From 6d4d44caf69e06ea51937da9c74e78af0ac99a62 Mon Sep 17 00:00:00 2001
From: Tom Ross <midgey@rockbox.org>
Date: Sat, 17 Oct 2009 05:24:11 +0000
Subject: Add support to genlang to sort strings based on their user. This is
 useful for translatable plugins. Currently, the .lng that is generated is
 hard-coded to only contain strings marked as core. The output of this version
 of genlang should be the same as non-sorting, so we don't need to change the
 version number.

git-svn-id: svn://svn.rockbox.org/rockbox/trunk@23218 a1c6a512-1295-4272-9138-f99709370657
---
 tools/genlang | 121 ++++++++++++++++++++++++++++++++++------------------------
 1 file changed, 71 insertions(+), 50 deletions(-)

diff --git a/tools/genlang b/tools/genlang
index 44dfcc6fdf..c5fb403801 100755
--- a/tools/genlang
+++ b/tools/genlang
@@ -22,9 +22,6 @@ my $LANGUAGE_FLAG_RTL = 0x01;
 # the ability. But, starting now, this ability is no longer provided since I
 # figured it was boring and unnecessary to write support for now since we
 # don't use it anymore.
-
-my %user2num =
-  ('core' => 1);
    
 
 if(!$ARGV[0]) {
@@ -131,6 +128,9 @@ my %source; # id string to source phrase hash
 my %dest; # id string to dest phrase hash
 my %voice; # id string to voice phrase hash
 
+my %users =
+  ('core' => 0);
+
 my $input = $ARGV[0];
 
 my @m;
@@ -214,16 +214,17 @@ if($english) {
     # it before we read the translated file. For -b it isn't necessary, but for
     # -u it is convenient.
 
-    my $idnum=0; # start with a true number
-    my $vidnum=0x8000; # first voice id
+    my @idnum = ((0));       # start with a true number
+    my @vidnum = ((0x8000)); # first voice id
     open(ENG, "<$english") || die "Error: can't open $english";
     my @phrase;
     my $id;
     my $maybeid;
     my $user;
-    my $maybeuser;
     my $withindest;
     my $numphrases = 0;
+    my $numusers = 1; # core is already in the users map
+    
     while(<ENG>) {
 
         # get rid of DOS newlines
@@ -239,14 +240,14 @@ if($english) {
                 # voice-only entries get a difference range
                 if($id =~ /^VOICE_/) {
                     # Assign an ID number to this entry
-                    $idmap{$id}=$vidnum;
-                    $vidnum++;
+                    $idmap[$user]{$id}=$vidnum[$user];
+                    $vidnum[$user]++;
                 }
                 else {
                     # Assign an ID number to this entry
-                    $idmap{$id}=$idnum;
-                    $idnum++;
-   #                 print STDERR "DEST: bumped idnum to $idnum\n";
+                    $idmap[$user]{$id}=$idnum[$user];
+                    $idnum[$user]++;
+   #                 print STDERR "DEST: bumped idnum to $idnum[$user]\n";
                 }
 
                 # this is the end of a phrase, add it to the english hash
@@ -269,7 +270,6 @@ if($english) {
                     # we unconditionally always use all IDs when the "update"
                     # feature is used
                     $id = $maybeid;
-                    $user = $user2num{$maybeuser};
     #                print "DEST: use this id $id\n";
                 }
                 else {
@@ -295,7 +295,11 @@ if($english) {
             $sortorder{$maybeid}=$numphrases++;
         }
         if($_ =~ /^ *user: ([^ \t\n]+)/i) {
-            $maybeuser=$1;
+            $user = $users{$1};
+            if(!(defined $user)) {
+                $user = ++$numusers;
+                $users{$1} = $user;
+            }
         }
     }
     close(ENG);
@@ -378,8 +382,13 @@ sub compare {
     print @show;
 }
 
-my $idcount;        # counter for lang ID numbers
-my $voiceid=0x8000; # counter for voice-only ID numbers
+my @idcount;           # counter for lang ID numbers
+my @voiceid;           # counter for voice-only ID numbers
+
+foreach $i (keys %users) {
+    push @idcount, 0;
+    push @voiceid, 0x8000;
+}
 
 #
 # Now start the scanning of the selected language string
@@ -465,10 +474,22 @@ while(<LANG>) {
                         print STDERR "$input:$line:1: warning: empty dest before line in non-deprecated phrase!\n";
                     }
                     
-                    my $user = $user2num{trim($phrase{'user'})};
-                    if(!$user) {
-                        print STDERR "$input:$line:1: warning: unknown user!\n";
-                        $user = 1;
+                    my $userstr = trim($phrase{'user'});
+                    my $user = $users{$userstr};
+                    if ($userstr eq "") {
+                        print STDERR "$input:$line:1: warning: missing user!\n";
+                        $user = $users{"core"};
+                    }
+                    elsif(!(defined $user)) {
+                        if($english) {
+                           print STDERR "$input:$line:1: warning: user was not found in $english!\n";
+                           $user = keys %users;  # set to an invalid user so it won't be added
+                        }
+                        else {
+                            # we found a new user, add it to the usermap
+                            $user = ++$numusers;
+                            $users{$userstr} = $user;
+                        }
                     }
 
                     # Use the ID name to figure out which id number range we
@@ -476,14 +497,14 @@ while(<LANG>) {
                     # separated.
 
                     if($idstr =~ /^VOICE/) {
-                        $idnum = $voiceid++;
+                        $idnum = $voiceid[$user]++;
                     }
                     else {
-                        $idnum = $idcount++;
+                        $idnum = $idcount[$user]++;
                     }
                     
                     $id{$idstr} = $idnum;
-                    $idnum[$idnum]=$idstr;
+                    $idnum[$user][$idnum]=$idstr;
                     
                     $source{$idstr}=$src;
                     $dest{$idstr}=$dest;
@@ -570,16 +591,16 @@ if ($sortfile) {
 if($prefix) {
     # We create a .c and .h file
 
-    open(HFILE, ">$prefix/lang.h") ||
+    open(HFILE_CORE, ">$prefix/lang.h") ||
         die "Error: couldn't create file $prefix/lang.h\n";
-    open(CFILE, ">$prefix/lang_core.c") ||
+    open(CFILE_CORE, ">$prefix/lang_core.c") ||
         die "Error: couldn't create file $prefix/lang_core.c\n";        
 
    # get header file name
    $headername = "$prefix/lang.h";
    $headername =~ s/(.*\/)*//;
 
-    print HFILE <<MOO
+    print HFILE_CORE <<MOO
 /* This file was automatically generated using genlang */
 /*
  * The str() macro/functions is how to access strings that might be
@@ -599,8 +620,8 @@ enum \{
 MOO
     ;
 
-    print CFILE <<MOO
-/* This file was automaticly generated using genlang, the strings come
+    print CFILE_CORE <<MOO
+/* This file was automatically generated using genlang, the strings come
    from "$input" */
    
 #include "$headername"
@@ -612,17 +633,17 @@ MOO
 
     # Output the ID names for the enum in the header file
     my $i;
-    for $i (1 .. $idcount) {
-        my $name=$idnum[$i - 1]; # get the ID name
+    for $i (1 .. $idcount[$users{"core"}]) {
+        my $name=$idnum[$users{"core"}][$i - 1]; # get the ID name
         
         $name =~ s/\"//g; # cut off the quotes
         
-        printf HFILE ("    %s, /* %d */\n", $name, $i-1);
+        printf HFILE_CORE ("    %s, /* %d */\n", $name, $i-1);
     }
 
 # Output separation marker for last string ID and the upcoming voice IDs
 
-    print HFILE <<MOO
+    print HFILE_CORE <<MOO
     LANG_LAST_INDEX_IN_ARRAY, /* this is not a string, this is a marker */
     /* --- below this follows voice-only strings --- */
     VOICEONLY_DELIMITER = 0x8000,
@@ -630,20 +651,20 @@ MOO
     ;
 
 # Output the ID names for the enum in the header file
-    for $i (0x8000 .. ($voiceid-1)) {
-        my $name=$idnum[$i]; # get the ID name
+    for $i (0x8000 .. ($voiceid[$users{"core"}]-1)) {
+        my $name=$idnum[$users{"core"}][$i]; # get the ID name
         
         $name =~ s/\"//g; # cut off the quotes
         
-        printf HFILE ("    %s,\n", $name);
+        printf HFILE_CORE ("    %s,\n", $name);
     }
 
     # Output end of enum
-    print HFILE "\n};\n/* end of generated enum list */\n";
+    print HFILE_CORE "\n};\n/* end of generated enum list */\n";
 
     # Output the target phrases for the source file
-    for $i (1 .. $idcount) {
-        my $name=$idnum[$i - 1]; # get the ID
+    for $i (1 .. $idcount[$users{"core"}]) {
+        my $name=$idnum[$users{"core"}][$i - 1]; # get the ID
         my $dest = $dest{$name}; # get the destination phrase
         
         $dest =~ s:\"$:\\0\":; # insert a \0 before the second quote
@@ -653,18 +674,18 @@ MOO
             $dest = '"\0"';
         }
 
-        printf CFILE ("    %s\n", $dest);
+        printf CFILE_CORE ("    %s\n", $dest);
     }
 
 # Output end of string chunk
-    print CFILE <<MOO
+    print CFILE_CORE <<MOO
 ;
 /* end of generated string list */
 MOO
 ;
 
-    close(HFILE);
-    close(CFILE);
+    close(HFILE_CORE);
+    close(CFILE_CORE);
 } # end of the c/h file generation
 elsif($binary) {
     # Creation of a binary lang file was requested
@@ -680,15 +701,15 @@ elsif($binary) {
       $langoptions); # magic lang file header
 
     # loop over the target phrases
-    for $i (1 .. $idcount) {
-        my $name=$idnum[$i - 1]; # get the ID
+    for $i (1 .. $idcount[$users{"core"}]) {
+        my $name=$idnum[$users{"core"}][$i - 1]; # get the ID
         my $dest = $dest{$name}; # get the destination phrase
 
         if($dest) {
             $dest =~ s/^\"(.*)\"\s*$/$1/g; # cut off quotes
 
             # Now, make sure we get the number from the english sort order:
-            $idnum = $idmap{$name};
+            $idnum = $idmap[$users{"core"}]{$name};
 
             printf OUTF ("%c%c%s\x00", ($idnum>>8), ($idnum&0xff), $dest);
         }
@@ -701,22 +722,22 @@ elsif($voiceout) {
     my @engl;
 
     # This loops over the strings in the translated language file order
-    my @ids = ((0 .. ($idcount-1)));
-    push @ids, (0x8000 .. ($voiceid-1));
+    my @ids = ((0 .. ($idcount[$users{"core"}]-1)));
+    push @ids, (0x8000 .. ($voiceid[$users{"core"}]-1));
 
     #for my $id (@ids) {
     #    print "$id\n";
     #}
 
     for $i (@ids) {
-        my $name=$idnum[$i]; # get the ID
+        my $name=$idnum[$users{"core"}][$i]; # get the ID
         my $dest = $voice{$name}; # get the destination voice string
 
         if($dest) {
             $dest =~ s/^\"(.*)\"\s*$/$1/g; # cut off quotes
 
             # Now, make sure we get the number from the english sort order:
-            $idnum = $idmap{$name};
+            $idnum = $idmap[$users{"core"}]{$name};
 
             if(length($idnum)) {
                 $engl[$idnum] = $i;
@@ -739,7 +760,7 @@ elsif($voiceout) {
             next;
         }
 
-        my $name=$idnum[$o]; # get the ID
+        my $name=$idnum[$users{"core"}][$o]; # get the ID
         my $dest = $voice{$name}; # get the destination voice string
         
         print "#$i ($o)\nid: $name\nvoice: $dest\n";
@@ -749,7 +770,7 @@ elsif($voiceout) {
 
 
 if($verbose) {
-    printf("%d ID strings scanned\n", $idcount);
+    printf("%d ID strings scanned\n", $idcount[$users{"core"}]);
 
     print "* head *\n";
     for(keys %head) {
-- 
cgit v1.2.3