From e881000941ea96b6b6f0010f929c5c5c789522f0 Mon Sep 17 00:00:00 2001 From: Daniel Stenberg Date: Fri, 10 Aug 2007 10:40:21 +0000 Subject: [PATCH] initial add from my private repo, 'roffit' was internal revision 1.17 --- CHANGES | 56 +++++++ release.pl | 22 +++ roffit | 467 +++++++++++++++++++++++++++++++++++++++++++++++++++++ roffit.1 | 58 +++++++ 4 files changed, 603 insertions(+) create mode 100644 CHANGES create mode 100755 release.pl create mode 100755 roffit create mode 100644 roffit.1 diff --git a/CHANGES b/CHANGES new file mode 100644 index 0000000..a29ef66 --- /dev/null +++ b/CHANGES @@ -0,0 +1,56 @@ +* (Jan 2, 2005) +- Steve Benson made roffit translate all '\-' sequenes to just '-'. + +* (April 5, 2005) +- Make a bare .BR do a "
" +- Make .pp start a new paragraph +- Ignore the weird '\" comment latex2man adds +- Ignore de, ft and "..". I *believe* they are use to set a "macro". + +* (March 15, 2005) +- Added comments about .nf and .fi based on + http://www.rfc-editor.org/nroff.html + +* (June 21, 2004) +- make .fi properly end a .nf section + +* (April 12, 2004) +- convert <> properly within the .BR data as well + +* (March 5, 2004) +- make links for ftp:// as well +- strip \& completely instead of making "&zerosp;" (which isn't good html) + +0.6 (February 27, 2004) + - added --mandir and --hrefdir + - made it convert http:// URLs to links + - added --version + +0.5 (February 3, 2004) + - Added support for more &-combos + - Better treatment of .nf to make 
 sections, and we now do newlines
+   better within such a 
 section
+ - .IR is now treated equally as .I
+ - keywords and their data can be tab-separated
+ - .SS is now treated as .SH
+
+0.4 (December 5, 2003)
+ - Andrew Dixon filed a bug report with a man page generated with the
+   help2man tool. roffit now deals with \fR as well as \fP as an end
+   marker for special word formatting.
+ - I moved the changelog into this file
+ - Wrote a first basic man page for roffit
+
+0.3 (November 14, 2003)
+ Daniel Fandrich brought:
+ - deals with .lp lines
+ - .TH needs no section portion anymore
+ - added generator meta tag in the header
+
+0.2 (November 7, 2003)
+ - fixed the  name for the SH section
+ - added  links from all words within \fIthis\fP or \fBthis\fP
+   that has the same text as a .SH or .IP.
+
+0.1 (November 5, 2003)
+ - Initial public release
diff --git a/release.pl b/release.pl
new file mode 100755
index 0000000..70ad098
--- /dev/null
+++ b/release.pl
@@ -0,0 +1,22 @@
+#!/usr/bin/perl
+
+my @files=('CHANGES', 'roffit', 'roffit.1');
+
+my $version=$ARGV[0];
+
+if(!$version) {
+    print "Usage: release.pl \n";
+    exit;
+}
+
+my $dir = "roffit-$version";
+
+mkdir $dir;
+
+for(@files) {
+    `cp $_ $dir`;
+}
+
+`tar czf roffit-${version}.tar.gz $dir`;
+
+`rm -rf $dir`;
diff --git a/roffit b/roffit
new file mode 100755
index 0000000..3e78d3c
--- /dev/null
+++ b/roffit
@@ -0,0 +1,467 @@
+#!/usr/bin/env perl 
+#
+# roffit: convert man page source files to HTML
+#
+# Read an nroff file. Output a HTML file.
+#
+# This is a very simple script, but I use it on very simple man pages as I've
+# found no decent script that makes really nice web pages.
+#
+# $Id: roffit,v 1.1 2007/08/10 10:40:22 bagder Exp $
+#
+my $version = "0.7";
+# Author:  Daniel Stenberg 
+# Please mail me improvements.
+#
+
+use strict;
+#use warnings;
+
+my $InFH = \*STDIN;
+my $OutFH = \*STDOUT;
+my $debugFH = \*STDERR;
+
+my %manpage;
+my @out;
+
+my $indentlevel=0; # logical levels, not columns
+my @p;
+my $within_tp;
+my $standalone=1; # by default we make stand-alone HTML pages
+my $pre;
+my %anchor; # hash with all anchors
+
+my $mandir;
+my $hrefdir=".";
+
+while($ARGV[0]) {
+    if($ARGV[0] eq "--bare") {
+        # don't include headers and stuff
+        $standalone=0;
+        shift @ARGV;
+    }
+    elsif($ARGV[0] eq "--version") {
+        print "roffit $version (http://daniel.haxx.se/projects/roffit/)\n";
+        exit;
+    }
+    elsif($ARGV[0] =~ /^--mandir=(.*)/) {
+        # check for other HTMLized man pages in this given dir
+        $mandir=$1;
+        shift @ARGV;
+    }
+    elsif($ARGV[0] =~ /^--hrefdir=(.*)/) {
+        # if another manpage is found in the --mandir, this is the dir we
+        # prefix the HTML file name with in the output HTML
+        $hrefdir=$1;
+        shift @ARGV;
+    }
+    else {
+        printf $debugFH "unknown option: %s\n", $ARGV[0] if($ARGV[0] ne "-h");
+        print $debugFH "Usage: roffit [options] < infile > outfile\n",
+        "Options:\n",
+        " --version       display roffit version and exit\n",
+        " --bare          do not put in HTML, HEAD, BODY tags\n",
+        " --mandir=  check for other HTMLized man pages in this dir\n",
+        " --hrefdir= if a manpage is found in the --mandir, this is the\n",
+        "                 dir we prefix the HTML file name with in the output\n";
+        exit;
+    }
+}
+
+sub showp {
+    my @p = @_;
+    push @out, "\n" if(!$pre);
+    push @out, "
", @p;
+}
+
+sub defaultcss {
+    print $OutFH <
+P.level0 {
+ padding-left: 2em;
+}
+
+P.level1 {
+ padding-left: 4em;
+}
+
+P.level2 {
+ padding-left: 6em;
+}
+
+span.emphasis {
+ font-style: italic;
+}
+
+span.bold {
+ font-weight: bold;
+}
+
+span.manpage {
+ font-weight: bold;
+}
+
+h2.nroffsh {
+ background-color: #e0e0e0;
+}
+
+span.nroffip {
+ font-weight: bold;
+ font-size: 120%;
+ font-family: monospace;
+}
+
+p.roffit {
+ text-align: center;
+ font-size: 80%;
+}
+
+ENDOFCSS
+    ;
+}
+
+sub text2name {
+    my ($text) = @_;
+    $text =~ s/^ *([^ ]*).*/$1/g;
+    $text =~ s/[^a-zA-Z0-9-]//g;
+    return $text;
+}
+
+# scan through the file and check for  sections we should convert
+# to proper links
+sub linkfile {
+    my @new;
+    for(@out) {
+        my $line=$_;
+        my $l;
+        while($line =~ s/([^<]*)<\/span>/[]/) {
+            my ($style, $name)=($1, $2);
+            
+            $l = text2name($name);
+
+            #printf $debugFH "$style - $name - %s - %d\n",
+            #$l, $anchor{$l};
+
+            my $link;
+            if($anchor{$l}) {
+                $link="$name";
+            }
+            else {
+                if(($name =~ /^ *(\w*) *\( *(\d*) *\) *$/) && ($mandir)) {
+
+                    # this looks like a reference to another man page, and we
+                    # have asked for this feature! We check for the specified
+                    # nroff file and not the HTML version, to avoid depending
+                    # on which order the set of files are converted to HTML!
+                    my ($symbol, $section)=($1, $2);
+                    my $file = "$mandir/$symbol.$section";
+
+                    if(-r $file) {
+                        my $html = "$hrefdir/$symbol.html";
+                        # there is such a HTML file present, produce a link
+                        # to it!
+                        $link="$name";
+                    }
+                }
+                if(!$link) {
+                    $link="$name";
+                }
+            }
+            $line =~ s/\[\]/$link/;
+        }
+        # convert lowercase ftp:// or http:// links to  links
+        $line =~ s/((http|ftp):\/\/([a-z0-9.\/_%-]*[a-z\/]))/$1<\/a>/g;
+        push @new, $line;
+    }
+    return @new;
+}
+
+sub parsefile {
+    my $lineno;
+    while(<$InFH>) {
+        $lineno++;
+        my $in = $_;
+        my $out;
+  #     print $debugFH "DEBUG IN: $_";
+
+        $in =~ s/[\r\n]//g if(!$pre); # tear off newlines
+
+        $in =~ s/\r//g; # remote carriage return always
+
+        # substitue all '\-' sequences with '-'.
+        # '\-' might be in the source file to specify a literal
+        # '-' rather than a hyphen.
+        $in =~ s/\\-/-/g;
+
+        if($in =~ /^\'\\\"/) {
+            # ignore this kind of weird comment
+            $out ="";
+        }
+        elsif($in =~ /^\.([^ \t\n]*)(.*)/) {
+            # this is a line starting with a dot, that means it is special
+            my ($keyword, $rest) = ($1, $2);
+            $out = "";
+            
+            # cut off initial spaces
+            $rest =~ s/^ +//g;
+            
+            if($keyword eq "\\\"") {
+                # this is a comment, skip this line
+            }
+            elsif($keyword =~ /^TH$/i) {
+                # man page header:
+                # curl 1 "22 Oct 2003" "Curl 7.10.8" "Curl Manual"
+                # NAME SECTION DATE VERSION MANUAL
+                if($rest =~ /([^ ]*) (\d+) \"([^\"]*)\" \"([^\"]*)\"(\"([^\"]*)\")?/) {
+                    # strict matching only so far
+                    $manpage{'name'} = $1;
+                    $manpage{'section'} = $2;
+                    $manpage{'date'} = $3;
+                    $manpage{'version'} = $4;
+                    $manpage{'manual'} = $6;
+                }
+            }
+            elsif($keyword =~ /^S[HS]$/i) {
+                # SS is treated the same as SH
+                # Section Header
+                showp(@p);
+                @p="";
+                if($pre) {
+                    push @out, "
\n";
+                    $pre = 0;
+                }
+
+                my $name = text2name($rest);
+                $anchor{$name}=1;
+
+                $rest =~ s/\"//g; # cut off quotes
+                $rest =~ s//>/g;
+                $out = "
$rest
";
+                $indentlevel=0;
+                $within_tp=0;
+            }
+            elsif(($keyword =~ /^B$/i) || ($keyword =~ /^BI$/i)) {
+                # Make B and BI the same for simplicity
+                $rest =~ s/\"//g; # cut off quotes
+                $rest =~ s//>/g;
+
+                # This is pretty lame, but since a .B section and itself
+                # contain a \fI[blabla]\fP section, we cut off all the \f
+                # occurances within the .B text to make it easier for ourself.
+                $rest =~ s/\\f[IPR]//g;
+
+                push @p, "$rest ";
+                if($pre) {
+                    push @p, "\n";
+                }
+            }
+            elsif(($keyword =~ /^I$/i) || ($keyword =~ /^IR$/i)) {
+                $rest =~ s/\"//g; # cut off quotes
+                $rest =~ s//>/g;
+                push @p, "$rest ";
+            }
+            elsif($keyword =~ /^RS$/i) {
+                # the start of another indent-level. for inlined tables
+                # within an "IP"
+                showp(@p);
+                @p="";
+                $indentlevel++;
+            }
+            elsif($keyword =~ /^RE$/i) {
+                # end of the RS section
+                showp(@p);
+                @p="";
+                $indentlevel--;
+            }
+            elsif($keyword =~ /^NF$/i) {
+                # We let nf start a 
 section .nf = no-fill, you use for
+                # graphs or text that you don't want spaces to be ignored
+
+                showp(@p);
+                @p="";
+                push @out, "
\n";
+                $pre=1
+            }
+            elsif($keyword =~ /^TP$/i) {
+                # Used within an "RS" section to make a new line. The first
+                # TP as a column indicator, but we decide to do that
+                # controlling in the CSS instead.
+                $within_tp=1;
+                showp(@p);
+                @p="";                
+            }
+            elsif($keyword =~ /^IP$/i) {
+                # start of a new paragraph coming up
+                showp(@p);
+                @p="";
+
+                my $name= text2name($rest);
+                $anchor{$name}=1;
+
+                $rest =~ s/\"//g; # cut off quotes
+                $rest =~ s//>/g;
+                
+                $indentlevel-- if ($indentlevel);
+                push @p, "$rest ";
+                # make this a single-line title
+                showp(@p);
+                @p="";
+                $indentlevel++;
+                $within_tp=0;
+            }
+            elsif($keyword =~ /^ad$/i) {
+                showp(@p);
+                @p="";
+            }
+            elsif($keyword =~ /^sp$/i) {
+                showp(@p);
+                @p="";
+            }
+            elsif($keyword =~ /^lp$/i) {
+                # marks end of a paragraph
+                showp(@p);
+                @p="";
+            }
+            elsif($keyword =~ /^pp$/i) {
+                # PP ends a TP section, but some TP sections don't use it
+                # Often used to separate paragraphs
+                $within_tp=0;
+                showp(@p);
+                @p="";
+            }
+            elsif($keyword =~ /^fi$/i) {
+                # .fi = fill-in, extra space will be ignored and text that is
+                # entered like this, the fill-in command will continue until
+                # you enter a .nf command and vice-versa
+
+                showp(@p);
+                @p="";
+                if($pre) {
+                    # if this is the end of a .nf (
) section
+                    push @out, "
\n";
+                    $pre=0; # disable pre again
+                }
+
+            }
+            elsif($keyword =~ /^(de|ft|\.)$/i) {
+                # no idea what these do!
+            }
+            elsif($keyword =~ /^so$/i) {
+                # This keyword refers to a different man page, named in the
+                # $rest.
+                # We don't support this
+                push @out, "See the $rest man page.\n";
+            }
+            elsif($keyword =~ /^BR$/i) {
+                # I'm not sure what this does exactly, but this is commonly
+                # used to include pointers to other man pages. Let's assume
+                # it only does that for now.
+                # blabla (3)
+                # or "blabla (3)"
+                # or strcmp "(3), " strcasecmp "(3)"
+                # etc
+                
+                $rest =~ s/\"//g; # cut off quotes
+                $rest =~ s//>/g;
+
+                if(!$rest) {
+                    # A stand-alone .br will become a line break
+                    #print $debugFH "ALERT: bare BR\n";
+                    push @p, "
";
+                }
+                else {
+                    my @all = split /,/, $rest;
+                    for(@all) {
+                        if(/([^ ]*) *\((\d+)\)/) {
+                            # TODO: this looks like a man page, check if
+                            # there's a HTML file for it and if so make a link
+                            # to it
+                        }
+                        
+                        push @p, "$_ ";
+                    }
+                }
+            }
+            else {
+                showp(@p);
+                @p="";
+                print $debugFH "ALERT: unknown keyword \"$keyword\" on line $lineno\n";
+            }
+        }
+        else {
+            # text line, decode \-stuff
+            my $txt = $in;
+
+            $txt =~ s//>\;/g;
+            $txt =~ s/\\&//g;
+            $txt =~ s/\\-/-/g;
+            $txt =~ s/\\ / \;/g;
+            $txt =~ s/\\\'/´\;/g;
+            $txt =~ s/\\\(co/©\;/g;
+            $txt =~ s/\\fI//g;
+            $txt =~ s/\\fB//g;
+            $txt =~ s/\\f[PR]/<\/span>/g;
+
+            # replace backslash [something] with just [something]
+            $txt =~ s/\\(.)/$1/g;
+
+            if($txt =~ /^[ \t\r\n]*$/) {
+                # no contents, marks end of a paragraph
+                showp(@p);
+                @p="";
+            }
+            else {
+                $txt =~ s/^ /\ \;/g;
+                push @p, "$txt ";
+            }
+            $out ="";
+        }
+
+        if($out) {
+            push @out, $out;
+   #         print $debugFH "DEBUG OUT: $out\n";
+        }
+        else {
+   #         print $debugFH "DEBUG OUT: [withheld]\n";
+        }
+    }
+    showp(@p);
+}
+
+parsefile();
+
+my @conv = linkfile();
+
+my $title=sprintf("%s man page",
+                  $manpage{'name'}?$manpage{'name'}:"secret");
+
+if($standalone) {
+    print $OutFH <
+$title
+
+MOO
+    ;
+    defaultcss();
+    print "\n";
+}
+else {
+    print "\n";
+}
+
+print $OutFH @conv;
+print $OutFH <
+ This HTML page was made with roffit.
+ROFFIT
+    ;
+
+if($standalone) {
+    print "\n";
+}
diff --git a/roffit.1 b/roffit.1
new file mode 100644
index 0000000..291886b
--- /dev/null
+++ b/roffit.1
@@ -0,0 +1,58 @@
+.\" You can view this file with:
+.\" nroff -man roffit.1
+.\" Written by Daniel Stenberg
+.\"
+.TH roffit 1 "27 Feb 2004" "roffit 0.6" "roffit Manual"
+.SH NAME
+roffit \- convert nroff to HTML
+.SH SYNOPSIS
+.B roffit [options] < inputfile > outputfile
+.SH DESCRIPTION
+.B roffit
+converts the \fIinputfile\fP to \fIoutputfile\fP. The \fIinputfile\fP must be
+an nroff formatted man page, and the \fIoutputfile\fP will be an HTML
+document.
+.SH OPTIONS
+.IP --bare
+The output HTML will not include any HTML, HEAD or BODY tags. Also not that
+when this is selected, there will be no inlined CSS but you will have to
+define the necessary classes yourself.
+.IP --version
+Display version number and exit
+.IP --mandir=
+Set a directory in which \fIroffit\fP will check for other man pages (in nroff
+[name].[num] format) that this one refers to. If found, a  link will
+be made to that page with a html extension instead of the number. The file
+name in the generated link will be prefixed by the dir given with
+\fI--hrefdir\fP.
+
+This works for references specified as \fImanpage(3)\fP (within the emhpasis
+foformatting) and in a plain \.BR section (often used in the SEE ALSO
+section).
+.IP --hrefdir=
+Specify a directory to prefix generated href links created with the --mandir
+option. This defaults to ".".
+.SH "CSS CLASSES"
+.IP h2.nroffsh
+The nroff ".SH" section. These are normally the "headlines" before each sub
+section within the man page.
+.IP p.nroffip
+The nroff ".IP" section. These are normally the bullet kind used to list each
+option out of several in a list. Usually followed by a \fIp.level0\fP
+containining the descriptive text.
+.IP p.level0
+Normal text at the first level. Normally not indented.
+.IP p.level1
+Text at next logical indent level. Normally indented one level right.
+.IP p.level2
+Text at the second logical indent level. Normally indented two levels right.
+.IP span.bold
+Text marked to be bold.
+.IP span.manpage
+Text marked as a reference to another man page.
+.IP span.emphasis
+Text marked to be emphasized.
+.IP p.roffit
+Used for the advertising final paragraph.
+.SH WWW
+http://daniel.haxx.se/projects/roffit