LSlang: improve internal message extrators from PHP files

- add support of multiple lines messages
- detect and ignore commented code block
- improving code tracing

Note: still no support of advanced gettext functions (ngettext,
dcgettext, ...).

src/includes/class/class.LSlang.php

@@ -545,59 +545,185 @@ function _cli_find_and_parse_template_file($dir) {
 function _cli_parse_php_file($file) {
   global $LSlang_cli_logger;
   $LSlang_cli_logger -> debug("Looking for string to translate in '$file' PHP file");
-  $count = 0;
+  $quote = '';
-  $quote='';
+  $res = '';
-  $res='';
+  $comment = null;
-  foreach(file($file) as $line) {
+  $line_break_continue = false;
-    $count++;
+  $inside_keyword = false;
-    $LSlang_cli_logger -> trace("Handle line #$count of '$file' PHP file");
+  $inside_comment = false;
-    $offset=0;
+  $end_comment_pos = false;
-    while ($pos = strpos($line,'__(',$offset)) {
+  $keyword = null;
-      $LSlang_cli_logger -> trace("$file:$count: detect keyword at position #$pos ('$line')");
+  foreach(file($file) as $line_count => $line) {
-      for ($i=$pos+3;$i<strlen($line);$i++) {
+    $line = preg_replace("/\n$/", "", $line);
-        $offset=$i; // Always increase offset to avoid infinity-loop
+    $line_count++;
-        if (empty($quote)) {
+    $LSlang_cli_logger -> trace("Handle line #$line_count of '$file' PHP file : '$line'");
+    $offset = 0;
+    while (
+      $line_break_continue
+      || (!$inside_comment && preg_match('#^([ \t])*(//|/\*)#', substr($line, $offset), $comment))
+      || ($inside_comment && $end_comment_pos = strpos(substr($line, $offset), '*/'))
+      || preg_match('/[^a-zA-Z0-9\'\"]_{1,3}\(/', substr($line, $offset), $keyword, PREG_OFFSET_CAPTURE)
+    ) {
+      if ($line_break_continue) {
+        $LSlang_cli_logger -> trace("$file:$line_count: continue from previous line after a line break");
+        $line_break_continue = false;
+        $offset = 0;
+        $inside_keyword = true;
+      }
+      elseif ($comment) {
+        $pos = $offset + strlen($comment[1]);
+        if ($comment[2] == '//') {
+          $LSlang_cli_logger -> trace(
+            "$file:$line_count: entering comment detected '".$comment[2]."' at position #$pos, ".
+            "ignore the line ('$line', offset=$offset)");
+          break;
+        }
+        $LSlang_cli_logger -> trace(
+          "$file:$line_count: entering comment detected '".$comment[2]."' at position #$pos ".
+          "('$line', offset=$offset)");
+        $inside_comment = true;
+        $offset = $pos + strlen($comment[2]);
+        $comment = null;
+      }
+      elseif ($inside_comment) {
+        if ($end_comment_pos) {
+          $pos = $offset + $end_comment_pos;
+          $LSlang_cli_logger -> trace(
+            "$file:$line_count: end of comment detected at position #$pos, continue ".
+            "('$line', offset=$offset)");
+          $offset = $pos;
+          $end_comment_pos = null;
+          $inside_comment = false;
+          continue;
+        }
+        if ($keyword) {
+          $pos = $offset + $keyword[0][1];
+          $LSlang_cli_logger -> trace(
+            "$file:$line_count: detect keyword '".substr($keyword[0][0], 1).")' at position #$pos,".
+            " but inside comment => ignore it ('$line', offset=$offset)");
+          $offset += strlen($keyword[0][0]) + $keyword[0][1];
+        }
+        else {
+          $LSlang_cli_logger -> fatal(
+            "$file:$line_count: not supported case (inside comment & !keyword), stop");
+          return;
+        }
+      }
+      else if ($keyword) {
+        $pos = $offset + $keyword[0][1];
+        $LSlang_cli_logger -> trace(
+          "$file:$line_count: detect keyword '".substr($keyword[0][0], 1).")' at position #$pos ".
+          "('$line', offset=$offset)");
+        $offset += strlen($keyword[0][0]) + $keyword[0][1];
+        $closed = false;
+        $inside_keyword = true;
+      }
+      else {
+        $LSlang_cli_logger -> fatal(
+          "$file:$line_count: not supported case, stop");
+        return;
+      }
+      $ignore = false;
+      $concatenation_need = false;
+      for ($i=$offset; $i<strlen($line); $i++) {
+        $offset = $i; // Always increase offset to avoid infinity-loop
+        if ($inside_comment) {
+          if (substr($line, $i, 2) == '*/') {
+            $LSlang_cli_logger -> trace(
+              "$file:$line_count: End of comment detected at position #$i ('$line')");
+            $inside_comment = false;
+          }
+          $i++;
+        }
+        elseif (empty($quote)) {
           // Quote char not detected : try to detect it
           if ($line[$i]=='\\' || $line[$i]==" " || $line[$i]=="\t") {
             // Space or escape char : pass
             $i++;
           }
-          elseif ($line[$i]=='"' || $line[$i]=="'") {
+          elseif (in_array($line[$i], array('"', "'")) && !$concatenation_need) {
             // Quote detected
+            $LSlang_cli_logger -> trace(
+              "$file:$line_count: Quote (".$line[$i].") detected at position #$i ('$line')");
             $quote=$line[$i];
           }
-          elseif ($line[$i]=='$' || $line[$i]==')') {
+          elseif ($line[$i]=='$') {
             // Variable translation not possible or end function call detected
+            $LSlang_cli_logger -> trace(
+              "$file:$line_count: Variable translation detected at position #$i, ignore it ".
+              "('$line')");
+            $ignore = true;
+            break;
+          }
+          elseif ($line[$i]==')' && $inside_keyword) {
+            // End function call detected
+            $LSlang_cli_logger -> trace(
+              "$file:$line_count: End function call detected at position #$i ('$line')");
+            $inside_keyword = false;
+            break;
+          }
+          elseif ($line[$i] == '.' && $concatenation_need) {
+            // Concatenation char detected
+            $LSlang_cli_logger -> trace(
+              "$file:$line_count: Concatenation char detected at position #$i ('$line')");
+            $concatenation_need = false;
+          }
+          elseif (preg_match('/[A-Za-z]/', $line[$i])) {
+            // Constant or function call detected
+            $LSlang_cli_logger -> trace(
+              "$file:$line_count: Constant or function call detected at position #$i, ignore it ('$line')");
+            $ignore = true;
             break;
           }
           else {
             // Unknown case : continue
+            $LSlang_cli_logger -> trace(
+              "$file:$line_count: Unknown case before quote at position #$i ('".$line[$i]."'), ".
+              "ignore it and continue ('$line')");
             $i++;
           }
         }
         elseif ($quote) {
           // Quote char already detected : try to detect end quote char
-          if ($line[$i]=='\\') {
+          if ($line[$i] == '\\') {
-            // Escape char detected : pass this char and the following one
+            // Escape char detected : keep it and the following one
-            $res.=$line[$i];
+            $LSlang_cli_logger -> trace(
+              "$file:$line_count: Escape char detected inside quote at position #$i, keep it and ".
+              "the following one ('$line')");
+            $res .= $line[$i];
             $i++;
-            $res.=$line[$i];
+            $res .= $line[$i];
           }
-          elseif ($line[$i]==$quote) {
+          elseif ($line[$i] == $quote) {
-            // End quote char detected : reset quote char detection and break detection
+            // End quote char detected : reset quote and set concatenation as need
-            $quote='';
+            $LSlang_cli_logger -> trace(
-            break;
+              "$file:$line_count: End quote char detected at position #$i, reset quote and set ".
+              "concatenation as need ('$line')");
+            $quote = '';
+            $concatenation_need = true;
           }
           else {
             // End quote char not detected : append current char to result
-            $res.=$line[$i];
+            $res .= $line[$i];
           }
         }
       }
       // Include detected string if not empty and quote char was detected and reseted
-      if (!empty($res) && empty($quote)) {
+      if ($inside_comment) {
-        _cli_add_str_to_translate($res, _cli_absolute2relative_path($file).":$count");
+        $LSlang_cli_logger -> trace(
-        $res='';
+          "$file:$line_count: line break detected inside comment ('$line')");
+      }
+      elseif (!empty($res) && empty($quote) && !$inside_keyword) {
+        _cli_add_str_to_translate($res, _cli_absolute2relative_path($file).":$line_count");
+        $res = '';
+      }
+      else if (!$ignore) {
+        $LSlang_cli_logger -> trace(
+          "$file:$line_count: line break detected inside keyword, continue on next line ('$line')");
+        $line_break_continue = true;
+        if ($quote)
+          $res .= "\n";
+        break;
       }
     }
   }
@@ -609,12 +735,24 @@ function _cli_parse_php_file($file) {
  * @param  string $dir  The directory path
  * @return void
  */
-function _cli_find_and_parse_php_file($dir, $filename_regex) {
+function _cli_find_and_parse_php_file($dir, $filename_regex=null, $recursive=false, $ignore_dirs=null) {
+  global $LSlang_cli_logger;
+  $filename_regex = $filename_regex?$filename_regex:'/^(.+)\.php$/';
   if (is_dir($dir)) {
+    $LSlang_cli_logger -> debug("Looking for string to translate in PHP files of directory '$dir'");
     if ($dh = opendir($dir)) {
       while (($file = readdir($dh)) !== false) {
-        if (preg_match($filename_regex, $file)) {
+        if ($file == '.' || $file == '..') continue;
-          _cli_parse_php_file($dir.'/'.$file);
+        if (is_link("$dir/$file")) continue;
+        if (
+          is_dir("$dir/$file")
+          && $recursive
+          && (!is_array($ignore_dirs) || !in_array("$dir/$file", $ignore_dirs))
+        ) {
+          _cli_find_and_parse_php_file("$dir/$file", $filename_regex, true, $ignore_dirs);
+        }
+        elseif (is_file($dir."/".$file) && preg_match($filename_regex, $file)) {
+          _cli_parse_php_file("$dir/$file");
         }
       }
       closedir($dh);
@@ -1018,8 +1156,8 @@ function cli_generate_lang_file($command_args) {
   // Manage includes files
   if (!in_array('includes', $withouts) && (!$only || $only == 'includes')) {
     // Note: Upstream code most only use gettext translation, do not handle it here
-    if ($include_upstream) _cli_find_and_parse_php_file(LS_ROOT_DIR.'/'.LS_INCLUDE_DIR, '/^(.+)\.php$/');
+    if ($include_upstream) _cli_find_and_parse_php_file(LS_ROOT_DIR.'/'.LS_INCLUDE_DIR);
-    _cli_find_and_parse_php_file(LS_ROOT_DIR.'/'.LS_LOCAL_DIR.LS_INCLUDE_DIR, '/^(.+)\.php$/');
+    _cli_find_and_parse_php_file(LS_ROOT_DIR.'/'.LS_LOCAL_DIR.LS_INCLUDE_DIR);
     if ($include_upstream) _cli_find_and_parse_php_file(LS_ROOT_DIR.'/'.LS_CLASS_DIR, '/^class\.(.+)\.php$/');
     _cli_find_and_parse_php_file(LS_ROOT_DIR.'/'.LS_LOCAL_DIR.LS_CLASS_DIR, '/^class\.(.+)\.php$/');
   }