#!/usr/local/bin/perl
#***************************************************************************
# _ _ ____ _
# Project ___| | | | _ \| |
# / __| | | | |_) | |
# | (__| |_| | _ <| |___
# \___|\___/|_| \_\_____|
#
# Copyright (C) 1998 - 2011, Daniel Stenberg, <daniel@haxx.se>, et al.
#
# This software is licensed as described in the file COPYING, which
# you should have received as part of this distribution. The terms
# are also available at http://curl.haxx.se/docs/copyright.html.
#
# You may opt to use, copy, modify, merge, publish, distribute and/or sell
# copies of the Software, and permit persons to whom the Software is
# furnished to do so, under the terms of the COPYING file.
#
# This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
# KIND, either express or implied.
#
###########################################################################
# Yeah, I know, probably 1000 other persons already wrote a script like
# this, but I'll tell ya:
# THEY DON'T FIT ME :-)
# Get readme file as parameter:
if($ARGV[0] eq "-c") {
$c=1;
shift @ARGV;
}
my $README = $ARGV[0];
if($README eq "") {
print "usage: mkreadme.pl [-c] <README> < manpage\n";
exit;
}
push @out, " _ _ ____ _\n";
push @out, " Project ___| | | | _ \\| |\n";
push @out, " / __| | | | |_) | |\n";
push @out, " | (__| |_| | _ <| |___\n";
push @out, " \\___|\\___/|_| \\_\\_____|\n";
my $olen=0;
while (<STDIN>) {
my $line = $_;
# this should be removed:
$line =~ s/(.|_)//g;
if($line =~ /^([ \t]*\n|curl)/i) {
# cut off headers and empty lines
$wline++; # count number of cut off lines
next;
}
my $text = $line;
$text =~ s/^\s+//g; # cut off preceding...
$text =~ s/\s+$//g; # and trailing whitespaces
$tlen = length($text);
if($wline && ($olen == $tlen)) {
# if the previous line with contents was exactly as long as
# this line, then we ignore the newlines!
# We do this magic because a header may abort a paragraph at
# any line, but we don't want that to be noticed in the output
# here
$wline=0;
}
$olen = $tlen;
if($wline) {
# we only make one empty line max
$wline = 0;
push @out, "\n";
}
push @out, $line;
}
push @out, "\n"; # just an extra newline
open(READ, "<$README") ||
die "couldn't read the README infile $README";
while(<READ>) {
push @out, $_;
}
close(READ);
# if compressed
if($c) {
my @test = `gzip --version 2>&1`;
if($test[0] =~ /gzip/) {
open(GZIP, ">dumpit") ||
die "can't create the dumpit file, try without -c";
binmode GZIP;
for(@out) {
print GZIP $_;
$gzip += length($_);
}
close(GZIP);
system("gzip --best --no-name dumpit");
open(GZIP, "<dumpit.gz") ||
die "can't read the dumpit.gz file, try without -c";
binmode GZIP;
while(<GZIP>) {
push @gzip, $_;
$gzipped += length($_);
}
close(GZIP);
unlink("dumpit.gz");
}
else {
# no gzip, no compression!
undef $c;
print STDERR "MEEEP: Couldn't find gzip, disable compression\n";
}
}
$now = localtime;
print <<HEAD
/*
* NEVER EVER edit this manually, fix the mkhelp.pl script instead!
* Generation time: $now
*/
#ifdef USE_MANUAL
#include "tool_hugehelp.h"
HEAD
;
if($c) {
print <<HEAD
#include <zlib.h>
#include "memdebug.h" /* keep this as LAST include */
static const unsigned char hugehelpgz[] = {
/* This mumbo-jumbo is the huge help text compressed with gzip.
Thanks to this operation, the size of this data shrunk from $gzip
to $gzipped bytes. You can disable the use of compressed help
texts by NOT passing -c to the mkhelp.pl tool. */
HEAD
;
my $c=0;
print " ";
for(@gzip) {
my @all=split(//, $_);
for(@all) {
my $num=ord($_);
printf(" 0x%02x,", 0+$num);
if(++$c>11) {
print "\n ";
$c=0;
}
}
}
print "\n};\n";
print <<EOF
#define BUF_SIZE 0x10000
static voidpf zalloc_func(voidpf opaque, unsigned int items, unsigned int size)
{
(void) opaque;
/* not a typo, keep it calloc() */
return (voidpf) calloc(items, size);
}
static void zfree_func(voidpf opaque, voidpf ptr)
{
(void) opaque;
free(ptr);
}
/* Decompress and send to stdout a gzip-compressed buffer */
void hugehelp(void)
{
unsigned char* buf;
int status,headerlen;
z_stream z;
/* Make sure no gzip options are set */
if (hugehelpgz[3] & 0xfe)
return;
headerlen = 10;
memset(&z, 0, sizeof(z_stream));
z.zalloc = (alloc_func)zalloc_func;
z.zfree = (free_func)zfree_func;
z.avail_in = (unsigned int)(sizeof(hugehelpgz) - headerlen);
z.next_in = (unsigned char *)hugehelpgz + headerlen;
if (inflateInit2(&z, -MAX_WBITS) != Z_OK)
return;
buf = malloc(BUF_SIZE);
if (buf) {
while(1) {
z.avail_out = BUF_SIZE;
z.next_out = buf;
status = inflate(&z, Z_SYNC_FLUSH);
if (status == Z_OK || status == Z_STREAM_END) {
fwrite(buf, BUF_SIZE - z.avail_out, 1, stdout);
if (status == Z_STREAM_END)
break;
}
else
break; /* Error */
}
free(buf);
}
inflateEnd(&z);
}
EOF
;
foot();
exit;
}
else {
print <<HEAD
void hugehelp(void)
{
fputs(
HEAD
;
}
$outsize=0;
for(@out) {
chop;
$new = $_;
$outsize += length($new)+1; # one for the newline
$new =~ s/\\/\\\\/g;
$new =~ s/\"/\\\"/g;
# gcc 2.96 claims ISO C89 only is required to support 509 letter strings
if($outsize > 500) {
# terminate and make another fputs() call here
print ", stdout);\n fputs(\n";
$outsize=length($new)+1;
}
printf("\"%s\\n\"\n", $new);
}
print ", stdout) ;\n}\n";
foot();
sub foot {
print <<FOOT
#endif /* USE_MANUAL */
FOOT
;
}
Submit an article or tip