Change 18822: Re: [perl #21395] rcatline doesn't grok utf8

[Jarkko Hietaniemi]

Change 18822 by [EMAIL PROTECTED] on 2003/03/04 05:14:13

        Subject: Re: [perl #21395] rcatline doesn't grok utf8
        From: Enache Adrian <[EMAIL PROTECTED]>
        Date: Sun, 2 Mar 2003 05:43:54 +0200
        Message-ID: <[EMAIL PROTECTED]>

Affected files ...

... //depot/perl/sv.c#638 edit
... //depot/perl/t/io/utf8.t#27 edit

Differences ...

==== //depot/perl/sv.c#638 (text) ====
Index: perl/sv.c
--- perl/sv.c#637~18813~        Sun Mar  2 21:33:21 2003
+++ perl/sv.c   Mon Mar  3 21:14:13 2003
@@ -6247,7 +6247,27 @@
     (void)SvUPGRADE(sv, SVt_PV);
 
     SvSCREAM_off(sv);
-    SvPOK_only(sv);    /* Validate pointer */
+
+    if (append) {
+       if (PerlIO_isutf8(fp)) {
+           if (!SvUTF8(sv)) {
+               sv_utf8_upgrade_nomg(sv);
+               sv_pos_u2b(sv,&append,0);
+           }
+       } else if (SvUTF8(sv)) {
+           SV *tsv = NEWSV(0,0);
+           sv_gets(tsv, fp, 0);
+           sv_utf8_upgrade_nomg(tsv);
+           SvCUR_set(sv,append);
+           sv_catsv(sv,tsv);
+           sv_free(tsv);
+           goto return_string_or_null;
+       }
+    }
+
+    SvPOK_only(sv);
+    if (PerlIO_isutf8(fp))
+       SvUTF8_on(sv);
 
     if (PL_curcop == &PL_compiling) {
        /* we always read code in line mode */
@@ -6290,7 +6310,7 @@
 #endif
       SvCUR_set(sv, bytesread += append);
       buffer[bytesread] = '\0';
-      goto check_utf8_and_return;
+      goto return_string_or_null;
     }
     else if (RsPARA(PL_rs)) {
        rsptr = "\n\n";
@@ -6543,12 +6563,7 @@
        }
     }
 
-check_utf8_and_return:
-    if (PerlIO_isutf8(fp))
-       SvUTF8_on(sv);
-    else
-       SvUTF8_off(sv);
-
+return_string_or_null:
     return (SvCUR(sv) - append) ? SvPVX(sv) : Nullch;
 }
 

==== //depot/perl/t/io/utf8.t#27 (xtext) ====
Index: perl/t/io/utf8.t
--- perl/t/io/utf8.t#26~17172~  Mon Jun 10 12:27:23 2002
+++ perl/t/io/utf8.t    Mon Mar  3 21:14:13 2003
@@ -12,7 +12,7 @@
 no utf8; # needed for use utf8 not griping about the raw octets
 
 $| = 1;
-print "1..31\n";
+print "1..33\n";
 
 open(F,"+>:utf8",'a');
 print F chr(0x100).'�';
@@ -271,6 +271,28 @@
     close(F);
 
     print $@ =~ /Wide character in print/ ? "ok 31\n" : "not ok 31\n";
+}
+
+{
+    open F, ">:bytes","a"; print F "\xde"; close F;
+
+    open F, "<:bytes", "a";
+    my $b = chr 0x100;
+    $b .= <F>;
+    print $b eq chr(0x100).chr(0xde) ? "ok 32" : "not ok 32";
+    print " \#21395 '.= <>' utf8 vs. bytes\n";
+    close F;
+}
+
+{
+    open F, ">:utf8","a"; print F chr 0x100; close F;
+
+    open F, "<:utf8", "a";
+    my $b = "\xde";
+    $b .= <F>;
+    print $b eq chr(0xde).chr(0x100) ? "ok 33" : "not ok 33";
+    print " \#21395 '.= <>' bytes vs. utf8\n";
+    close F;
 }
 
 # sysread() and syswrite() tested in lib/open.t since Fnctl is used
End of Patch.