# starts abcdeffedbcaabcdef... In the examples `_' shows where a
# space would be printed.
#
-# Small2 Small3 ... Small10
+# Small2 Small3 ... Small9 Small10 Small12 Small14 Small16
# A single line containing as much of the count will fit, eg:
# Small5 3456*
# Small8 _*123456
# will fit without makign the output ambiguous:
# Small9 ab*123456
#
-# Small2S Small4S ... Small10S
-# Small3T Small9T
+# Small2S Small4S ... Small16S
+# Small3T Small9T Small12T
# Same as Small but split into two lines (S)
# or three lines (T). Eg:
# Small4S *4 Small6T _*
# 56 34
# 56
-# Git2 Git3 ... Git10
-# Git4S Git6S ... Git10S
-# Git6T Git9T
+# Git2 Git3 ... Git9 Git10 Git12 Git14 Git16
+# Git4S Git6S ... Git16S
+# Git6T Git9T Git12T
# Just the commit object hash, in one, two (S) or three (T)
# lines. E.g.:
# Git5 abcd*
# Full4 Full6 ... Full20:
# The commit object hash plus the commit count, on
# separate lines, eg:
-# Full6 abcdef Full8 abcdeffe
+# Full12 abcdef Full16 abcdeffe
# *23456 _*123456
#
# Full6T Full9T ... Full30T
$SIG{__WARN__} = sub { die @_; };
+our $debug=0;
+
+if (@ARGV && $ARGV[0] =~ m/^-(D+)$/) {
+ $debug = length $1;
+ shift @ARGV;
+}
+
sub p { print @_ or die $!; }
+sub p_debug { print STDERR @_ if $debug; }
+
p <<'END';
// *** AUTOGENERATED - DO NOT EDIT *** //
function Commitid_pixelsz() =
our @demo;
+
+our $prcount;
+
+sub debug_simplify_begin ($) {
+ my ($chr) = @_;
+
+ return unless $debug;
+
+ open S, ">commitid-DEBUG-simplify-$chr.ps";
+ print S "%!\n";
+ print S "(Courier-Bold) findfont 15 scalefont setfont\n";
+
+ $prcount=0;
+}
+
+sub debug_simplify_done () {
+ return unless $debug;
+ print S "showpage\n";
+ close S or die $!;
+}
+
+sub debug_simplify_pr ($$$) {
+ my ($chr,$polys,$why) = @_;
+
+ return unless $debug;
+
+ print STDERR "PR $chr $why\n";
+ my $ct_x = 10000 * ($prcount % 6);
+ my $ct_y = 18000 * int($prcount / 6);
+ printf S "0 setgray\n";
+ printf S "%d %d moveto\n", map {$_/100 + 10} $ct_x,$ct_y;
+ printf S "(%s) show\n", $why;
+ my $pr_recur;
+
+ $pr_recur = sub {
+ my ($tpolys, @levels) = @_;
+ return unless @$tpolys;
+ foreach my $i (0..$#$tpolys) {
+ printf STDERR "P@levels %02d :", $i;
+ my $pinfo = $tpolys->[$i];
+ my $p = $pinfo->{E};
+ printf STDERR "@$p\n";
+ my $lw = 5 - 4*($i / ($#$tpolys || 1));
+ my $pp = sub {
+ my $spec = $p->[$_[0]];
+ $spec =~ m/^\d{5}/;
+ sprintf "%d %d",map { $_/100 }
+ 1000 + $ct_x + $&,
+ 5000 + $ct_y + $';
+ };
+ printf S "%s setrgbcolor\n", (@levels==0 ? '0 0 0' :
+ @levels==1 ? '0 0 1'
+ : '1 1 0');
+ foreach my $eai (0..$#$p) {
+ my $ebi = ($eai + 1) % @$p;
+ printf S <<END, $lw, $pp->($eai), $pp->($ebi);
+ %f setlinewidth
+ %s moveto
+ %s lineto
+ stroke
+END
+ }
+ $pr_recur->($pinfo->{Holes}, @levels, $i);
+ }
+ };
+ $pr_recur->($polys,0);
+
+ $prcount++;
+}
+
+sub simplify ($$) {
+ my ($chr,$polys) = @_;
+ use Data::Dumper;
+
+ return unless @$polys;
+
+ my $count=0;
+ my $pr = sub { };
+
+ if ($debug) {
+ debug_simplify_begin($chr);
+ }
+
+ $pr->("start");
+
+ AGAIN: while(1) {
+ my %edges;
+ my $found_hole;
+
+ foreach my $pi (0..$#$polys) {
+ my $p = $polys->[$pi]{E};
+ foreach my $ei (0..$#$p) {
+ my $e = $p->[$ei].$p->[($ei+1) % @$p];
+ die if $edges{$e};
+ $edges{$e} = [ $p, $pi, $ei ];
+ }
+ }
+ p_debug "AGAIN $count\n";
+ my $merge = sub {
+ my ($pa, $pai, $eai, $pb, $pbi, $ebi) = @_;
+ p_debug "# merging $pai:$eai.. $pbi:$ebi..\n";
+ splice @$pa, $eai, 1,
+ ((@$pb)[$ebi+1..$#$pb], (@$pb)[0..$ebi-1]);
+ @$pb = ( );
+ };
+ foreach my $pai (0..$#$polys) {
+ my $painfo = $polys->[$pai];
+ my $pa = $painfo->{E};
+ foreach my $eai (0..$#$pa) {
+ my $ear = $pa->[ ($eai+1) % @$pa ].$pa->[$eai];
+ my $ebi = $edges{$ear};
+ next unless $ebi;
+ my ($pb,$pbi);
+ ($pb, $pbi, $ebi) = @$ebi;
+ # $pai:($eai+1)..$eai and $pbi:$ebi..($ebi+1) are identical
+ # so we want to remove them.
+ if ($pai==$pbi) {
+ # we're making a hole! we make an assumption:
+ # holes have fewer line segments than the
+ # outlines. This is almost always true because of
+ # the way we construct our figures.
+ if (($ebi - $eai + @$pa) % @$pa > @$pa/2) {
+ # We arrange that $eai..$ebi is the hole
+ ($ebi,$eai) = ($eai,$ebi);
+ }
+ p_debug "HOLE $eai $ebi\n";
+ # we want to make the smallest hole, to avoid
+ # making a hole that itself needs simplifying
+ my $holesz = ($ebi - $eai + @$pa) % @$pa;
+ $found_hole = [ $pa,$pai,$eai, $ebi, $holesz ]
+ unless $found_hole && $found_hole->[4] < $holesz;
+ } else {
+ $merge->($pa,$pai,$eai,$pb,$pbi,$ebi);
+ debug_simplify_pr($chr,$polys,"after $count");
+ next AGAIN;
+ }
+ }
+ # we process hole joining last, so that the whole of the
+ # edge of the hole must be part of the same polygon
+ if ($found_hole) {
+ p_debug "HOLE DOING @$found_hole\n";
+ my ($pa,$pai,$eai,$ebi) = @$found_hole;
+ # simplify the indexing
+ @$pa = ((@$pa)[$eai..$#$pa], (@$pa)[0..$eai-1]);
+ $ebi -= $eai; $ebi += @$pa; $ebi %= @$pa;
+ $eai = 0;
+ push @{ $painfo->{Holes} }, {
+ E => [ (@$pa)[$eai+1..$ebi-1] ],
+ Holes => [ ],
+ };
+ splice @$pa, $eai, $ebi-$eai+1;
+ debug_simplify_pr($chr,$polys,"hole $count");
+ next AGAIN;
+ }
+ }
+ last;
+ }
+
+ debug_simplify_done();
+}
+
+sub p_edgelist ($$$) {
+ my ($points,$vecs,$p) = @_;
+ my @vec;
+ foreach my $pt (@$p) {
+ $pt =~ s{\d{5}}{$&,};
+ $pt =~ s{\b\d}{$&.}g;
+ push @$points, "[$pt]";
+ push @vec, $#$points;
+ }
+ push @$vecs, \@vec;
+}
+
sub parsefont () {
my %cellmap;
for (;;) {
$cellmap{$1} = $_;
}
my %chrpolys;
+ # $chrs{$chr}[$poly] = $poly
+ # $poly->{E} = [ "012345012345", ... ]
+ # $poly->{Holes} = $poly2
while (<DATA>) {
next unless m/\S/;
chomp;
} elsif (s{^\S}{}) {
my $f = $cellmap{$&};
die unless $f;
- $f =~ s/\b\d/ sprintf '%05d', $col*2000 + $&*1025 /ge;
- $f =~ s/\d\b/ sprintf '%05d', $row*2000 + $&*1025 /ge;
- push @{ $chrpolys{$chr} }, [ split / /, $f ];
+ $f =~ s/\b\d/ sprintf '%05d', $col*2000 + $&*1000 /ge;
+ $f =~ s/\d\b/ sprintf '%05d', $row*2000 + $&*1000 /ge;
+ push @{ $chrpolys{$chr} }, { E => [ split / /, $f ] };
} else {
die "$_ ?";
}
my $demo = '';
my $democols = 6;
foreach my $chr (sort keys %chrpolys) {
+
+ my $polys = $chrpolys{$chr};
+ $_->{Holes} = [] foreach @$polys;
+
+ simplify($chr,$polys);
+
my $mod = chrmodname $chr;
p "module $mod () {\n";
- foreach my $poly (@{ $chrpolys{$chr} }) {
- p " polygon([";
- my $delim = "";
- foreach my $pt (@$poly) {
- p $delim;
- $pt =~ s{\d{5}}{$&,};
- $pt =~ s{\b\d}{$&.}g;
- p "[$pt]";
- $delim = ',';
+ foreach my $poly (@$polys) {
+ p " polygon(";
+ my $holes = $poly->{Holes};
+ my (@points, @vecs);
+ p_edgelist(\@points, \@vecs, $poly->{E});
+ foreach my $hole (@$holes) {
+ p_edgelist(\@points, \@vecs, $hole->{E});
}
- p "]);\n";
+ p "points=[".(join ",",@points)."],";
+ if (@$holes) {
+ p ",paths=[".(join ",",
+ map { "[".(join ",",@$_)."]" }
+ @vecs)."],";
+ }
+ p "convexity=4);\n";
}
p "}\n";
$demo .= $chr;
our $argcounter;
our @forms;
+our %included; # 0 = not at all; 1 = truncated; 2 = full
sub rjustt ($$$;$) {
# right justify and truncate (ie, pad and truncate at left)
# always includes prefix
+ # sets $included{$what}
my ($sz, $what, $whole, $prefix) = @_;
$prefix //= '';
my $lw = length $whole;
my $spare = $sz - $lw - (length $prefix);
+ $included{$what}= 1 + ($spare > 0);
return
($spare > 0 ? (' ' x $spare) : '').
$prefix.
my ($sz, $what, $whole, $suffix) = @_;
$suffix //= '';
$sz -= length $suffix;
+ $included{$what} = 1 + ($sz >= length $whole);
return sprintf "%-${sz}.${sz}s%s", $whole, $suffix;
}
+sub genform_prep() {
+ $included{$_}=0 foreach qw(Objid Count);
+}
+
sub genform ($@) {
my ($form, @lines) = @_;
gentextmodule($form, @lines);
Chars => (length join '', @lines),
Lines => (scalar @lines),
Ambiguous => ($form =~ m/Full/ && !grep { m/\W/ } @lines),
+ Included => { %included },
};
push @forms, $f;
}
print STDERR join ' ', map { $_ // '?' }
"-- commitid", $git_object, $git_dirty, $git_count, "--\n";
- foreach my $sz (2..10) {
+ foreach my $sz (2..10, qw(12 14 16)) {
gentextmodule_demo_start_batch();
if (defined($git_count)) {
+ genform_prep();
my $smallstr = rjustt($sz, 'Count', $git_count, $git_dirty);
my $forgitobj = $sz - length($git_count) - 1;
if (defined($git_object) && $forgitobj >= 2) {
genform_plusq("Small$sz", $smallstr);
}
+ genform_prep();
genform_plusq("Git$sz", ljustt($sz, 'Objid', $git_object, $git_dirty))
if defined $git_object;
- if (defined $git_count && defined $git_object) {
+ if (defined $git_count && defined $git_object && $sz<=10) {
+ genform_prep();
genform("Full".($sz*2),
ljustt($sz, 'Objid', $git_object),
rjustt($sz, 'Count', $git_count, $git_dirty));
+ genform_prep();
my $e = $sz;
genform("Full".($e*3)."T",
ljustt($e*2, 'Objid', $git_object, $git_dirty)
my @do;
foreach my $f (
sort {
+ $b->{Included}{$bestwhat} <=> $a->{Included}{$bestwhat} or
$b->{Chars} <=> $a->{Chars} or
$a->{Lines} <=> $b->{Chars}
}