post->post_title ) . "\n";
+    // file contains the literal character. On pull-back WP re-encodes correctly.
+    $fm .= 'title: ' . str_replace( [ "\r", "\n" ], ' ', html_entity_decode( $post->post_title, ENT_QUOTES | ENT_HTML5, 'UTF-8' ) ) . "\n";
     $fm .= 'slug: '   . $post->post_name    . "\n";
     $fm .= 'status: ' . $post->post_status  . "\n";
 
@@ -874,27 +875,27 @@ function oribi_sync_generate_post_markdown( WP_Post $post ): string {
         $fm .= 'author: ' . $author->user_login . "\n";
     }
 
-    // Categories
+    // Categories — decode HTML entities stored by WP
     $cats = get_the_category( $post->ID );
     if ( ! empty( $cats ) ) {
         $fm .= "categories:\n";
         foreach ( $cats as $cat ) {
-            $fm .= '  - ' . $cat->name . "\n";
+            $fm .= '  - ' . html_entity_decode( $cat->name, ENT_QUOTES | ENT_HTML5, 'UTF-8' ) . "\n";
         }
     }
 
-    // Tags
+    // Tags — decode HTML entities stored by WP
     $post_tags = get_the_tags( $post->ID );
     if ( ! empty( $post_tags ) ) {
         $fm .= "tags:\n";
         foreach ( $post_tags as $tag ) {
-            $fm .= '  - ' . $tag->name . "\n";
+            $fm .= '  - ' . html_entity_decode( $tag->name, ENT_QUOTES | ENT_HTML5, 'UTF-8' ) . "\n";
         }
     }
 
-    // Excerpt
+    // Excerpt — decode HTML entities stored by WP
     if ( ! empty( $post->post_excerpt ) ) {
-        $fm .= 'excerpt: ' . str_replace( [ "\r", "\n" ], ' ', $post->post_excerpt ) . "\n";
+        $fm .= 'excerpt: ' . str_replace( [ "\r", "\n" ], ' ', html_entity_decode( $post->post_excerpt, ENT_QUOTES | ENT_HTML5, 'UTF-8' ) ) . "\n";
     }
 
     // Featured image (absolute URL so it round-trips cleanly)

includes/push-client.php

@@ -104,7 +104,7 @@ function oribi_sync_api_request( string $method, string $url, array $body, strin
         'method'  => $method,
         'timeout' => 30,
         'headers' => $headers,
-        'body'    => wp_json_encode( $body, JSON_UNESCAPED_UNICODE ),
+        'body'    => wp_json_encode( $body, JSON_UNESCAPED_UNICODE | JSON_UNESCAPED_SLASHES ),
     ];
 
     $response = wp_remote_request( $url, $args );
@@ -152,9 +152,15 @@ function oribi_sync_gitea_get_file_meta( string $api_base, string $branch, strin
         return $result;
     }
 
+    // Gitea inserts \n every 60 chars in base64 — strip before decoding.
+    $raw_b64 = $result['content'] ?? '';
+    $content = ! empty( $raw_b64 )
+        ? base64_decode( str_replace( [ "\r", "\n", " " ], '', $raw_b64 ), true )
+        : '';
+
     return [
         'sha'     => $result['sha'] ?? '',
-        'content' => isset( $result['content'] ) ? base64_decode( $result['content'] ) : '',
+        'content' => ( $content !== false ) ? $content : '',
     ];
 }
 
@@ -182,6 +188,11 @@ function oribi_sync_gitea_put_file(
     ?string $sha = null,
     string $message = ''
 ) {
+    // Validate and fix UTF-8 encoding before base64-encoding
+    if ( ! mb_check_encoding( $content, 'UTF-8' ) ) {
+        $content = mb_convert_encoding( $content, 'UTF-8', 'UTF-8, ISO-8859-1, Windows-1252' );
+    }
+    
     $encoded_path = implode( '/', array_map( 'rawurlencode', explode( '/', $filepath ) ) );
     $url = $api_base . '/contents/' . $encoded_path;