bug 180266 : hard-code precompiled CCMaps to speed up start-up.

This perl script is not a part of build, but is only for information
because it's used to gen. hard-coded CCMaps.
This commit is contained in:
jshin%mailaps.org 2003-03-05 03:42:38 +00:00
Родитель aac4dd7004
Коммит aeffe82333
1 изменённых файлов: 385 добавлений и 0 удалений

Просмотреть файл

@ -0,0 +1,385 @@
#!/usr/bin/perl -w
# This script is used to generate precompiled CCMap files.
# See bug 180266 for details.
use strict;
use vars qw($fill_fmt $fu_sz);
use vars qw($e_mid_offset $e_pg_offset $f_mid_offset $f_pg_offset);
(@ARGV < 1 ) and usage();
my $ifn = $ARGV[0];
my ($ifh, $class);
open $ifh , "< $ifn" or die "Cannot open $ifn";
if (@ARGV >= 2) {
$class = $ARGV[1];
printf STDERR
"$0:\n\t CLASS $class is specified in the command line.\n" .
"\t The class spec. in the input file will be ignored.\n";
}
use constant PLANE_SZ => 0x10000;
use constant MID_SZ => PLANE_SZ / 16;
use constant PG_SZ => MID_SZ / 16;
# Unlike FillInfo() method in Mozilla, let's use 16bit integer
# to pack the character coverage/representability. This way,
# we can just copy fillinfo to fill up page maps later.
use constant {
FILL_SZ => PLANE_SZ / 16,
MID_FILL_SZ => MID_SZ / 16,
PG_FILL_SZ => PG_SZ / 16
};
# network byte order short. actually, byte order doesn't matter.
$fill_fmt = "n1";
$fu_sz = length(pack $fill_fmt, 0); # fillinfo unit size in byte (size of short)
$e_mid_offset = 16;
$e_pg_offset = 32;
$f_pg_offset = 0;
$f_mid_offset = 0;
my @ccmap = ();
my %pg_offsets = ();
my @fillinfo = ();
my %comments = ();
#get all upper pointers to point at empty mid pointers
push @ccmap, ($e_mid_offset) x 16;
#get all mid-pointers to point at empty page.
push @ccmap, ($e_pg_offset) x 16;
push @ccmap, (0) x 16; # empty pg
&read_input(\@fillinfo,$ifh,\%comments);
if (!defined($class) && !defined($comments{'CLASS'}))
{
printf STDERR "Class name is not specified in the cmd line. " .
"Neither is it found in the input file.\n\n" ;
usage();
}
$class = $comments{'CLASS'} if (! defined($class));
foreach my $mid (0 .. 15)
{
my(@mid_fill) = splice @fillinfo, 0, MID_FILL_SZ;
# convert 256(MID_FILL_SZ) 16bit integers to a string of 256 * $fu_sz
# characters.
my($mid_str) = pack $fill_fmt x MID_FILL_SZ, @mid_fill;
# for an empty mid, upper-pointer is already pointing to the empty mid.
next if ($mid_str eq "\0" x ($fu_sz * MID_FILL_SZ));
# for a full mid, add full mid if necessary.
if ($mid_str eq "\xff" x ($fu_sz * MID_FILL_SZ)) {
($f_mid_offset, $f_pg_offset) =
add_full_mid(\@ccmap, $f_pg_offset) unless ($f_mid_offset);
$ccmap[$mid] = $f_mid_offset;
next;
}
my $mid_offset = add_new_mid(\@ccmap,$mid);
foreach my $pg (0 .. 15) {
my(@pg_fill) = splice @mid_fill, 0, PG_FILL_SZ;
my($pg_str) = pack $fill_fmt x PG_FILL_SZ, @pg_fill;
# for an empty pg, mid-pointer is already pointing to the empty page.
next if ($pg_str eq "\x0" x ($fu_sz * PG_FILL_SZ));
# for a full pg, add the full pg if necessary.
# and set the mid-pointer to the full pg offset.
if ($pg_str eq "\xff" x ($fu_sz * PG_FILL_SZ)) {
if (! $f_pg_offset) {
$f_pg_offset = @ccmap;
#for the full pg, endianess and ALU size are immaterial.
push @ccmap, (0xffff) x 16;
}
$ccmap[$mid_offset + $pg ] = $f_pg_offset;
next;
}
$ccmap[$mid_offset + $pg] = @ccmap;
# 'Flag' the offset as the beginning of a page with actual data as
# opposed to pointer sections.
$pg_offsets{scalar @ccmap} = @ccmap;
push @ccmap, @pg_fill;
}
}
&print_ccmap(\@ccmap, \%pg_offsets, $class . ".ccmap", \%comments);
#&print_ccmap(\@ccmap, "32BE", \%pg_offsets, $class . ".32be", \%comments);
#&print_ccmap(\@ccmap, "64BE", \%pg_offsets, $class . ".64be", \%comments);
exit 0;
# END of Main
sub usage
{
print STDERR <<USAGE;
Usage: $0 input_file [class]
The output file "class.ccmap" will be generated with
all three cases LE(16/32/64bit)/BE(16bit), BE(32bit), and BE(64bit)
put together.
When 'class' is omitted, it has to be specified in the input file with
the following syntax
CLASS:: class_name
USAGE
exit 1;
}
sub read_input
{
my($fillinfo_p, $input, $comments_p) = @_;
@$fillinfo_p = (0) x FILL_SZ;
my($lc)=0;
while (<$input>)
{
$lc++;
chomp;
/^CLASS::/ and
($comments_p->{'CLASS'} = $_) =~ s/^CLASS::\s*([a-zA-Z0-9_]+).*$/$1/,
next;
/^DESCRIPTION::/ and
($comments_p->{'DESC'} = $_) =~ s/^DESCRIPTION::\s*//, next;
/^FILE::/ and
($comments_p->{'FILE'} = $_) =~ s/^FILE::\s*//, next;
next unless /^\s*0[Xx][0-9A-Fa-f]{4}/;
/^\s*(.*)\s*$/;
my ($u, $comment) = split /\s+/, $1, 2;
$u =~ s/,//g;
$u =~ tr/A-Z/a-z/;
next if /^0x.*[^0-9a-f]+.*/;
my $usv = oct $u;
if ( 0xd800 <= $usv && $usv <= 0xdfff) { #surrogate pair
printf STDERR "Invalid input $u at %4d\n", $lc;
}
$fillinfo_p->[($usv >> 4) & 0xfff] |= (1 << ($usv & 0x0f));
# printf STDERR "input %s(%04x) \@line %d : put %04x @ %04x\n",
# $u,$usv, $lc, (1 << ($usv & 0x0f)), ($usv >> 4) & 0xfff;
$comments_p->{$usv} = "";
# Remove '/*' and '*/' (C style comment) or '//' (C++ style comment)
# or ':' and store only the textual content of the comment.
if (defined($comment)) {
($comments_p->{$usv} = $comment)
=~ s !
(?:/\*|//|:)? # '/*', '//' or ':' or NULL. Do not store.
\s* # zero or more of white space(s)
([^*]+) # one or more of non-white space(s).Store it
# in $1 for the reference in replace part.
\s* # zero or more of white space(s)
(?:\*/)? # '*/' or NONE. Do not store
!$1!sx # replace the whole match with $1 stored above.
}
}
}
sub add_full_mid
{
my($ccmap_p, $f_pg_offset) = @_;
# add a full page if not yet added.
if (! $f_pg_offset) {
$f_pg_offset = @$ccmap_p;
push @$ccmap_p, (0xffff) x 16;
}
# add the full mid-pointer array with all the pointers pointing to the full page.
my $f_mid_offset = @$ccmap_p;
push @$ccmap_p, ($f_pg_offset) x 16;
return ($f_mid_offset, $f_pg_offset);
}
sub add_new_mid
{
my($ccmap_p, $mid) = @_;
my $mid_offset = @$ccmap_p;
$ccmap_p->[$mid] = $mid_offset;
#by default, all mid-pointers point to the empty page.
push @$ccmap_p, ($e_pg_offset) x 16;
return $mid_offset;
}
sub print_ccmap
{
my($ccmap_p,$pg_offset_p, $ofn, $comments_p) = @_;
my(@idxlist);
open OUT, "> $ofn" or
die "cannot open $ofn for output\n";
print OUT print_preamble($class);
print OUT "\n/*\n";
defined ($comments_p->{'CLASS'}) and
print OUT " CLASS:: $comments_p->{'CLASS'}\n";
defined ($comments_p->{'DESC'}) and
print OUT " DESCRIPTION:: $comments_p->{'DESC'}\n";
defined ($comments_p->{'FILE'}) and
print OUT " FILE:: $comments_p->{'FILE'}\n";
print OUT "\n";
for my $usv (sort keys %$comments_p) {
next if ($usv =~ /CLASS/);
printf OUT " 0X%04X : %s\n", $usv, $comments_p->{$usv};
}
printf OUT "*/\n\n";
# When CCMap is accessed, (PRUint16 *) is casted to
# the pointer type of the ALU of a machine.
# For little endian machines, the size of the ALU
# doesn't matter (16, 32, 64). For Big endian
# machines with 32/64 bit ALU, two/four 16bit words
# have to be rearranged to be interpreted correctly
# as 32bit or 64bit integers with the 16bit word
# at the lowest addr. taking the highest place value.
# This shuffling is NOT necessary for the upper pointer section
# and mid-pointer sections.
foreach my $fmt ("LE", "32BE", "64BE")
{
for ($fmt) {
/LE/ and do {
@idxlist = (0, 1, 2, 3);
print OUT "#if (defined(IS_LITTLE_ENDIAN) || ALU_SIZE == 16)\n" .
"// Precompiled CCMap for Big Endian(16bit)/Little" .
" Endian(16/32/64bit) \n";
last;
};
/32BE/ and do {
@idxlist = (1, 0, 3, 2);
print OUT "#elif (ALU_SIZE == 32)\n" .
"// Precompiled CCMap for Big Endian(32bit)\n";
last;
};
/64BE/ and do {
@idxlist = (3, 2, 1, 0);
print OUT "#elif (ALU_SIZE == 64)\n" .
"// Precompiled CCMap for Big Endian(64bit)\n";
last;
};
}
my($offset) = 0;
while ($offset < @$ccmap_p) {
printf OUT "/* %04x */ ", $offset;
for my $i (0 .. 3) {
for my $j (defined($pg_offset_p->{$offset}) ? @idxlist : (0,1,2,3)) {
printf OUT "0x%04X,", $ccmap_p->[$offset + $i * 4 + $j];
}
print OUT "\n " if $i==1;
}
print OUT "\n";
$offset += 16;
}
}
print OUT <<END;
#else
#error "We don't support this architecture."
#endif
END
close OUT;
}
sub print_preamble
{
my($class) = @_;
sprintf <<PREAMBLE;
/* ***** BEGIN LICENSE BLOCK *****
* Version: NPL 1.1/GPL 2.0/LGPL 2.1
*
* The contents of this file are subject to the Netscape Public License
* Version 1.1 (the "License"); you may not use this file except in
* compliance with the License. You may obtain a copy of the License at
* http://www.mozilla.org/NPL/
*
* Software distributed under the License is distributed on an "AS IS" basis,
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
* for the specific language governing rights and limitations under the
* License.
*
* The Original Code is Mozilla Communicator client code.
*
* Contributor(s): Jungshik Shin <jshin\@mailaps.org> (Original developer)
*
*
* Alternatively, the contents of this file may be used under the terms of
* either the GNU General Public License Version 2 or later (the "GPL"), or
* the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
* in which case the provisions of the GPL or the LGPL are applicable instead
* of those above. If you wish to allow use of your version of this file only
* under the terms of either the GPL or the LGPL, and not to allow others to
* use your version of this file under the terms of the NPL, indicate your
* decision by deleting the provisions above and replace them with the notice
* and other provisions required by the GPL or the LGPL. If you do not delete
* the provisions above, a recipient may use your version of this file under
* the terms of any one of the NPL, the GPL or the LGPL.
*
* ***** END LICENSE BLOCK ***** */
/*========================================================
This file contains a precompiled CCMap for a class of Unicode
characters ($class) to be identified quickly by Mozilla.
It was generated by ccmapbin.pl which you can find under
mozilla/intl/unicharutil/tools.
Enumerated below are characters included in the precompiled CCMap
which is human-readable but not so human-friendly. If you
needs to modify the list of characters belonging to "$class",
you have to make a new file (with the name of your choice)
listing characters (one character per line) you want to put
into "$class" in the format
0xuuuu // comment
In addition, the input file can have the following optional lines that
read
CLASS::$class
DESCRIPTION:: description of a character class
FILE:: mozilla source file to include output files
Then, run the following in the current directory.
perl ccmapbin.pl input_file [$class]
which will generate $class.ccmap.
(see bug 180266 and bug 167136)
*/
PREAMBLE
}