[Po4a-commits] po4a/lib/Locale/Po4a Man.pm,1.86,1.87

[Nicolas FRANCOIS]

Update of /cvsroot/po4a/po4a/lib/Locale/Po4a
In directory haydn:/tmp/cvs-serv24545/lib/Locale/Po4a

Modified Files:
	Man.pm 
Log Message:
There were a problem with the added/removed non-breaking spaces.
This could be a general issue (not only for non-breaking spaces).

This was triggered by a French translation containing a "à" (agrave).
When the output document is UTF-8, this character is encoded as 0xC3 0xA0,
and this last 0xA0 was changed to "\ " in post_trans.

This fix consists in using the non-breaking spaces from the input document
charset (with a fall back to a non probable string) before the translation.
After the translation, we change the non-breaking spaces to the groff one
only if the output charset allows non-breaking spaces.


Index: Man.pm
===================================================================
RCS file: /cvsroot/po4a/po4a/lib/Locale/Po4a/Man.pm,v
retrieving revision 1.86
retrieving revision 1.87
diff -u -d -r1.86 -r1.87
--- Man.pm	10 Apr 2005 19:23:18 -0000	1.86
+++ Man.pm	23 Apr 2005 15:50:52 -0000	1.87
@@ -293,6 +293,12 @@
 # .    - a single char (e.g. B, I, R, P, 1, 2, 3, 4, etc.)
 my $FONT_RE = "\\\\f(?:\\[[^\\]]*\\]|\\(..|[^\\(\\[])";
 
+# Variable used to identify non breaking spaces.
+# These non breaking spaces are used to ease the parsing, and a
+# translator can use them in her translation (and they will be translated
+# into the groff non-breaking space.
+my $nbs;
+
 #########################
 #### DEBUGGING STUFF ####
 #########################
@@ -588,8 +594,8 @@
     $str =~ s/\\\(dq/"/sg;
     
     # non-breaking spaces
-    # some \xA0 may have been added during the parsing
-    $str =~ s|\xA0|\\ |sg;
+    # some non-breaking spaces may have been added during the parsing
+    $str =~ s/\Q$nbs/\\ /sg;
 
     print STDERR "$str\n" if ($debug{'pretrans'});
     return $str;
@@ -628,11 +634,14 @@
     $str =~ s/\n([.'])/ $1/mg;
 
     # Change ascii non-breaking space to groff one
-    $str =~ s|\xA0|\\ |sg;
+    my $nbs_out = "\xA0";
+    my $enc_length = Encode::from_to($nbs_out, "latin1",
+                                               $self->get_out_charset);
+    $str =~ s/\Q$nbs_out/\\ /sg if defined $enc_length;
     # No nbsp (said "\ " in groff on the last pos of the line, or groff adds
     # an extra space
     $str =~ s/\\ \n/\\ /sg;
-	
+
     # Make sure we compute internal sequences right.
     # think about: B<AZE E<lt> EZA E<gt>>
     while ($str =~ m/^(.*)(CW|[RBI])<(.*)$/s) {
@@ -729,6 +738,15 @@
                             # MACRONO: don't wrap because we saw the nf macro. It stays so 
                             #          until the next fi macro.
 
+
+    # change the non-breaking space according to the input document charset
+    $nbs = "\xA0";
+    my $enc_length = Encode::from_to($nbs, "latin-1",
+                                           $self->{TT}{'file_in_charset'});
+    # fall back solution
+    $nbs = "PO4A:VERY_IMPROBABLE_STRING_USEDFOR_NON-BREAKING-SPACES";
+        unless defined $enc_length;
+
   LINE:
     undef $self->{type};
     ($line,$ref)=$self->shiftline();
@@ -751,7 +769,7 @@
 
 	    # Split on spaces for arguments, but not spaces within double quotes
 	    my @args=();
-	    $line =~ s/\\ /\xA0/g; # This is probably not needed
+	    $line =~ s/\\ /$nbs/g; # This is probably not needed
 	    push @args,$arg1;
 	    push @args, splitargs($ref,$arguments);
 
@@ -887,7 +905,7 @@
     # will be translated will have the same change again (in pre_trans and
     # post_trans), but the ones which won't get translated are not changed
     # anymore. Let's play safe.
-    $arguments =~ s/\\ /\xA0/g;
+    $arguments =~ s/\\ /$nbs/g;
     $arguments =~ s/^ +//;
     foreach my $elem (split (/ +/,$arguments)) {
         print STDERR ">>Seen $elem(buffer=$buffer;esc=$escaped)\n"
@@ -904,8 +922,8 @@
                 print STDERR "End of quote, with stuff after it\n"
                     if ($debug{'splitargs'});
                 my ($a,$b)=($1,$2);
-                $a =~ s/\xA0/\\ /g;
-                $b =~ s/\xA0/\\ /g;
+                $a =~ s/\Q$nbs/\\ /g;
+                $b =~ s/\Q$nbs/\\ /g;
                 push @args,$a;
                 push @args,$b;
                 $buffer = "";
@@ -913,7 +931,7 @@
                 print STDERR "End of a quote\n"
                     if ($debug{'splitargs'});
                 my $a = $1;
-                $a =~ s/\xA0/\\ /g;
+                $a =~ s/\Q$nbs/\\ /g;
                 push @args,$a;
                 $buffer = "";
             } elsif ($escaped) {
@@ -921,9 +939,9 @@
                     if ($debug{'splitargs'});
                 unless(length($elem)){
                     die wrap_ref_mod($ref, "po4a::man", dgettext("po4a",
-			"Escaped space at the end of macro arg. With high probability, it won't do the trick with po4a (because of wrapping). You may want to remove it and use the .nf/.fi groff macro to control the wrapping."));
+                        "Escaped space at the end of macro arg. With high probability, it won't do the trick with po4a (because of wrapping). You may want to remove it and use the .nf/.fi groff macro to control the wrapping."));
                 }
-                $buffer =~ s/\xA0/\\ /g;
+                $buffer =~ s/\Q$nbs/\\ /g;
                 push @args,$buffer;
                 $buffer = "";
                 $escaped = 0;
@@ -934,7 +952,7 @@
             my $a = $1;
             # Inside a quoted argument, "" indicates a double quote
             $a =~ s/""/\\(dq/g;
-            $a =~ s/\xA0/\\ /g;
+            $a =~ s/\Q$nbs/\\ /g;
             push @args,$a;
         } elsif ($elem =~ m/^"/) { #") {
             print STDERR "Begin of a quoting arg\n"
@@ -957,7 +975,7 @@
         # Inside a quoted argument, "" indicates a double quote
         while ($buffer =~ s/^"([^"]*?)""/"$1\\(dq/) {}
         $buffer=~ s/"//g;
-        $buffer =~ s/\xA0/\\ /g;
+        $buffer =~ s/\Q$nbs/\\ /g;
         push @args,$buffer;
     }
     if ($debug{'splitargs'}) {