Remove backports, alert for insensitive file systems, and bundled file entity for directories (#213)

* Remove backport.
* Detect build breakage on case insensitive file systems
* Do not generate individual entity files for directories
* Setting documentURI cause fails on Windows

Author: alfsb

configure.php

@@ -827,10 +827,7 @@ function xml_configure()
 checking('whether to save an invalid .manual.xml');
 checkvalue($ac['FORCE_DOM_SAVE']);
 
-
-$dom = new DOMDocument();
-
-function dom_load( DOMDocument $dom , string $filename ) : bool
+function dom_load( DOMDocument $dom , string $filename , string $baseURI = "" ) : bool
 {
     $filename = realpath( $filename );
     $options = LIBXML_NOENT | LIBXML_COMPACT | LIBXML_BIGLINES | LIBXML_PARSEHUGE;
@@ -849,6 +846,7 @@ function dom_saveload( DOMDocument $dom , string $filename = "" ) : string
 }
 
 echo "Loading and parsing {$ac["INPUT_FILENAME"]}... ";
+$dom = new DOMDocument();
 
 if ( dom_load( $dom , "{$ac['srcdir']}/{$ac["INPUT_FILENAME"]}" ) )
 {
@@ -862,7 +860,6 @@ function dom_saveload( DOMDocument $dom , string $filename = "" ) : string
     errors_are_bad(1);
 }
 
-
 echo "Running XInclude/XPointer... ";
 
 $total  = xinclude_run_byid( $dom );

scripts/file-entities.php

@@ -20,7 +20,7 @@
 This script creates various "file entities", that is, DTD entities that
 point to files and file listings, named and composed of:
 
-- dir.dir.file : pulls in a dir/dir/file.xml file
+- dir.dir.file : pulls in a dir/dir/file.xml
 - dir.dif.entities.dir : pulls in all files of dir/dir/dir/*.xml
 
 In the original file-entities.php.in, the files are created at:
@@ -33,24 +33,8 @@
 - doc-base/temp/file-entites.ent
 - doc-base/temp/file-entites.dir.dir.ent
 
-# TODO
-
-1. Leave this running in new idempotent mode for a few months, before
-erasing the const BACKPORT, that exists only to replicate the old
-style build.
-
-2. Istead of creating ~thousand doc-base/temp/file-entites.*.ent files,
-output an solid XML grouped file (per github.com/php/doc-base/pull/183)
-so it would be possible to detect accidental overwriting of structural
-entities, the "list of entities" moved to/as normal entity text. PS: This
-will NOT work, with libxml recusing to load .manuxal.xml.in because of an
-"Detected an entity reference loop", that does not exist. Sigh. PPS: May
-be possible with LIBXML_PARSEHUGE after all.
-
 */
 
-const BACKPORT = false; // TODO remove, see above.
-
 // Setup
 
 ini_set( 'display_errors' , 1 );
@@ -86,15 +70,15 @@
 
 echo "Creating file-entities.ent... ";
 
-$entities = []; // See pushEntity()
+$entities = [];
+$mixedCase = [];
 
 generate_file_entities( $root , "en" );
 generate_list_entities( $root , "en" );
 
 if ( $lang != "" )
     generate_file_entities( $root , $lang );
 
-// TODO BACKPORT: Fixed relative path, move this directly into manual.xml.in
 pushEntity( "global.function-index", path: realpain( __DIR__ . "/.." ) . "/funcindex.xml" );
 
 if ( ! $chmonly )
@@ -113,13 +97,7 @@
 
 fputs( $file , "<!-- DON'T TOUCH - AUTOGENERATED BY file-entities.php -->\n\n" );
 
-if ( BACKPORT )
-    fputs( $file , "\n" );
-
-if ( BACKPORT )
-    asort( $entities );
-else
-    ksort( $entities );
+ksort( $entities );
 
 foreach ( $entities as $ent )
     writeEntity( $file , $ent );
@@ -138,6 +116,7 @@ function __construct( public string $name , public string $text , public string
 function pushEntity( string $name , string $text = '' , string $path = '' )
 {
     global $entities;
+    global $mixedCase;
 
     $name = str_replace( '_' , '-' , $name );
     $path = str_replace( '\\' , '/' , $path );
@@ -147,8 +126,25 @@ function pushEntity( string $name , string $text = '' , string $path = '' )
     if ( ( $text == "" && $path == "" ) || ( $text != "" && $path != "" ) )
     {
         echo "Something went wrong on file-entities.php.\n";
-        exit(-1);
+        exit( 1 );
     }
+
+    $lname = strtolower( $name );
+    if ( isset( $mixedCase[ $lname ] ) && $mixedCase[ $lname ] != $name )
+    {
+        echo "\n\n";
+        echo "BROKEN BUILD on case insensitive file systems!\n";
+        echo "Detected distinct file entities only by case:\n";
+        echo " - {$mixedCase[ $lname ]}\n";
+        echo " - $name \n";
+        echo "This will PERMANENTLY BRICK manual build on Windows machines!\n";
+        echo "Avoid committing this on repository, and if it's already committed,\n";
+        echo "revert and send a heads up on mailinst how to fix the issue.\n\n";
+        echo "See https://github.com/php/doc-en/pull/4330#issuecomment-2557306828";
+        echo "\n\n";
+        exit( 1 );
+    }
+    $mixedCase[ $lname ] = $name;
 }
 
 function generate_file_entities( string $root , string $lang )
@@ -158,7 +154,7 @@ function generate_file_entities( string $root , string $lang )
     if ( $test === false || is_dir( $path ) == false )
     {
         echo "Language directory not found: $path\n.";
-        exit(-1);
+        exit( 1 );
     }
     $path = $test;
 
@@ -210,13 +206,10 @@ function generate_list_entities( string $root , string $lang )
     if ( $test === false || is_dir( $path ) == false )
     {
         echo "Language directory not found: $path\n.";
-        exit(-1);
+        exit( 1 );
     }
     $path = $test;
 
-    if ( BACKPORT ) // Spurious file generated outside reference/
-        pushEntity( "language.predefined.entities.weakreference", path: "$root/$lang/language/predefined/entities.weakreference.xml" );
-
     $dirs = [ "reference" ];
     list_entities_recurse( $path , $dirs );
 }
@@ -235,8 +228,6 @@ function list_entities_recurse( string $root , array $dirs )
             continue;
         if ( $file[0] == "." )
             continue;
-        if ( BACKPORT && str_starts_with( $file , "entities.") )
-            continue;
 
         $path = "$dir/$file";
 
@@ -265,18 +256,15 @@ function list_entities_recurse( string $root , array $dirs )
     $text = implode( "\n" , $list );
 
     if ( $text != "" )
-    {
-        // pushEntity( $name , text: $text ); // See TODO item 2 // LIBXML_PARSEHUGE
+        pushEntity( $name , text: $text );
 
-        if ( BACKPORT )
-            $path = "$dir/../entities.$last.xml";
-        else
-            $path = __DIR__ . "/../temp/file-entities." . implode( '.' , $dirs ) . ".ent";
-
-        file_put_contents( $path , $text );
-        $path = realpain( $path );
-        pushEntity( $name , path: $path );
-    }
+//  Old style, pre LIBXML_PARSEHUGE, "directory" entity as external file
+//
+//        $path = __DIR__ . "/../temp/file-entities." . implode( '.' , $dirs ) . ".ent";
+//        file_put_contents( $path , $text );
+//        $path = realpain( $path );
+//        pushEntity( $name , path: $path );
+//
 
     foreach( $subdirs as $subdir )
     {
@@ -292,20 +280,10 @@ function writeEntity( $file , Entity $ent )
     $text = $ent->text;
     $path = $ent->path;
 
-    if ( BACKPORT )
-    {
-        if ( $path == "" )
-            $line = sprintf("<!ENTITY %-40s        ''>\n" , $name ); // was on original, possibly unused
-        else
-            $line = sprintf("<!ENTITY %-40s SYSTEM 'file:///%s'>\n" , $name , $path );
-    }
+    if ( $path == "" )
+        $line = "<!ENTITY $name '$text'>\n";
     else
-    {
-        if ( $path == "" )
-            $line = "<!ENTITY $name '$text'>\n";
-        else
-            $line = "<!ENTITY $name SYSTEM '$path'>\n";
-    }
+        $line = "<!ENTITY $name SYSTEM '$path'>\n";
 
     fwrite( $file , $line );
 }