#! /usr/bin/env perl
-# Copyright 2002-2016 The OpenSSL Project Authors. All Rights Reserved.
+# Copyright 2002-2019 The OpenSSL Project Authors. All Rights Reserved.
#
-# Licensed under the OpenSSL license (the "License"). You may not use
+# Licensed under the Apache License 2.0 (the "License"). You may not use
# this file except in compliance with the License. You can obtain a copy
# in the file LICENSE in the source distribution or at
# https://www.openssl.org/source/license.html
require 5.10.0;
use warnings;
use strict;
+
use Pod::Checker;
use File::Find;
use File::Basename;
use lib catdir(dirname($0), "perl");
use OpenSSL::Util::Pod;
+# Set to 1 for debug output
+my $debug = 0;
+
# Options.
+our($opt_d);
+our($opt_e);
our($opt_s);
-our($opt_u);
+our($opt_o);
our($opt_h);
-our($opt_n);
our($opt_l);
+our($opt_n);
+our($opt_p);
+our($opt_u);
+our($opt_v);
+our($opt_c);
-sub help()
-{
+# Print usage message and exit.
+sub help {
print <<EOF;
Find small errors (nits) in documentation. Options:
+ -c List undocumented commands and options
+ -d Detailed list of undocumented (implies -u)
+ -e Detailed list of new undocumented (implies -v)
+ -h Print this help message
-l Print bogus links
-n Print nits in POD pages
- -s Also print missing sections in POD pages (implies -n)
- -u List undocumented functions
- -h Print this help message
+ -o Causes -e/-v to count symbols added since 1.1.1 as new (implies -v)
+ -u Count undocumented functions
+ -v Count new undocumented functions
EOF
exit;
}
+getopts('cdehlnouv');
+
+help() if $opt_h;
+$opt_u = 1 if $opt_d;
+$opt_v = 1 if $opt_o || $opt_e;
+die "Cannot use both -u and -v"
+ if $opt_u && $opt_v;
+die "Cannot use both -d and -e"
+ if $opt_d && $opt_e;
+
+# We only need to check c, l, n, u and v.
+# Options d, e, o imply one of the above.
+die "Need one of -[cdehlnouv] flags.\n"
+ unless $opt_c or $opt_l or $opt_n or $opt_u or $opt_v;
+
+
my $temp = '/tmp/docnits.txt';
my $OUT;
-
-my %mandatory_sections =
- ( '*' => [ 'NAME', 'DESCRIPTION', 'COPYRIGHT' ],
- 1 => [ 'SYNOPSIS', 'OPTIONS' ],
- 3 => [ 'SYNOPSIS', 'RETURN VALUES' ],
- 5 => [ ],
- 7 => [ ] );
+my %public;
+my $status = 0;
+
+my %mandatory_sections = (
+ '*' => [ 'NAME', 'DESCRIPTION', 'COPYRIGHT' ],
+ 1 => [ 'SYNOPSIS', 'OPTIONS' ],
+ 3 => [ 'SYNOPSIS', 'RETURN VALUES' ],
+ 5 => [ ],
+ 7 => [ ]
+);
+
+
+# Print error message, set $status.
+sub err {
+ print join(" ", @_), "\n";
+ $status = 1
+}
# Cross-check functions in the NAME and SYNOPSIS section.
-sub name_synopsis()
-{
+sub name_synopsis {
my $id = shift;
my $filename = shift;
my $contents = shift;
return unless $contents =~ /=head1 NAME(.*)=head1 SYNOPSIS/ms;
my $tmp = $1;
$tmp =~ tr/\n/ /;
- print "$id trailing comma before - in NAME\n" if $tmp =~ /, *-/;
- $tmp =~ s/-.*//g;
- $tmp =~ s/,//g;
+ err($id, "trailing comma before - in NAME")
+ if $tmp =~ /, *-/;
+ $tmp =~ s/ -.*//g;
+ err($id, "POD markup among the names in NAME")
+ if $tmp =~ /[<>]/;
+ $tmp =~ s/ */ /g;
+ err($id, "missing comma in NAME")
+ if $tmp =~ /[^,] /;
my $dirname = dirname($filename);
- my $simplename = basename($filename);
- $simplename =~ s/.pod$//;
+ my $simplename = basename(basename($filename, ".in"), ".pod");
my $foundfilename = 0;
my %foundfilenames = ();
my %names;
- foreach my $n ( split ' ', $tmp ) {
+ foreach my $n ( split ',', $tmp ) {
+ $n =~ s/^\s+//;
+ $n =~ s/\s+$//;
+ err($id, "the name '$n' contains white-space")
+ if $n =~ /\s/;
$names{$n} = 1;
$foundfilename++ if $n eq $simplename;
$foundfilenames{$n} = 1
if -f "$dirname/$n.pod" && $n ne $simplename;
}
- print "$id the following exist as other .pod files:\n",
- join(" ", sort keys %foundfilenames), "\n"
+ err($id, "the following exist as other .pod files:",
+ sort keys %foundfilenames)
if %foundfilenames;
- print "$id $simplename (filename) missing from NAME section\n",
+ err($id, "$simplename (filename) missing from NAME section")
unless $foundfilename;
+ if ( $filename !~ /internal/ ) {
+ foreach my $n ( keys %names ) {
+ err($id, "$n is not public")
+ if !defined $public{$n};
+ }
+ }
# Find all functions in SYNOPSIS
return unless $contents =~ /=head1 SYNOPSIS(.*)=head1 DESCRIPTION/ms;
my $syn = $1;
foreach my $line ( split /\n+/, $syn ) {
+ next unless $line =~ /^\s/;
my $sym;
$line =~ s/STACK_OF\([^)]+\)/int/g;
+ $line =~ s/SPARSE_ARRAY_OF\([^)]+\)/int/g;
$line =~ s/__declspec\([^)]+\)//;
if ( $line =~ /env (\S*)=/ ) {
# environment variable env NAME=...
$sym = $1;
} elsif ( $line =~ /typedef.*\(\*(\S+)\)\(.*/ ) {
- # a callback function: typedef ... (*NAME)(...
+ # a callback function pointer: typedef ... (*NAME)(...
+ $sym = $1;
+ } elsif ( $line =~ /typedef.* (\S+)\(.*/ ) {
+ # a callback function signature: typedef ... NAME(...
$sym = $1;
} elsif ( $line =~ /typedef.* (\S+);/ ) {
# a simple typedef: typedef ... NAME;
} elsif ( $line =~ /enum (\S*) \{/ ) {
# an enumeration: enum ... {
$sym = $1;
- } elsif ( $line =~ /#define ([A-Za-z0-9_]+)/ ) {
+ } elsif ( $line =~ /#(?:define|undef) ([A-Za-z0-9_]+)/ ) {
$sym = $1;
} elsif ( $line =~ /([A-Za-z0-9_]+)\(/ ) {
$sym = $1;
else {
next;
}
- print "$id $sym missing from NAME section\n"
+ err($id, "$sym missing from NAME section")
unless defined $names{$sym};
$names{$sym} = 2;
# Do some sanity checks on the prototype.
- print "$id prototype missing spaces around commas: $line\n"
+ err($id, "prototype missing spaces around commas: $line")
if ( $line =~ /[a-z0-9],[^ ]/ );
}
foreach my $n ( keys %names ) {
next if $names{$n} == 2;
- print "$id $n missing from SYNOPSIS\n";
+ err($id, "$n missing from SYNOPSIS")
}
}
-sub check()
-{
+# Check if SECTION ($3) is located before BEFORE ($4)
+sub check_section_location {
+ my $id = shift;
+ my $contents = shift;
+ my $section = shift;
+ my $before = shift;
+
+ return unless $contents =~ /=head1 $section/
+ and $contents =~ /=head1 $before/;
+ err($id, "$section should appear before $before section")
+ if $contents =~ /=head1 $before.*=head1 $section/ms;
+}
+
+# Check if a =head1 is duplicated, or a =headX is duplicated within a
+# =head1. Treats =head2 =head3 as equivalent -- it doesn't reset the head3
+# sets if it finds a =head2 -- but that is good enough for now. Also check
+# for proper capitalization, trailing periods, etc.
+sub check_head_style {
+ my $id = shift;
+ my $contents = shift;
+ my %head1;
+ my %subheads;
+
+ foreach my $line ( split /\n+/, $contents ) {
+ next unless $line =~ /^=head/;
+ if ( $line =~ /head1/ ) {
+ err($id, "duplicate section $line")
+ if defined $head1{$line};
+ $head1{$line} = 1;
+ %subheads = ();
+ } else {
+ err($id, "duplicate subsection $line")
+ if defined $subheads{$line};
+ $subheads{$line} = 1;
+ }
+ err($id, "period in =head")
+ if $line =~ /\.[^\w]/ or $line =~ /\.$/;
+ err($id, "not all uppercase in =head1")
+ if $line =~ /head1.*[a-z]/;
+ err($id, "all uppercase in subhead")
+ if $line =~ /head[234][ A-Z0-9]+$/;
+ }
+}
+
+# Because we have options and symbols with extra markup, we need
+# to take that into account, so we need a regexp that extracts
+# markup chunks, including recursive markup.
+# please read up on /(?R)/ in perlre(1)
+# (note: order is important, (?R) needs to come before .)
+# (note: non-greedy is important, or something like 'B<foo> and B<bar>'
+# will be captured as one item)
+my $markup_re =
+ qr/( # Capture group
+ [BIL]< # The start of what we recurse on
+ (?:(?-1)|.)*? # recurse the whole regexp (refering to
+ # the last opened capture group, i.e. the
+ # start of this regexp), or pick next
+ # character. Do NOT be greedy!
+ > # The end of what we recurse on
+ )/x; # (the x allows this sort of split up regexp)
+
+# Options must start with a dash, followed by a letter, possibly
+# followed by letters, digits, dashes and underscores, and the last
+# character must be a letter or a digit.
+# We do also accept the single -? or -n, where n is a digit
+my $option_re =
+ qr/(?:
+ \? # Single question mark
+ |
+ \d # Single digit
+ |
+ - # Single dash (--)
+ |
+ [[:alpha:]](?:[-_[:alnum:]]*?[[:alnum:]])?
+ )/x;
+
+# Helper function to check if a given $thing is properly marked up
+# option. It returns one of these values:
+# undef if it's not an option
+# "" if it's a malformed option
+# $unwrapped the option with the outermost B<> wrapping removed.
+sub normalise_option {
+ my $id = shift;
+ my $filename = shift;
+ my $thing = shift;
+
+ my $unwrapped = $thing;
+ my $unmarked = $thing;
+
+ # $unwrapped is the option with the outer B<> markup removed
+ $unwrapped =~ s/^B<//;
+ $unwrapped =~ s/>$//;
+ # $unmarked is the option with *all* markup removed
+ $unmarked =~ s/[BIL]<|>//msg;
+
+
+ # If we found an option, check it, collect it
+ if ( $unwrapped =~ /^\s*-/ ) {
+ return $unwrapped # return option with outer B<> removed
+ if $unmarked =~ /^-${option_re}$/;
+ return ""; # Malformed option
+ }
+ return undef; # Something else
+}
+
+# Checks of command option (man1) formatting. The man1 checks are
+# restricted to the SYNOPSIS and OPTIONS sections, the rest is too
+# free form, we simply cannot be too strict there.
+
+sub option_check {
+ my $id = shift;
+ my $filename = shift;
+ my $contents = shift;
+
+ my $synopsis = ($contents =~ /=head1\s+SYNOPSIS(.*?)=head1/s, $1);
+
+ # Some pages have more than one OPTIONS section, let's make sure
+ # to get them all
+ my $options = '';
+ while ( $contents =~ /=head1\s+[A-Z ]*?OPTIONS$(.*?)(?==head1)/msg ) {
+ $options .= $1;
+ }
+
+ # Look for options with no or incorrect markup
+ while ( $synopsis =~
+ /(?<![-<[:alnum:]])-(?:$markup_re|.)*(?![->[:alnum:]])/msg ) {
+ err($id, "Malformed option [1] in SYNOPSIS: $&");
+ }
+
+ while ( $synopsis =~ /$markup_re/msg ) {
+ my $found = $&;
+ print STDERR "$id:DEBUG[option_check] SYNOPSIS: found $found\n"
+ if $debug;
+ my $option_uw = normalise_option($id, $filename, $found);
+ err($id, "Malformed option [2] in SYNOPSIS: $found")
+ if defined $option_uw && $option_uw eq '';
+ }
+
+ # In OPTIONS, we look for =item paragraphs.
+ # (?=^\s*$) detects an empty line.
+ while ( $options =~ /=item\s+(.*?)(?=^\s*$)/msg ) {
+ my $item = $&;
+
+ while ( $item =~ /(\[\s*)?($markup_re)/msg ) {
+ my $found = $2;
+ print STDERR "$id:DEBUG[option_check] OPTIONS: found $&\n"
+ if $debug;
+ err($id, "Unexpected bracket in OPTIONS =item: $item")
+ if ($1 // '') ne '' && $found =~ /^B<\s*-/;
+
+ my $option_uw = normalise_option($id, $filename, $found);
+ err($id, "Malformed option in OPTIONS: $found")
+ if defined $option_uw && $option_uw eq '';
+ }
+ }
+}
+
+# Normal symbol form
+my $symbol_re = qr/[[:alpha:]_][_[:alnum:]]*?/;
+
+# Checks of function name (man3) formatting. The man3 checks are
+# easier than the man1 checks, we only check the names followed by (),
+# and only the names that have POD markup.
+sub functionname_check {
+ my $id = shift;
+ my $filename = shift;
+ my $contents = shift;
+
+ while ( $contents =~ /($markup_re)\(\)/msg ) {
+ print STDERR "$id:DEBUG[functionname_check] SYNOPSIS: found $&\n"
+ if $debug;
+
+ my $symbol = $1;
+ my $unmarked = $symbol;
+ $unmarked =~ s/[BIL]<|>//msg;
+
+ err($id, "Malformed symbol: $symbol")
+ unless $symbol =~ /^B<.*>$/ && $unmarked =~ /^${symbol_re}$/
+ }
+
+ # We can't do the kind of collecting coolness that option_check()
+ # does, because there are too many things that can't be found in
+ # name repositories like the NAME sections, such as symbol names
+ # with a variable part (typically marked up as B<foo_I<TYPE>_bar>
+}
+
+# This is from http://man7.org/linux/man-pages/man7/man-pages.7.html
+my %preferred_words = (
+ 'bitmask' => 'bit mask',
+ 'builtin' => 'built-in',
+ #'epoch' => 'Epoch', # handled specially, below
+ 'file name' => 'filename',
+ 'file system' => 'filesystem',
+ 'host name' => 'hostname',
+ 'i-node' => 'inode',
+ 'lower case' => 'lowercase',
+ 'lower-case' => 'lowercase',
+ 'non-zero' => 'nonzero',
+ 'path name' => 'pathname',
+ 'pseudo-terminal' => 'pseudoterminal',
+ 'reserved port' => 'privileged port',
+ 'system port' => 'privileged port',
+ 'realtime' => 'real-time',
+ 'real time' => 'real-time',
+ 'runtime' => 'run time',
+ 'saved group ID'=> 'saved set-group-ID',
+ 'saved set-GID' => 'saved set-group-ID',
+ 'saved user ID' => 'saved set-user-ID',
+ 'saved set-UID' => 'saved set-user-ID',
+ 'set-GID' => 'set-group-ID',
+ 'setgid' => 'set-group-ID',
+ 'set-UID' => 'set-user-ID',
+ 'setuid' => 'set-user-ID',
+ 'super user' => 'superuser',
+ 'super-user' => 'superuser',
+ 'super block' => 'superblock',
+ 'super-block' => 'superblock',
+ 'time stamp' => 'timestamp',
+ 'time zone' => 'timezone',
+ 'upper case' => 'uppercase',
+ 'upper-case' => 'uppercase',
+ 'useable' => 'usable',
+ 'userspace' => 'user space',
+ 'user name' => 'username',
+ 'zeroes' => 'zeros'
+);
+
+# Search manpage for words that have a different preferred use.
+sub wording {
+ my $id = shift;
+ my $contents = shift;
+
+ foreach my $k ( keys %preferred_words ) {
+ # Sigh, trademark
+ next if $k eq 'file system'
+ and $contents =~ /Microsoft Encrypted File System/;
+ err($id, "found '$k' should use '$preferred_words{$k}'")
+ if $contents =~ /\b\Q$k\E\b/i;
+ }
+ err($id, "found 'epoch' should use 'Epoch'")
+ if $contents =~ /\bepoch\b/;
+}
+
+# Perform all sorts of nit/error checks on a manpage
+sub check {
my $filename = shift;
my $dirname = basename(dirname($filename));
}
my $id = "${filename}:1:";
+ check_head_style($id, $contents);
+
+ # Check ordering of some sections in man3
+ if ( $filename =~ m|man3/| ) {
+ check_section_location($id, $contents, "RETURN VALUES", "EXAMPLES");
+ check_section_location($id, $contents, "SEE ALSO", "HISTORY");
+ check_section_location($id, $contents, "EXAMPLES", "SEE ALSO");
+ }
+
+ # Make sure every link has a section.
+ while ( $contents =~ /$markup_re/msg ) {
+ my $target = $1;
+ next unless $target =~ /^L</; # Skip if not L<...>, or
+ next if $target =~ /::/; # links to a Perl module, or
+ next if $target =~ m@L</@; # links within the page, or
+ next if $target =~ /^L<https?:/; # is a URL link, or
+ next if $target =~ m@\([1357]\)>$@; # it has a section, or
+ next if $target =~ m@\([1357]\)/.*>$@; # it has a section/anchor
+ err($id, "Section missing in $target")
+ }
+ # Check for proper in-man-3 API links.
+ while ( $contents =~ /L<([^>]*)\(3\)(?:\/.*)?>/g ) {
+ my $target = $1;
+ err($id, "Bad L<$target>")
+ unless $target =~ /^[_[:alpha:]][_[:alnum:]]*$/
+ }
+
+ unless ( $contents =~ /=for openssl generic/ ) {
+ if ( $filename =~ m|man3/| ) {
+ name_synopsis($id, $filename, $contents);
+ functionname_check($id, $filename, $contents);
+ } elsif ( $filename =~ m|man1/| ) {
+ option_check($id, $filename, $contents)
+ }
+ }
- &name_synopsis($id, $filename, $contents)
- unless $contents =~ /=for comment generic/
- or $filename =~ m@man[157]/@;
+ wording($id, $contents);
- print "$id doesn't start with =pod\n"
+ err($id, "doesn't start with =pod")
if $contents !~ /^=pod/;
- print "$id doesn't end with =cut\n"
+ err($id, "doesn't end with =cut")
if $contents !~ /=cut\n$/;
- print "$id more than one cut line.\n"
+ err($id, "more than one cut line.")
if $contents =~ /=cut.*=cut/ms;
- print "$id missing copyright\n"
+ err($id, "EXAMPLE not EXAMPLES section.")
+ if $contents =~ /=head1 EXAMPLE[^S]/;
+ err($id, "WARNING not WARNINGS section.")
+ if $contents =~ /=head1 WARNING[^S]/;
+ err($id, "missing copyright")
if $contents !~ /Copyright .* The OpenSSL Project Authors/;
- print "$id copyright not last\n"
+ err($id, "copyright not last")
if $contents =~ /head1 COPYRIGHT.*=head/ms;
- print "$id head2 in All uppercase\n"
+ err($id, "head2 in All uppercase")
if $contents =~ /head2\s+[A-Z ]+\n/;
- print "$id extra space after head\n"
+ err($id, "extra space after head")
if $contents =~ /=head\d\s\s+/;
- print "$id period in NAME section\n"
+ err($id, "period in NAME section")
if $contents =~ /=head1 NAME.*\.\n.*=head1 SYNOPSIS/ms;
- print "$id POD markup in NAME section\n"
- if $contents =~ /=head1 NAME.*[<>].*=head1 SYNOPSIS/ms;
- print "$id Duplicate $1 in L<>\n"
+ err($id, "Duplicate $1 in L<>")
if $contents =~ /L<([^>]*)\|([^>]*)>/ && $1 eq $2;
- print "$id Bad =over $1\n"
- if $contents =~ /=over([^ ][^4])/;
-
- # Look for multiple consecutive openssl #include lines.
- # Consecutive because of files like md5.pod. Sometimes it's okay
- # or necessary, as in ssl/SSL_set1_host.pod
- if ( $contents !~ /=for comment multiple includes/ ) {
+ err($id, "Bad =over $1")
+ if $contents =~ /=over([^ ][^24])/;
+ err($id, "Possible version style issue")
+ if $contents =~ /OpenSSL version [019]/;
+
+ if ( $contents !~ /=for openssl multiple includes/ ) {
+ # Look for multiple consecutive openssl #include lines
+ # (non-consecutive lines are okay; see man3/MD5.pod).
if ( $contents =~ /=head1 SYNOPSIS(.*)=head1 DESCRIPTION/ms ) {
my $count = 0;
foreach my $line ( split /\n+/, $1 ) {
if ( $line =~ m@include <openssl/@ ) {
- if ( ++$count == 2 ) {
- print "$id has multiple includes\n";
- }
+ err($id, "has multiple includes")
+ if ++$count == 2;
} else {
$count = 0;
}
}
}
- return unless $opt_s;
-
- # Find what section this page is in. If run from "." assume
- # section 3.
- my $section = 3;
- $section = $1 if $dirname =~ /man([1-9])/;
-
- foreach ((@{$mandatory_sections{'*'}}, @{$mandatory_sections{$section}})) {
- print "$id: missing $_ head1 section\n"
- if $contents !~ /^=head1\s+${_}\s*$/m;
- }
-
open my $OUT, '>', $temp
or die "Can't open $temp, $!";
podchecker($filename, $OUT);
}
close $OUT;
unlink $temp || warn "Can't remove $temp, $!";
-}
-my %dups;
+ # Find what section this page is in; assume 3.
+ my $section = 3;
+ $section = $1 if $dirname =~ /man([1-9])/;
-sub parsenum()
-{
+ foreach ( (@{$mandatory_sections{'*'}}, @{$mandatory_sections{$section}}) ) {
+ err($id, "missing $_ head1 section")
+ if $contents !~ /^=head1\s+${_}\s*$/m;
+ }
+}
+
+# Parse libcrypto.num, etc., and return sorted list of what's there.
+sub parsenum {
my $file = shift;
my @apis;
or die "Can't open $file, $!, stopped";
while ( <$IN> ) {
+ next if /^#/;
next if /\bNOEXIST\b/;
- next if /\bEXPORT_VAR_AS_FUNC\b/;
- push @apis, $1 if /([^\s]+).\s/;
+ my @fields = split();
+ die "Malformed line $_"
+ if scalar @fields != 2 && scalar @fields != 4;
+ push @apis, $fields[0];
}
close $IN;
- print "# Found ", scalar(@apis), " in $file\n";
return sort @apis;
}
-sub getdocced()
+# Parse all the manpages, getting return map of what they document
+# (by looking at their NAME sections).
+sub getdocced
{
my $dir = shift;
my %return;
+ my %dups;
foreach my $pod ( glob("$dir/*.pod") ) {
my %podinfo = extract_pod_info($pod);
foreach my $n ( @{$podinfo{names}} ) {
$return{$n} = $pod;
- print "# Duplicate $n in $pod and $dups{$n}\n"
+ err("# Duplicate $n in $pod and $dups{$n}")
if defined $dups{$n} && $dups{$n} ne $pod;
$dups{$n} = $pod;
}
return %return;
}
+# Map of documented functions; function => manpage
my %docced;
+# Map of links in each POD file; filename => [ "foo(1)", "bar(3)", ... ]
+my %link_map = ();
+# Map of names in each POD file; "name(s)" => filename
+my %name_map = ();
-sub printem()
+# Load file of symbol names that we know aren't documented.
+sub loadmissing($)
{
+ my $missingfile = shift;
+ my @missing;
+
+ open FH, $missingfile
+ || die "Can't open $missingfile";
+ while ( <FH> ) {
+ chomp;
+ next if /^#/;
+ push @missing, $_;
+ }
+ close FH;
+
+ return @missing;
+}
+
+# Check for undocumented macros; ignore those in the "missing" file
+# and do simple check for #define in our header files.
+sub checkmacros {
+ my $count = 0;
+ my %seen;
+ my @missing;
+
+ if ( $opt_o ) {
+ @missing = loadmissing('util/missingmacro111.txt');
+ } elsif ( $opt_v ) {
+ @missing = loadmissing('util/missingmacro.txt');
+ }
+
+ foreach my $f ( glob('include/openssl/*.h') ) {
+ # Skip some internals we don't want to document yet.
+ next if $f eq 'include/openssl/asn1.h';
+ next if $f eq 'include/openssl/asn1t.h';
+ next if $f eq 'include/openssl/err.h';
+ open(IN, $f) || die "Can't open $f, $!";
+ while ( <IN> ) {
+ next unless /^#\s*define\s*(\S+)\(/;
+ my $macro = $1;
+ next if $docced{$macro} || defined $seen{$macro};
+ next if $macro =~ /i2d_/
+ || $macro =~ /d2i_/
+ || $macro =~ /DEPRECATEDIN/
+ || $macro =~ /IMPLEMENT_/
+ || $macro =~ /DECLARE_/;
+
+ # Skip macros known to be missing
+ next if $opt_v && grep( /^$macro$/, @missing);
+
+ err("$f:", "macro $macro undocumented")
+ if $opt_d || $opt_e;
+ $count++;
+ $seen{$macro} = 1;
+ }
+ close(IN);
+ }
+ err("# $count macros undocumented (count is approximate)")
+ if $count > 0;
+}
+
+# Find out what is undocumented (filtering out the known missing ones)
+# and display them.
+sub printem {
my $libname = shift;
my $numfile = shift;
+ my $missingfile = shift;
my $count = 0;
+ my %seen;
+
+ my @missing = loadmissing($missingfile) if ( $opt_v );
- foreach my $func ( &parsenum($numfile) ) {
- next if $docced{$func};
+ foreach my $func ( parsenum($numfile) ) {
+ next if $docced{$func} || defined $seen{$func};
# Skip ASN1 utilities
next if $func =~ /^ASN1_/;
- print "$libname:$func\n";
+ # Skip functions known to be missing
+ next if $opt_v && grep( /^$func$/, @missing);
+
+ err("$libname:", "function $func undocumented")
+ if $opt_d || $opt_e;
$count++;
+ $seen{$func} = 1;
}
- print "# Found $count missing from $numfile\n\n";
+ err("# $count in $numfile are not documented")
+ if $count > 0;
}
-
-# Collection of links in each POD file.
-# filename => [ "foo(1)", "bar(3)", ... ]
-my %link_collection = ();
-# Collection of names in each POD file.
-# "name(s)" => filename
-my %name_collection = ();
-
+# Collect all the names in a manpage.
sub collectnames {
my $filename = shift;
$filename =~ m|man(\d)/|;
$contents =~ /=head1 NAME([^=]*)=head1 /ms;
my $tmp = $1;
- unless (defined $tmp) {
- print "$id weird name section\n";
+ unless ( defined $tmp ) {
+ err($id, "weird name section");
return;
}
$tmp =~ tr/\n/ /;
- $tmp =~ s/-.*//g;
-
- my @names = map { s/\s+//g; $_ } split(/,/, $tmp);
- unless (grep { $simplename eq $_ } @names) {
- print "$id missing $simplename\n";
+ $tmp =~ s/ -.*//g;
+
+ my @names =
+ map { s|/|-|g; $_ } # Treat slash as dash
+ map { s/^\s+//g; s/\s+$//g; $_ } # Trim prefix and suffix blanks
+ split(/,/, $tmp);
+ unless ( grep { $simplename eq $_ } @names ) {
+ err($id, "missing $simplename");
push @names, $simplename;
}
foreach my $name (@names) {
next if $name eq "";
+ if ( $name =~ /\s/ ) {
+ err($id, "'$name' contains white space")
+ }
my $name_sec = "$name($section)";
- if (! exists $name_collection{$name_sec}) {
- $name_collection{$name_sec} = $filename;
- } else { #elsif ($filename ne $name_collection{$name_sec}) {
- print "$id $name_sec also in $name_collection{$name_sec}\n";
+ if ( !exists $name_map{$name_sec} ) {
+ $name_map{$name_sec} = $filename;
+ } elsif ( $filename eq $name_map{$name_sec} ) {
+ err($id, "$name_sec repeated in NAME section of",
+ $name_map{$name_sec});
+ } else {
+ err($id, "$name_sec also in NAME section of",
+ $name_map{$name_sec});
}
}
my @foreign_names =
map { map { s/\s+//g; $_ } split(/,/, $_) }
$contents =~ /=for\s+comment\s+foreign\s+manuals:\s*(.*)\n\n/;
- foreach (@foreign_names) {
- $name_collection{$_} = undef; # It still exists!
+ foreach ( @foreign_names ) {
+ $name_map{$_} = undef; # It still exists!
}
my @links = $contents =~ /L<
# then remove 'something'. Note that 'something'
# may contain POD codes as well...
(?:(?:[^\|]|<[^>]*>)*\|)?
- # we're only interested in referenses that have
+ # we're only interested in references that have
# a one digit section number
([^\/>\(]+\(\d\))
/gx;
- $link_collection{$filename} = [ @links ];
+ $link_map{$filename} = [ @links ];
}
+# Look for L<> ("link") references that point to files that do not exist.
sub checklinks {
- foreach my $filename (sort keys %link_collection) {
- foreach my $link (@{$link_collection{$filename}}) {
- print "${filename}:1: reference to non-existing $link\n"
- unless exists $name_collection{$link};
+ foreach my $filename (sort keys %link_map) {
+ foreach my $link (@{$link_map{$filename}}) {
+ err("${filename}:1:", "reference to non-existing $link")
+ unless exists $name_map{$link};
}
}
}
-getopts('lnshu');
+# Load the public symbol/macro names
+sub publicize {
+ foreach my $name ( parsenum('util/libcrypto.num') ) {
+ $public{$name} = 1;
+ }
+ foreach my $name ( parsenum('util/libssl.num') ) {
+ $public{$name} = 1;
+ }
+ foreach my $name ( parsenum('util/other.syms') ) {
+ $public{$name} = 1;
+ }
+}
-&help() if ( $opt_h );
+# Cipher/digests to skip if they show up as "not implemented"
+# because they are, via the "-*" construct.
+my %skips = (
+ 'aes128' => 1,
+ 'aes192' => 1,
+ 'aes256' => 1,
+ 'aria128' => 1,
+ 'aria192' => 1,
+ 'aria256' => 1,
+ 'camellia128' => 1,
+ 'camellia192' => 1,
+ 'camellia256' => 1,
+ 'des' => 1,
+ 'des3' => 1,
+ 'idea' => 1,
+ 'cipher' => 1,
+ 'digest' => 1,
+);
+
+# Check the flags of a command and see if everything is in the manpage
+sub checkflags {
+ my $cmd = shift;
+ my $doc = shift;
+ my %cmdopts;
+ my %docopts;
+ my %localskips;
+
+ # Get the list of options in the command.
+ open CFH, "./apps/openssl list --options $cmd|"
+ || die "Can list options for $cmd, $!";
+ while ( <CFH> ) {
+ chop;
+ s/ .$//;
+ $cmdopts{$_} = 1;
+ }
+ close CFH;
+
+ # Get the list of flags from the synopsis
+ open CFH, "<$doc"
+ || die "Can't open $doc, $!";
+ while ( <CFH> ) {
+ chop;
+ last if /DESCRIPTION/;
+ if ( /=for openssl ifdef (.*)/ ) {
+ foreach my $f ( split / /, $1 ) {
+ $localskips{$f} = 1;
+ }
+ next;
+ }
+ next unless /\[B<-([^ >]+)/;
+ my $opt = $1;
+ $opt = $1 if $opt =~ /I<(.*)/;
+ $docopts{$1} = 1;
+ }
+ close CFH;
-die "Need one of -l -n -s or -u flags.\n"
- unless $opt_l or $opt_n or $opt_s or $opt_u;
+ # See what's in the command not the manpage.
+ my @undocced = sort grep { !defined $docopts{$_} } keys %cmdopts;
+ foreach ( @undocced ) {
+ next if /-/; # Skip the -- end-of-flags marker
+ err("$doc: undocumented option -$_");
+ }
-if ( $opt_n or $opt_s ) {
- foreach (@ARGV ? @ARGV : glob('doc/*/*.pod')) {
- &check($_);
+ # See what's in the command not the manpage.
+ my @unimpl = sort grep { !defined $cmdopts{$_} } keys %docopts;
+ foreach ( @unimpl ) {
+ next if defined $skips{$_} || defined $localskips{$_};
+ err("$cmd documented but not implemented -$_");
}
}
+##
+## MAIN()
+## Do the work requested by the various getopt flags.
+## The flags are parsed in alphabetical order, just because we have
+## to have *some way* of listing them.
+##
+
+if ( $opt_c ) {
+ my @commands = ();
+
+ # Get list of commands.
+ open FH, "./apps/openssl list -1 -commands|"
+ || die "Can't list commands, $!";
+ while ( <FH> ) {
+ chop;
+ push @commands, $_;
+ }
+ close FH;
+
+ # See if each has a manpage.
+ foreach my $cmd ( @commands ) {
+ next if $cmd eq 'help' || $cmd eq 'exit';
+ my $doc = "doc/man1/$cmd.pod";
+ $doc = "doc/man1/openssl-$cmd.pod" if -f "doc/man1/openssl-$cmd.pod";
+ if ( ! -f "$doc" ) {
+ err("$doc does not exist");
+ } else {
+ checkflags($cmd, $doc);
+ }
+ }
+
+ # See what help is missing.
+ open FH, "./apps/openssl list --missing-help |"
+ || die "Can't list missing help, $!";
+ while ( <FH> ) {
+ chop;
+ my ($cmd, $flag) = split;
+ err("$cmd has no help for -$flag");
+ }
+ close FH;
+
+ exit $status;
+}
+
if ( $opt_l ) {
- foreach (@ARGV ? @ARGV : glob('doc/*/*.pod')) {
+ foreach ( glob('doc/*/*.pod doc/internal/*/*.pod') ) {
collectnames($_);
}
checklinks();
}
-if ( $opt_u ) {
- my %temp = &getdocced('doc/man3');
+if ( $opt_n ) {
+ publicize();
+ foreach ( @ARGV ? @ARGV : glob('doc/*/*.pod doc/internal/*/*.pod') ) {
+ check($_);
+ }
+
+ # If not given args, check that all man1 commands are named properly.
+ if ( scalar @ARGV == 0 ) {
+ foreach (glob('doc/man1/*.pod')) {
+ next if /CA.pl/ || /openssl\.pod/ || /tsget\.pod/;
+ err("$_ doesn't start with openssl-") unless /openssl-/;
+ }
+ }
+}
+
+if ( $opt_u || $opt_v) {
+ my %temp = getdocced('doc/man3');
foreach ( keys %temp ) {
$docced{$_} = $temp{$_};
}
- &printem('crypto', 'util/libcrypto.num');
- &printem('ssl', 'util/libssl.num');
+ if ( $opt_o ) {
+ printem('crypto', 'util/libcrypto.num', 'util/missingcrypto111.txt');
+ printem('ssl', 'util/libssl.num', 'util/missingssl111.txt');
+ } else {
+ printem('crypto', 'util/libcrypto.num', 'util/missingcrypto.txt');
+ printem('ssl', 'util/libssl.num', 'util/missingssl.txt');
+ }
+ checkmacros();
}
-exit;
+exit $status;