From: Jarkko Hietaniemi Date: Tue, 9 Apr 2002 23:41:01 +0000 (+0000) Subject: Retract the UTF-8 filenames patch. This may be X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=commitdiff_plain;h=1e2c6ed7b1d7c43be9a4422286be5333f4f3b7f8;p=p5sagit%2Fp5-mst-13.2.git Retract the UTF-8 filenames patch. This may be better dealt with in Perl level, a la File::Spec (since the whole mess is strongly filesystem-specific). p4raw-id: //depot/perl@15835 --- diff --git a/MANIFEST b/MANIFEST index 2a742f9..f917d39 100644 --- a/MANIFEST +++ b/MANIFEST @@ -385,7 +385,6 @@ ext/File/Glob/t/basic.t See if File::Glob works ext/File/Glob/t/case.t See if File::Glob works ext/File/Glob/t/global.t See if File::Glob works ext/File/Glob/t/taint.t See if File::Glob works -ext/File/Glob/t/utf8.t See if File::Glob works ext/File/Glob/TODO File::Glob extension todo list ext/Filter/t/call.t See if Filter::Util::Call works ext/Filter/Util/Call/Call.pm Filter::Util::Call extension module diff --git a/ext/File/Glob/Glob.pm b/ext/File/Glob/Glob.pm index 1ae894f..a704b56 100644 --- a/ext/File/Glob/Glob.pm +++ b/ext/File/Glob/Glob.pm @@ -31,7 +31,6 @@ use XSLoader (); GLOB_NOSPACE GLOB_QUOTE GLOB_TILDE - GLOB_UTF8 ); %EXPORT_TAGS = ( @@ -52,7 +51,6 @@ use XSLoader (); GLOB_NOSPACE GLOB_QUOTE GLOB_TILDE - GLOB_UTF8 glob bsd_glob ) ], @@ -64,11 +62,10 @@ sub import { require Exporter; my $i = 1; while ($i < @_) { - if ($_[$i] =~ /^:(case|nocase|globally|utf8)$/) { + if ($_[$i] =~ /^:(case|nocase|globally)$/) { splice(@_, $i, 1); $DEFAULT_FLAGS &= ~GLOB_NOCASE() if $1 eq 'case'; $DEFAULT_FLAGS |= GLOB_NOCASE() if $1 eq 'nocase'; - $DEFAULT_FLAGS |= GLOB_UTF8() if $1 eq 'utf8'; if ($1 eq 'globally') { local $^W; *CORE::GLOBAL::glob = \&File::Glob::csh_glob; @@ -206,19 +203,15 @@ File::Glob - Perl extension for BSD glob routine ## override the core glob (CORE::glob() does this automatically ## by default anyway, since v5.6.0) use File::Glob ':globally'; - my @sources = <*.{c,h,y}>; + my @sources = <*.{c,h,y}> ## override the core glob, forcing case sensitivity use File::Glob qw(:globally :case); - my @sources = <*.{c,h,y}>; + my @sources = <*.{c,h,y}> ## override the core glob forcing case insensitivity use File::Glob qw(:globally :nocase); - my @sources = <*.{c,h,y}>; - - ## override the core glob forcing UTF-8 names - use File::Glob qw(:globally :utf8); - my @sources = <*.{c,h,y}>; + my @sources = <*.{c,h,y}> =head1 DESCRIPTION @@ -328,17 +321,6 @@ order (case does not matter) rather than in ASCII order. =back -The following flag has been added in the Perl implementation for -Unicode compatibility: - -=over 4 - -=item C - -The filenames returned will be marked as being in UTF-8 encoding of -Unicode. Note that it is your responsibility to ascertain that the -filesystem you are globbing in returns valid UTF-8 filenames. - =head1 DIAGNOSTICS bsd_glob() returns a list of matching paths, possibly zero length. If an diff --git a/ext/File/Glob/Glob.xs b/ext/File/Glob/Glob.xs index 1bb5fc4..bc58b6a 100644 --- a/ext/File/Glob/Glob.xs +++ b/ext/File/Glob/Glob.xs @@ -63,8 +63,6 @@ PPCODE: strlen(pglob.gl_pathv[i]))); TAINT; SvTAINT(tmp); - if (pglob.gl_flags & GLOB_UTF8 && !IN_BYTES) - SvUTF8_on(tmp); PUSHs(tmp); } diff --git a/ext/File/Glob/Makefile.PL b/ext/File/Glob/Makefile.PL index a24b663..0d4267a 100644 --- a/ext/File/Glob/Makefile.PL +++ b/ext/File/Glob/Makefile.PL @@ -26,7 +26,7 @@ WriteConstants( NAME => 'File::Glob', NAMES => [qw(GLOB_ABEND GLOB_ALPHASORT GLOB_ALTDIRFUNC GLOB_BRACE GLOB_ERR GLOB_LIMIT GLOB_MARK GLOB_NOCASE GLOB_NOCHECK GLOB_NOMAGIC - GLOB_NOSORT GLOB_NOSPACE GLOB_QUOTE GLOB_TILDE GLOB_UTF8), + GLOB_NOSORT GLOB_NOSPACE GLOB_QUOTE GLOB_TILDE), {name=>"GLOB_ERROR", macro=>["#ifdef GLOB_ERROR\n\tdMY_CXT;\n\n","#endif\n"]}], BREAKOUT_AT => 8, ); diff --git a/ext/File/Glob/bsd_glob.h b/ext/File/Glob/bsd_glob.h index ef46c00..af92c04 100644 --- a/ext/File/Glob/bsd_glob.h +++ b/ext/File/Glob/bsd_glob.h @@ -76,7 +76,6 @@ typedef struct { #define GLOB_ALPHASORT 0x2000 /* Alphabetic, not ASCII sort, like csh. */ #define GLOB_LIMIT 0x4000 /* Limit pattern match output to ARG_MAX (usually from limits.h). */ -#define GLOB_UTF8 0x8000 /* Return UTF-8. (Perl extension.) */ #define GLOB_NOSPACE (-1) /* Malloc call failed. */ #define GLOB_ABEND (-2) /* Unignored error. */ diff --git a/ext/File/Glob/t/utf8.t b/ext/File/Glob/t/utf8.t deleted file mode 100644 index 97d0b2c..0000000 --- a/ext/File/Glob/t/utf8.t +++ /dev/null @@ -1,17 +0,0 @@ -use File::Glob qw(:globally :utf8); - -# Can't really depend on Tru64 UTF-8 filenames being so must just see -# that things don't crash and that *if* UTF-8 were to be received, it's -# valid. (Maybe later add checks that are run if we are on NTFS/HFS+.) -# (see also t/op/readdir.t) - -print "1..2\n"; - -my $a = <*>; - -print utf8::valid($a) ? "ok 1\n" : "not ok 1\n"; - -my @a=<*>; - -print utf8::valid($a[0]) ? "ok 2\n" : "not ok 2\n"; - diff --git a/opcode.h b/opcode.h index 01b89da..8556e1e 100644 --- a/opcode.h +++ b/opcode.h @@ -1744,7 +1744,7 @@ EXT U32 PL_opargs[] = { 0x0001368c, /* readlink */ 0x0012291c, /* mkdir */ 0x0001379c, /* rmdir */ - 0x0122c814, /* open_dir */ + 0x0002c814, /* open_dir */ 0x0000d600, /* readdir */ 0x0000d60c, /* telldir */ 0x0002c804, /* seekdir */ diff --git a/opcode.pl b/opcode.pl index fc90005..61cd0e8 100755 --- a/opcode.pl +++ b/opcode.pl @@ -814,7 +814,7 @@ rmdir rmdir ck_fun isTu% S? # Directory calls. -open_dir opendir ck_fun is@ F S S? +open_dir opendir ck_fun is@ F S readdir readdir ck_fun % F telldir telldir ck_fun st% F seekdir seekdir ck_fun s@ F S diff --git a/pod/perldelta.pod b/pod/perldelta.pod index 22e5641..60cb016 100644 --- a/pod/perldelta.pod +++ b/pod/perldelta.pod @@ -319,12 +319,6 @@ contain the strings 'UTF-8' or 'UTF8' (case-insensitive matching), the default encoding of your STDIN, STDOUT, and STDERR, and of B, is UTF-8. -=item * - -If your filesystem supports returning UTF-8 encoded filenames, -it is possible to make Perl to understand that the filenames -returned by readdir() and glob() are in Unicode. - =back =head2 Safe Signals diff --git a/pod/perldiag.pod b/pod/perldiag.pod index ee3c617..f22aa80 100644 --- a/pod/perldiag.pod +++ b/pod/perldiag.pod @@ -3716,12 +3716,6 @@ what you are doing you can turn off this warning by C. (F) There are no byte-swapping functions for a machine with this byte order. -=item Unknown discipline '%s' - -(F) You specified an unknown I/O discipline. See L for valid -filehandle disciplines and L for valid directory -handle disciplines. - =item Unknown "re" subpragma '%s' (known ones are: %s) You tried to use an unknown subpragma of the "re" pragma. diff --git a/pod/perlfunc.pod b/pod/perlfunc.pod index 048ecfb..d20851f 100644 --- a/pod/perlfunc.pod +++ b/pod/perlfunc.pod @@ -3024,18 +3024,12 @@ them, and automatically close whenever and however you leave that scope: See L for some details about mixing reading and writing. -=item opendir DIRHANDLE,MODE,EXPR - =item opendir DIRHANDLE,EXPR Opens a directory named EXPR for processing by C, C, C, C, and C. Returns true if successful. DIRHANDLEs have their own namespace separate from FILEHANDLEs. -In three-argument form the middle argument may be C<:utf8> to force -the filenames returned by readdir() to be in UTF-8 encoding of Unicode. -This naturally works only if your filesystem returns UTF-8 filenames. - =item ord EXPR =item ord @@ -3734,10 +3728,6 @@ C there, it would have been testing the wrong file. @dots = grep { /^\./ && -f "$some_dir/$_" } readdir(DIR); closedir DIR; -In some filesystems it is possible to return UTF-8 encoded filenames. -To get readdir() to return such filenames, you must use C<:utf8> with -the three-argument form of opendir(), see L. - =item readline EXPR Reads from the filehandle whose typeglob is contained in EXPR. In scalar diff --git a/pod/perlunicode.pod b/pod/perlunicode.pod index d636700..26e704a 100644 --- a/pod/perlunicode.pod +++ b/pod/perlunicode.pod @@ -20,11 +20,6 @@ Other encodings can be converted to perl's encoding on input, or from perl's encoding on output by use of the ":encoding(...)" layer. See L. -In some filesystems (for example Microsoft NTFS and Apple HFS+) the -filenames are in UTF-8 . By using opendir() and File::Glob you can -make readdir() and glob() to return the filenames as Unicode, see -L and L for details. - To mark the Perl source itself as being in a particular encoding, see L. diff --git a/pod/perluniintro.pod b/pod/perluniintro.pod index 736a0e2..dd3064f 100644 --- a/pod/perluniintro.pod +++ b/pod/perluniintro.pod @@ -407,11 +407,6 @@ If you run this code twice, the contents of the F will be twice UTF-8 encoded. A C would have avoided the bug, or explicitly opening also the F for input as UTF-8. -In some filesystems (for example Microsoft NTFS and Apple HFS+) the -filenames are in UTF-8 . By using opendir() and File::Glob you can -make readdir() and glob() to return the filenames as Unicode, see -L and L for details. - B: the C<:utf8> and C<:encoding> features work only if your Perl has been built with the new "perlio" feature. Almost all Perl 5.8 platforms do use "perlio", though: you can see whether diff --git a/pp_sys.c b/pp_sys.c index aa8fb77..6ed8e0a 100644 --- a/pp_sys.c +++ b/pp_sys.c @@ -3747,39 +3747,17 @@ PP(pp_open_dir) dSP; STRLEN n_a; char *dirname = POPpx; - char *dscp = NULL; - GV *gv; - register IO *io; - bool want_utf8 = FALSE; - - if (MAXARG == 3) - dscp = POPpx; - - gv = (GV*)POPs; - io = GvIOn(gv); + GV *gv = (GV*)POPs; + register IO *io = GvIOn(gv); if (!io) goto nope; - if (dscp) { - if (*dscp == ':') { - if (strnEQ(dscp + 1, "utf8", 4)) - want_utf8 = TRUE; - else - Perl_croak(aTHX_ "Unknown discipline '%s'", dscp); - } - else - Perl_croak(aTHX_ "Unknown discipline '%s'", dscp); - } - if (IoDIRP(io)) PerlDir_close(IoDIRP(io)); if (!(IoDIRP(io) = PerlDir_open(dirname))) goto nope; - if (want_utf8) - IoFLAGS(io) |= IOf_DIR_UTF8; - RETPUSHYES; nope: if (!errno) @@ -3817,8 +3795,6 @@ PP(pp_readdir) if (!(IoFLAGS(io) & IOf_UNTAINT)) SvTAINTED_on(sv); #endif - if (IoFLAGS(io) & IOf_DIR_UTF8 && !IN_BYTES) - SvUTF8_on(sv); XPUSHs(sv_2mortal(sv)); } } @@ -3834,8 +3810,6 @@ PP(pp_readdir) if (!(IoFLAGS(io) & IOf_UNTAINT)) SvTAINTED_on(sv); #endif - if (IoFLAGS(io) & IOf_DIR_UTF8) - sv_utf8_upgrade(sv); XPUSHs(sv_2mortal(sv)); } RETURN; diff --git a/sv.h b/sv.h index 7a13905..92dec20 100644 --- a/sv.h +++ b/sv.h @@ -407,7 +407,6 @@ struct xpvio { #define IOf_UNTAINT 16 /* consider this fp (and its data) "safe" */ #define IOf_NOLINE 32 /* slurped a pseudo-line from empty file */ #define IOf_FAKE_DIRP 64 /* xio_dirp is fake (source filters kludge) */ -#define IOf_DIR_UTF8 128 /* readdir tries to return utf8 */ /* The following macros define implementation-independent predicates on SVs. */ diff --git a/t/op/readdir.t b/t/op/readdir.t index 8e67b65..7cfecdb 100755 --- a/t/op/readdir.t +++ b/t/op/readdir.t @@ -8,7 +8,7 @@ BEGIN { eval 'opendir(NOSUCH, "no/such/directory");'; if ($@) { print "1..0\n"; exit; } -print "1..6\n"; +print "1..3\n"; for $i (1..2000) { local *OP; @@ -43,25 +43,3 @@ while (@R && @G && "op/".$R[0] eq $G[0]) { shift(@G); } if (@R == 0 && @G == 0) { print "ok 3\n"; } else { print "not ok 3\n"; } - -# Can't really depend on Tru64 UTF-8 filenames being so must just see -# that things don't crash and that *if* UTF-8 were to be received, it's -# valid. (Maybe later add checks that are run if we are on NTFS/HFS+.) -# (see also ext/File/Glob/t/utf8.t) - -opendir(OP, ":utf8", "op"); - -my $a = readdir(OP); - -print utf8::valid($a) ? "ok 4\n" : "not ok 4\n"; - -my @a = readdir(OP); - -print utf8::valid($a[0]) ? "ok 5\n" : "not ok 5\n"; - -# But we can check for bogus mode arguments. - -eval { opendir(OP, ":foo", "op") }; - -print $@ =~ /Unknown discipline ':foo'/ ? "ok 6\n" : "not ok 6\n"; -