From a35beedfb3f25596b4ec9122742c1337083118f5 Mon Sep 17 00:00:00 2001 From: Brian Behlendorf Date: Wed, 30 Oct 2013 11:19:53 -0700 Subject: [PATCH] Add cstyle.pl utility and cstyle.1 man page Cstyle is the C source style checker used by Illumos. Since the original ZFS source was written using these style guidelines they must also be followed by ZoL for consistency. The checker has been added to the scripts directory and may be run on a per file basis. New patches should be careful to avoid introducing new style warnings. Additionally, the 'checkstyle' target has been added to the top level Makefile and can be used to check the entire source tree. While Zol has historically attempted to follow the SunOS style guide the lack of a rigorous style checker has allowed various warning to be introduced. Currently there are 2211 reported style violations and we want to gradually eliminate these from the tree. Note the cstyle.1 man page is provided under man/man1/cstyle.1 but since it is a developer utility it is not installed along with the other man pages. Signed-off-by: Brian Behlendorf --- Makefile.am | 3 + man/man1/Makefile.am | 2 +- man/man1/cstyle.1 | 167 ++++++++ scripts/Makefile.am | 2 +- scripts/cstyle.pl | 950 +++++++++++++++++++++++++++++++++++++++++++ 5 files changed, 1122 insertions(+), 2 deletions(-) create mode 100644 man/man1/cstyle.1 create mode 100755 scripts/cstyle.pl diff --git a/Makefile.am b/Makefile.am index 9c299a9ef..6c808b662 100644 --- a/Makefile.am +++ b/Makefile.am @@ -40,6 +40,9 @@ dist-hook: sed -i 's/Release:[[:print:]]*/Release: $(RELEASE)/' \ $(distdir)/META +checkstyle: + find ${top_srcdir} -name '*.[hc]' -type f -exec scripts/cstyle.pl {} \; + ctags: $(RM) $(top_srcdir)/tags find $(top_srcdir) -name .git -prune -o -name '*.[hc]' | xargs ctags diff --git a/man/man1/Makefile.am b/man/man1/Makefile.am index 9d4439817..476aeed54 100644 --- a/man/man1/Makefile.am +++ b/man/man1/Makefile.am @@ -1,5 +1,5 @@ man_MANS = zhack.1 zpios.1 ztest.1 -EXTRA_DIST = $(man_MANS) +EXTRA_DIST = $(man_MANS) cstyle.1 install-data-local: $(INSTALL) -d -m 0755 "$(DESTDIR)$(mandir)/man1" diff --git a/man/man1/cstyle.1 b/man/man1/cstyle.1 new file mode 100644 index 000000000..f467c55d4 --- /dev/null +++ b/man/man1/cstyle.1 @@ -0,0 +1,167 @@ +.\" Copyright 2009 Sun Microsystems, Inc. All rights reserved. +.\" Use is subject to license terms. +.\" +.\" CDDL HEADER START +.\" +.\" The contents of this file are subject to the terms of the +.\" Common Development and Distribution License (the "License"). +.\" You may not use this file except in compliance with the License. +.\" +.\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE +.\" or http://www.opensolaris.org/os/licensing. +.\" See the License for the specific language governing permissions +.\" and limitations under the License. +.\" +.\" When distributing Covered Code, include this CDDL HEADER in each +.\" file and include the License file at usr/src/OPENSOLARIS.LICENSE. +.\" If applicable, add the following below this CDDL HEADER, with the +.\" fields enclosed by brackets "[]" replaced with your own identifying +.\" information: Portions Copyright [yyyy] [name of copyright owner] +.\" +.\" CDDL HEADER END +.\" +.TH cstyle 1 "28 March 2005" +.SH NAME +.I cstyle +\- check for some common stylistic errors in C source files +.SH SYNOPSIS +\fBcstyle [-chpvCP] [-o constructs] [file...]\fP +.LP +.SH DESCRIPTION +.IX "OS-Net build tools" "cstyle" "" "\fBcstyle\fP" +.LP +.I cstyle +inspects C source files (*.c and *.h) for common sylistic errors. It +attempts to check for the cstyle documented in +\fIhttp://www.cis.upenn.edu/~lee/06cse480/data/cstyle.ms.pdf\fP. +Note that there is much in that document that +.I cannot +be checked for; just because your code is \fBcstyle(1)\fP clean does not +mean that you've followed Sun's C style. \fICaveat emptor\fP. +.LP +.SH OPTIONS +.LP +The following options are supported: +.TP 4 +.B \-c +Check continuation line indentation inside of functions. Sun's C style +states that all statements must be indented to an appropriate tab stop, +and any continuation lines after them must be indented \fIexactly\fP four +spaces from the start line. This option enables a series of checks +designed to find contination line problems within functions only. The +checks have some limitations; see CONTINUATION CHECKING, below. +.LP +.TP 4 +.B \-h +Performs heuristic checks that are sometimes wrong. Not generally used. +.LP +.TP 4 +.B \-p +Performs some of the more picky checks. Includes ANSI #else and #endif +rules, and tries to detect spaces after casts. Used as part of the +putback checks. +.LP +.TP 4 +.B \-v +Verbose output; includes the text of the line of error, and, for +\fB-c\fP, the first statement in the current continuation block. +.LP +.TP 4 +.B \-C +Ignore errors in header comments (i.e. block comments starting in the +first column). Not generally used. +.LP +.TP 4 +.B \-P +Check for use of non-POSIX types. Historically, types like "u_int" and +"u_long" were used, but they are now deprecated in favor of the POSIX +types uint_t, ulong_t, etc. This detects any use of the deprecated +types. Used as part of the putback checks. +.LP +.TP 4 +.B \-o \fIconstructs\fP +Allow a comma-seperated list of additional constructs. Available +constructs include: +.LP +.TP 10 +.B doxygen +Allow doxygen-style block comments (\fB/**\fP and \fB/*!\fP) +.LP +.TP 10 +.B splint +Allow splint-style lint comments (\fB/*@...@*/\fP) +.LP +.SH NOTES +.LP +The cstyle rule for the OS/Net consolidation is that all new files must +be \fB-pP\fP clean. For existing files, the following invocations are +run against both the old and new files: +.LP +.TP 4 +\fBcstyle file\fB +.LP +.TP 4 +\fBcstyle -p file\fB +.LP +.TP 4 +\fBcstyle -pP file\fB +.LP +If the old file gave no errors for one of the invocations, the new file +must also give no errors. This way, files can only become more clean. +.LP +.SH CONTINUATION CHECKING +.LP +The continuation checker is a resonably simple state machine that knows +something about how C is layed out, and can match parenthesis, etc. over +multiple lines. It does have some limitations: +.LP +.TP 4 +.B 1. +Preprocessor macros which cause unmatched parenthesis will confuse the +checker for that line. To fix this, you'll need to make sure that each +branch of the #if statement has balanced parenthesis. +.LP +.TP 4 +.B 2. +Some \fBcpp\fP macros do not require ;s after them. Any such macros +*must* be ALL_CAPS; any lower case letters will cause bad output. +.LP +The bad output will generally be corrected after the next \fB;\fP, +\fB{\fP, or \fB}\fP. +.LP +Some continuation error messages deserve some additional explanation +.LP +.TP 4 +.B +multiple statements continued over multiple lines +A multi-line statement which is not broken at statement +boundries. For example: +.RS 4 +.HP 4 +if (this_is_a_long_variable == another_variable) a = +.br +b + c; +.LP +Will trigger this error. Instead, do: +.HP 8 +if (this_is_a_long_variable == another_variable) +.br +a = b + c; +.RE +.LP +.TP 4 +.B +empty if/for/while body not on its own line +For visibility, empty bodies for if, for, and while statements should be +on their own line. For example: +.RS 4 +.HP 4 +while (do_something(&x) == 0); +.LP +Will trigger this error. Instead, do: +.HP 8 +while (do_something(&x) == 0) +.br +; +.RE + diff --git a/scripts/Makefile.am b/scripts/Makefile.am index 39d94ee82..bc8fe99cf 100644 --- a/scripts/Makefile.am +++ b/scripts/Makefile.am @@ -1,6 +1,6 @@ SUBDIRS = zpool-config zpios-test zpios-profile -EXTRA_DIST = dkms.mkconf dkms.postinst kmodtool zfs2zol-patch.sed +EXTRA_DIST = dkms.mkconf dkms.postinst kmodtool zfs2zol-patch.sed cstyle.pl pkgdatadir = $(datadir)/@PACKAGE@ dist_pkgdata_SCRIPTS = \ diff --git a/scripts/cstyle.pl b/scripts/cstyle.pl new file mode 100755 index 000000000..9f67b1d2e --- /dev/null +++ b/scripts/cstyle.pl @@ -0,0 +1,950 @@ +#!/usr/bin/perl -w +# +# CDDL HEADER START +# +# The contents of this file are subject to the terms of the +# Common Development and Distribution License (the "License"). +# You may not use this file except in compliance with the License. +# +# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE +# or http://www.opensolaris.org/os/licensing. +# See the License for the specific language governing permissions +# and limitations under the License. +# +# When distributing Covered Code, include this CDDL HEADER in each +# file and include the License file at usr/src/OPENSOLARIS.LICENSE. +# If applicable, add the following below this CDDL HEADER, with the +# fields enclosed by brackets "[]" replaced with your own identifying +# information: Portions Copyright [yyyy] [name of copyright owner] +# +# CDDL HEADER END +# +# +# Copyright 2008 Sun Microsystems, Inc. All rights reserved. +# Use is subject to license terms. +# +# @(#)cstyle 1.58 98/09/09 (from shannon) +#ident "%Z%%M% %I% %E% SMI" +# +# cstyle - check for some common stylistic errors. +# +# cstyle is a sort of "lint" for C coding style. +# It attempts to check for the style used in the +# kernel, sometimes known as "Bill Joy Normal Form". +# +# There's a lot this can't check for, like proper indentation +# of code blocks. There's also a lot more this could check for. +# +# A note to the non perl literate: +# +# perl regular expressions are pretty much like egrep +# regular expressions, with the following special symbols +# +# \s any space character +# \S any non-space character +# \w any "word" character [a-zA-Z0-9_] +# \W any non-word character +# \d a digit [0-9] +# \D a non-digit +# \b word boundary (between \w and \W) +# \B non-word boundary +# + +require 5.0; +use IO::File; +use Getopt::Std; +use strict; + +my $usage = +"usage: cstyle [-chpvCP] [-o constructs] file ... + -c check continuation indentation inside functions + -h perform heuristic checks that are sometimes wrong + -p perform some of the more picky checks + -v verbose + -C don't check anything in header block comments + -P check for use of non-POSIX types + -o constructs + allow a comma-seperated list of optional constructs: + doxygen allow doxygen-style block comments (/** /*!) + splint allow splint-style lint comments (/*@ ... @*/) +"; + +my %opts; + +if (!getopts("cho:pvCP", \%opts)) { + print $usage; + exit 2; +} + +my $check_continuation = $opts{'c'}; +my $heuristic = $opts{'h'}; +my $picky = $opts{'p'}; +my $verbose = $opts{'v'}; +my $ignore_hdr_comment = $opts{'C'}; +my $check_posix_types = $opts{'P'}; + +my $doxygen_comments = 0; +my $splint_comments = 0; + +if (defined($opts{'o'})) { + for my $x (split /,/, $opts{'o'}) { + if ($x eq "doxygen") { + $doxygen_comments = 1; + } elsif ($x eq "splint") { + $splint_comments = 1; + } else { + print "cstyle: unrecognized construct \"$x\"\n"; + print $usage; + exit 2; + } + } +} + +my ($filename, $line, $prev); # shared globals + +my $fmt; +my $hdr_comment_start; + +if ($verbose) { + $fmt = "%s: %d: %s\n%s\n"; +} else { + $fmt = "%s: %d: %s\n"; +} + +if ($doxygen_comments) { + # doxygen comments look like "/*!" or "/**"; allow them. + $hdr_comment_start = qr/^\s*\/\*[\!\*]?$/; +} else { + $hdr_comment_start = qr/^\s*\/\*$/; +} + +# Note, following must be in single quotes so that \s and \w work right. +my $typename = '(int|char|short|long|unsigned|float|double' . + '|\w+_t|struct\s+\w+|union\s+\w+|FILE)'; + +# mapping of old types to POSIX compatible types +my %old2posix = ( + 'unchar' => 'uchar_t', + 'ushort' => 'ushort_t', + 'uint' => 'uint_t', + 'ulong' => 'ulong_t', + 'u_int' => 'uint_t', + 'u_short' => 'ushort_t', + 'u_long' => 'ulong_t', + 'u_char' => 'uchar_t', + 'quad' => 'quad_t' +); + +my $lint_re = qr/\/\*(?: + ARGSUSED[0-9]*|NOTREACHED|LINTLIBRARY|VARARGS[0-9]*| + CONSTCOND|CONSTANTCOND|CONSTANTCONDITION|EMPTY| + FALLTHRU|FALLTHROUGH|LINTED.*?|PRINTFLIKE[0-9]*| + PROTOLIB[0-9]*|SCANFLIKE[0-9]*|CSTYLED.*? + )\*\//x; + +my $splint_re = qr/\/\*@.*?@\*\//x; + +my $warlock_re = qr/\/\*\s*(?: + VARIABLES\ PROTECTED\ BY| + MEMBERS\ PROTECTED\ BY| + ALL\ MEMBERS\ PROTECTED\ BY| + READ-ONLY\ VARIABLES:| + READ-ONLY\ MEMBERS:| + VARIABLES\ READABLE\ WITHOUT\ LOCK:| + MEMBERS\ READABLE\ WITHOUT\ LOCK:| + LOCKS\ COVERED\ BY| + LOCK\ UNNEEDED\ BECAUSE| + LOCK\ NEEDED:| + LOCK\ HELD\ ON\ ENTRY:| + READ\ LOCK\ HELD\ ON\ ENTRY:| + WRITE\ LOCK\ HELD\ ON\ ENTRY:| + LOCK\ ACQUIRED\ AS\ SIDE\ EFFECT:| + READ\ LOCK\ ACQUIRED\ AS\ SIDE\ EFFECT:| + WRITE\ LOCK\ ACQUIRED\ AS\ SIDE\ EFFECT:| + LOCK\ RELEASED\ AS\ SIDE\ EFFECT:| + LOCK\ UPGRADED\ AS\ SIDE\ EFFECT:| + LOCK\ DOWNGRADED\ AS\ SIDE\ EFFECT:| + FUNCTIONS\ CALLED\ THROUGH\ POINTER| + FUNCTIONS\ CALLED\ THROUGH\ MEMBER| + LOCK\ ORDER: + )/x; + +my $err_stat = 0; # exit status + +if ($#ARGV >= 0) { + foreach my $arg (@ARGV) { + my $fh = new IO::File $arg, "r"; + if (!defined($fh)) { + printf "%s: can not open\n", $arg; + } else { + &cstyle($arg, $fh); + close $fh; + } + } +} else { + &cstyle("", *STDIN); +} +exit $err_stat; + +my $no_errs = 0; # set for CSTYLED-protected lines + +sub err($) { + my ($error) = @_; + unless ($no_errs) { + printf $fmt, $filename, $., $error, $line; + $err_stat = 1; + } +} + +sub err_prefix($$) { + my ($prevline, $error) = @_; + my $out = $prevline."\n".$line; + unless ($no_errs) { + printf $fmt, $filename, $., $error, $out; + $err_stat = 1; + } +} + +sub err_prev($) { + my ($error) = @_; + unless ($no_errs) { + printf $fmt, $filename, $. - 1, $error, $prev; + $err_stat = 1; + } +} + +sub cstyle($$) { + +my ($fn, $filehandle) = @_; +$filename = $fn; # share it globally + +my $in_cpp = 0; +my $next_in_cpp = 0; + +my $in_comment = 0; +my $in_header_comment = 0; +my $comment_done = 0; +my $in_warlock_comment = 0; +my $in_function = 0; +my $in_function_header = 0; +my $in_declaration = 0; +my $note_level = 0; +my $nextok = 0; +my $nocheck = 0; + +my $in_string = 0; + +my ($okmsg, $comment_prefix); + +$line = ''; +$prev = ''; +reset_indent(); + +line: while (<$filehandle>) { + s/\r?\n$//; # strip return and newline + + # save the original line, then remove all text from within + # double or single quotes, we do not want to check such text. + + $line = $_; + + # + # C allows strings to be continued with a backslash at the end of + # the line. We translate that into a quoted string on the previous + # line followed by an initial quote on the next line. + # + # (we assume that no-one will use backslash-continuation with character + # constants) + # + $_ = '"' . $_ if ($in_string && !$nocheck && !$in_comment); + + # + # normal strings and characters + # + s/'([^\\']|\\[^xX0]|\\0[0-9]*|\\[xX][0-9a-fA-F]*)'/''/g; + s/"([^\\"]|\\.)*"/\"\"/g; + + # + # detect string continuation + # + if ($nocheck || $in_comment) { + $in_string = 0; + } else { + # + # Now that all full strings are replaced with "", we check + # for unfinished strings continuing onto the next line. + # + $in_string = + (s/([^"](?:"")*)"([^\\"]|\\.)*\\$/$1""/ || + s/^("")*"([^\\"]|\\.)*\\$/""/); + } + + # + # figure out if we are in a cpp directive + # + $in_cpp = $next_in_cpp || /^\s*#/; # continued or started + $next_in_cpp = $in_cpp && /\\$/; # only if continued + + # strip off trailing backslashes, which appear in long macros + s/\s*\\$//; + + # an /* END CSTYLED */ comment ends a no-check block. + if ($nocheck) { + if (/\/\* *END *CSTYLED *\*\//) { + $nocheck = 0; + } else { + reset_indent(); + next line; + } + } + + # a /*CSTYLED*/ comment indicates that the next line is ok. + if ($nextok) { + if ($okmsg) { + err($okmsg); + } + $nextok = 0; + $okmsg = 0; + if (/\/\* *CSTYLED.*\*\//) { + /^.*\/\* *CSTYLED *(.*) *\*\/.*$/; + $okmsg = $1; + $nextok = 1; + } + $no_errs = 1; + } elsif ($no_errs) { + $no_errs = 0; + } + + # check length of line. + # first, a quick check to see if there is any chance of being too long. + if (($line =~ tr/\t/\t/) * 7 + length($line) > 80) { + # yes, there is a chance. + # replace tabs with spaces and check again. + my $eline = $line; + 1 while $eline =~ + s/\t+/' ' x (length($&) * 8 - length($`) % 8)/e; + if (length($eline) > 80) { + err("line > 80 characters"); + } + } + + # ignore NOTE(...) annotations (assumes NOTE is on lines by itself). + if ($note_level || /\b_?NOTE\s*\(/) { # if in NOTE or this is NOTE + s/[^()]//g; # eliminate all non-parens + $note_level += s/\(//g - length; # update paren nest level + next; + } + + # a /* BEGIN CSTYLED */ comment starts a no-check block. + if (/\/\* *BEGIN *CSTYLED *\*\//) { + $nocheck = 1; + } + + # a /*CSTYLED*/ comment indicates that the next line is ok. + if (/\/\* *CSTYLED.*\*\//) { + /^.*\/\* *CSTYLED *(.*) *\*\/.*$/; + $okmsg = $1; + $nextok = 1; + } + if (/\/\/ *CSTYLED/) { + /^.*\/\/ *CSTYLED *(.*)$/; + $okmsg = $1; + $nextok = 1; + } + + # universal checks; apply to everything + if (/\t +\t/) { + err("spaces between tabs"); + } + if (/ \t+ /) { + err("tabs between spaces"); + } + if (/\s$/) { + err("space or tab at end of line"); + } + if (/[^ \t(]\/\*/ && !/\w\(\/\*.*\*\/\);/) { + err("comment preceded by non-blank"); + } + + # is this the beginning or ending of a function? + # (not if "struct foo\n{\n") + if (/^{$/ && $prev =~ /\)\s*(const\s*)?(\/\*.*\*\/\s*)?\\?$/) { + $in_function = 1; + $in_declaration = 1; + $in_function_header = 0; + $prev = $line; + next line; + } + if (/^}\s*(\/\*.*\*\/\s*)*$/) { + if ($prev =~ /^\s*return\s*;/) { + err_prev("unneeded return at end of function"); + } + $in_function = 0; + reset_indent(); # we don't check between functions + $prev = $line; + next line; + } + if (/^\w*\($/) { + $in_function_header = 1; + } + + if ($in_warlock_comment && /\*\//) { + $in_warlock_comment = 0; + $prev = $line; + next line; + } + + # a blank line terminates the declarations within a function. + # XXX - but still a problem in sub-blocks. + if ($in_declaration && /^$/) { + $in_declaration = 0; + } + + if ($comment_done) { + $in_comment = 0; + $in_header_comment = 0; + $comment_done = 0; + } + # does this looks like the start of a block comment? + if (/$hdr_comment_start/) { + if (!/^\t*\/\*/) { + err("block comment not indented by tabs"); + } + $in_comment = 1; + /^(\s*)\//; + $comment_prefix = $1; + if ($comment_prefix eq "") { + $in_header_comment = 1; + } + $prev = $line; + next line; + } + # are we still in the block comment? + if ($in_comment) { + if (/^$comment_prefix \*\/$/) { + $comment_done = 1; + } elsif (/\*\//) { + $comment_done = 1; + err("improper block comment close") + unless ($ignore_hdr_comment && $in_header_comment); + } elsif (!/^$comment_prefix \*[ \t]/ && + !/^$comment_prefix \*$/) { + err("improper block comment") + unless ($ignore_hdr_comment && $in_header_comment); + } + } + + if ($in_header_comment && $ignore_hdr_comment) { + $prev = $line; + next line; + } + + # check for errors that might occur in comments and in code. + + # allow spaces to be used to draw pictures in header comments. + if (/[^ ] / && !/".* .*"/ && !$in_header_comment) { + err("spaces instead of tabs"); + } + if (/^ / && !/^ \*[ \t\/]/ && !/^ \*$/ && + (!/^ \w/ || $in_function != 0)) { + err("indent by spaces instead of tabs"); + } + if (/^\t+ [^ \t\*]/ || /^\t+ \S/ || /^\t+ \S/) { + err("continuation line not indented by 4 spaces"); + } + if (/$warlock_re/ && !/\*\//) { + $in_warlock_comment = 1; + $prev = $line; + next line; + } + if (/^\s*\/\*./ && !/^\s*\/\*.*\*\// && !/$hdr_comment_start/) { + err("improper first line of block comment"); + } + + if ($in_comment) { # still in comment, don't do further checks + $prev = $line; + next line; + } + + if ((/[^(]\/\*\S/ || /^\/\*\S/) && + !(/$lint_re/ || ($splint_comments && /$splint_re/))) { + err("missing blank after open comment"); + } + if (/\S\*\/[^)]|\S\*\/$/ && + !(/$lint_re/ || ($splint_comments && /$splint_re/))) { + err("missing blank before close comment"); + } + if (/\/\/\S/) { # C++ comments + err("missing blank after start comment"); + } + # check for unterminated single line comments, but allow them when + # they are used to comment out the argument list of a function + # declaration. + if (/\S.*\/\*/ && !/\S.*\/\*.*\*\// && !/\(\/\*/) { + err("unterminated single line comment"); + } + + if (/^(#else|#endif|#include)(.*)$/) { + $prev = $line; + if ($picky) { + my $directive = $1; + my $clause = $2; + # Enforce ANSI rules for #else and #endif: no noncomment + # identifiers are allowed after #endif or #else. Allow + # C++ comments since they seem to be a fact of life. + if ((($1 eq "#endif") || ($1 eq "#else")) && + ($clause ne "") && + (!($clause =~ /^\s+\/\*.*\*\/$/)) && + (!($clause =~ /^\s+\/\/.*$/))) { + err("non-comment text following " . + "$directive (or malformed $directive " . + "directive)"); + } + } + next line; + } + + # + # delete any comments and check everything else. Note that + # ".*?" is a non-greedy match, so that we don't get confused by + # multiple comments on the same line. + # + s/\/\*.*?\*\///g; + s/\/\/.*$//; # C++ comments + + # delete any trailing whitespace; we have already checked for that. + s/\s*$//; + + # following checks do not apply to text in comments. + + if (/[^<>\s][!<>=]=/ || /[^<>][!<>=]=[^\s,]/ || + (/[^->]>[^,=>\s]/ && !/[^->]>$/) || + (/[^<]<[^,=<\s]/ && !/[^<]<$/) || + /[^<\s]<[^<]/ || /[^->\s]>[^>]/) { + err("missing space around relational operator"); + } + if (/\S>>=/ || /\S<<=/ || />>=\S/ || /<<=\S/ || /\S[-+*\/&|^%]=/ || + (/[^-+*\/&|^%!<>=\s]=[^=]/ && !/[^-+*\/&|^%!<>=\s]=$/) || + (/[^!<>=]=[^=\s]/ && !/[^!<>=]=$/)) { + # XXX - should only check this for C++ code + # XXX - there are probably other forms that should be allowed + if (!/\soperator=/) { + err("missing space around assignment operator"); + } + } + if (/[,;]\S/ && !/\bfor \(;;\)/) { + err("comma or semicolon followed by non-blank"); + } + # allow "for" statements to have empty "while" clauses + if (/\s[,;]/ && !/^[\t]+;$/ && !/^\s*for \([^;]*; ;[^;]*\)/) { + err("comma or semicolon preceded by blank"); + } + if (/^\s*(&&|\|\|)/) { + err("improper boolean continuation"); + } + if (/\S *(&&|\|\|)/ || /(&&|\|\|) *\S/) { + err("more than one space around boolean operator"); + } + if (/\b(for|if|while|switch|sizeof|return|case)\(/) { + err("missing space between keyword and paren"); + } + if (/(\b(for|if|while|switch|return)\b.*){2,}/ && !/^#define/) { + # multiple "case" and "sizeof" allowed + err("more than one keyword on line"); + } + if (/\b(for|if|while|switch|sizeof|return|case)\s\s+\(/ && + !/^#if\s+\(/) { + err("extra space between keyword and paren"); + } + # try to detect "func (x)" but not "if (x)" or + # "#define foo (x)" or "int (*func)();" + if (/\w\s\(/) { + my $s = $_; + # strip off all keywords on the line + s/\b(for|if|while|switch|return|case|sizeof)\s\(/XXX(/g; + s/#elif\s\(/XXX(/g; + s/^#define\s+\w+\s+\(/XXX(/; + # do not match things like "void (*f)();" + # or "typedef void (func_t)();" + s/\w\s\(+\*/XXX(*/g; + s/\b($typename|void)\s+\(+/XXX(/og; + if (/\w\s\(/) { + err("extra space between function name and left paren"); + } + $_ = $s; + } + # try to detect "int foo(x)", but not "extern int foo(x);" + # XXX - this still trips over too many legitimate things, + # like "int foo(x,\n\ty);" +# if (/^(\w+(\s|\*)+)+\w+\(/ && !/\)[;,](\s|)*$/ && +# !/^(extern|static)\b/) { +# err("return type of function not on separate line"); +# } + # this is a close approximation + if (/^(\w+(\s|\*)+)+\w+\(.*\)(\s|)*$/ && + !/^(extern|static)\b/) { + err("return type of function not on separate line"); + } + if (/^#define /) { + err("#define followed by space instead of tab"); + } + if (/^\s*return\W[^;]*;/ && !/^\s*return\s*\(.*\);/) { + err("unparenthesized return expression"); + } + if (/\bsizeof\b/ && !/\bsizeof\s*\(.*\)/) { + err("unparenthesized sizeof expression"); + } + if (/\(\s/) { + err("whitespace after left paren"); + } + # allow "for" statements to have empty "continue" clauses + if (/\s\)/ && !/^\s*for \([^;]*;[^;]*; \)/) { + err("whitespace before right paren"); + } + if (/^\s*\(void\)[^ ]/) { + err("missing space after (void) cast"); + } + if (/\S{/ && !/{{/) { + err("missing space before left brace"); + } + if ($in_function && /^\s+{/ && + ($prev =~ /\)\s*$/ || $prev =~ /\bstruct\s+\w+$/)) { + err("left brace starting a line"); + } + if (/}(else|while)/) { + err("missing space after right brace"); + } + if (/}\s\s+(else|while)/) { + err("extra space after right brace"); + } + if (/\b_VOID\b|\bVOID\b|\bSTATIC\b/) { + err("obsolete use of VOID or STATIC"); + } + if (/\b$typename\*/o) { + err("missing space between type name and *"); + } + if (/^\s+#/) { + err("preprocessor statement not in column 1"); + } + if (/^#\s/) { + err("blank after preprocessor #"); + } + if (/!\s*(strcmp|strncmp|bcmp)\s*\(/) { + err("don't use boolean ! with comparison functions"); + } + + # + # We completely ignore, for purposes of indentation: + # * lines outside of functions + # * preprocessor lines + # + if ($check_continuation && $in_function && !$in_cpp) { + process_indent($_); + } + if ($picky) { + # try to detect spaces after casts, but allow (e.g.) + # "sizeof (int) + 1", "void (*funcptr)(int) = foo;", and + # "int foo(int) __NORETURN;" + if ((/^\($typename( \*+)?\)\s/o || + /\W\($typename( \*+)?\)\s/o) && + !/sizeof\s*\($typename( \*)?\)\s/o && + !/\($typename( \*+)?\)\s+=[^=]/o) { + err("space after cast"); + } + if (/\b$typename\s*\*\s/o && + !/\b$typename\s*\*\s+const\b/o) { + err("unary * followed by space"); + } + } + if ($check_posix_types) { + # try to detect old non-POSIX types. + # POSIX requires all non-standard typedefs to end in _t, + # but historically these have been used. + if (/\b(unchar|ushort|uint|ulong|u_int|u_short|u_long|u_char|quad)\b/) { + err("non-POSIX typedef $1 used: use $old2posix{$1} instead"); + } + } + if ($heuristic) { + # cannot check this everywhere due to "struct {\n...\n} foo;" + if ($in_function && !$in_declaration && + /}./ && !/}\s+=/ && !/{.*}[;,]$/ && !/}(\s|)*$/ && + !/} (else|while)/ && !/}}/) { + err("possible bad text following right brace"); + } + # cannot check this because sub-blocks in + # the middle of code are ok + if ($in_function && /^\s+{/) { + err("possible left brace starting a line"); + } + } + if (/^\s*else\W/) { + if ($prev =~ /^\s*}$/) { + err_prefix($prev, + "else and right brace should be on same line"); + } + } + $prev = $line; +} + +if ($prev eq "") { + err("last line in file is blank"); +} + +} + +# +# Continuation-line checking +# +# The rest of this file contains the code for the continuation checking +# engine. It's a pretty simple state machine which tracks the expression +# depth (unmatched '('s and '['s). +# +# Keep in mind that the argument to process_indent() has already been heavily +# processed; all comments have been replaced by control-A, and the contents of +# strings and character constants have been elided. +# + +my $cont_in; # currently inside of a continuation +my $cont_off; # skipping an initializer or definition +my $cont_noerr; # suppress cascading errors +my $cont_start; # the line being continued +my $cont_base; # the base indentation +my $cont_first; # this is the first line of a statement +my $cont_multiseg; # this continuation has multiple segments + +my $cont_special; # this is a C statement (if, for, etc.) +my $cont_macro; # this is a macro +my $cont_case; # this is a multi-line case + +my @cont_paren; # the stack of unmatched ( and [s we've seen + +sub +reset_indent() +{ + $cont_in = 0; + $cont_off = 0; +} + +sub +delabel($) +{ + # + # replace labels with tabs. Note that there may be multiple + # labels on a line. + # + local $_ = $_[0]; + + while (/^(\t*)( *(?:(?:\w+\s*)|(?:case\b[^:]*)): *)(.*)$/) { + my ($pre_tabs, $label, $rest) = ($1, $2, $3); + $_ = $pre_tabs; + while ($label =~ s/^([^\t]*)(\t+)//) { + $_ .= "\t" x (length($2) + length($1) / 8); + } + $_ .= ("\t" x (length($label) / 8)).$rest; + } + + return ($_); +} + +sub +process_indent($) +{ + require strict; + local $_ = $_[0]; # preserve the global $_ + + s///g; # No comments + s/\s+$//; # Strip trailing whitespace + + return if (/^$/); # skip empty lines + + # regexps used below; keywords taking (), macros, and continued cases + my $special = '(?:(?:\}\s*)?else\s+)?(?:if|for|while|switch)\b'; + my $macro = '[A-Z_][A-Z_0-9]*\('; + my $case = 'case\b[^:]*$'; + + # skip over enumerations, array definitions, initializers, etc. + if ($cont_off <= 0 && !/^\s*$special/ && + (/(?:(?:\b(?:enum|struct|union)\s*[^\{]*)|(?:\s+=\s*)){/ || + (/^\s*{/ && $prev =~ /=\s*(?:\/\*.*\*\/\s*)*$/))) { + $cont_in = 0; + $cont_off = tr/{/{/ - tr/}/}/; + return; + } + if ($cont_off) { + $cont_off += tr/{/{/ - tr/}/}/; + return; + } + + if (!$cont_in) { + $cont_start = $line; + + if (/^\t* /) { + err("non-continuation indented 4 spaces"); + $cont_noerr = 1; # stop reporting + } + $_ = delabel($_); # replace labels with tabs + + # check if the statement is complete + return if (/^\s*\}?$/); + return if (/^\s*\}?\s*else\s*\{?$/); + return if (/^\s*do\s*\{?$/); + return if (/{$/); + return if (/}[,;]?$/); + + # Allow macros on their own lines + return if (/^\s*[A-Z_][A-Z_0-9]*$/); + + # cases we don't deal with, generally non-kosher + if (/{/) { + err("stuff after {"); + return; + } + + # Get the base line, and set up the state machine + /^(\t*)/; + $cont_base = $1; + $cont_in = 1; + @cont_paren = (); + $cont_first = 1; + $cont_multiseg = 0; + + # certain things need special processing + $cont_special = /^\s*$special/? 1 : 0; + $cont_macro = /^\s*$macro/? 1 : 0; + $cont_case = /^\s*$case/? 1 : 0; + } else { + $cont_first = 0; + + # Strings may be pulled back to an earlier (half-)tabstop + unless ($cont_noerr || /^$cont_base / || + (/^\t*(?: )?(?:gettext\()?\"/ && !/^$cont_base\t/)) { + err_prefix($cont_start, + "continuation should be indented 4 spaces"); + } + } + + my $rest = $_; # keeps the remainder of the line + + # + # The split matches 0 characters, so that each 'special' character + # is processed separately. Parens and brackets are pushed and + # popped off the @cont_paren stack. For normal processing, we wait + # until a ; or { terminates the statement. "special" processing + # (if/for/while/switch) is allowed to stop when the stack empties, + # as is macro processing. Case statements are terminated with a : + # and an empty paren stack. + # + foreach $_ (split /[^\(\)\[\]\{\}\;\:]*/) { + next if (length($_) == 0); + + # rest contains the remainder of the line + my $rxp = "[^\Q$_\E]*\Q$_\E"; + $rest =~ s/^$rxp//; + + if (/\(/ || /\[/) { + push @cont_paren, $_; + } elsif (/\)/ || /\]/) { + my $cur = $_; + tr/\)\]/\(\[/; + + my $old = (pop @cont_paren); + if (!defined($old)) { + err("unexpected '$cur'"); + $cont_in = 0; + last; + } elsif ($old ne $_) { + err("'$cur' mismatched with '$old'"); + $cont_in = 0; + last; + } + + # + # If the stack is now empty, do special processing + # for if/for/while/switch and macro statements. + # + next if (@cont_paren != 0); + if ($cont_special) { + if ($rest =~ /^\s*{?$/) { + $cont_in = 0; + last; + } + if ($rest =~ /^\s*;$/) { + err("empty if/for/while body ". + "not on its own line"); + $cont_in = 0; + last; + } + if (!$cont_first && $cont_multiseg == 1) { + err_prefix($cont_start, + "multiple statements continued ". + "over multiple lines"); + $cont_multiseg = 2; + } elsif ($cont_multiseg == 0) { + $cont_multiseg = 1; + } + # We've finished this section, start + # processing the next. + goto section_ended; + } + if ($cont_macro) { + if ($rest =~ /^$/) { + $cont_in = 0; + last; + } + } + } elsif (/\;/) { + if ($cont_case) { + err("unexpected ;"); + } elsif (!$cont_special) { + err("unexpected ;") if (@cont_paren != 0); + if (!$cont_first && $cont_multiseg == 1) { + err_prefix($cont_start, + "multiple statements continued ". + "over multiple lines"); + $cont_multiseg = 2; + } elsif ($cont_multiseg == 0) { + $cont_multiseg = 1; + } + if ($rest =~ /^$/) { + $cont_in = 0; + last; + } + if ($rest =~ /^\s*special/) { + err("if/for/while/switch not started ". + "on its own line"); + } + goto section_ended; + } + } elsif (/\{/) { + err("{ while in parens/brackets") if (@cont_paren != 0); + err("stuff after {") if ($rest =~ /[^\s}]/); + $cont_in = 0; + last; + } elsif (/\}/) { + err("} while in parens/brackets") if (@cont_paren != 0); + if (!$cont_special && $rest !~ /^\s*(while|else)\b/) { + if ($rest =~ /^$/) { + err("unexpected }"); + } else { + err("stuff after }"); + } + $cont_in = 0; + last; + } + } elsif (/\:/ && $cont_case && @cont_paren == 0) { + err("stuff after multi-line case") if ($rest !~ /$^/); + $cont_in = 0; + last; + } + next; +section_ended: + # End of a statement or if/while/for loop. Reset + # cont_special and cont_macro based on the rest of the + # line. + $cont_special = ($rest =~ /^\s*$special/)? 1 : 0; + $cont_macro = ($rest =~ /^\s*$macro/)? 1 : 0; + $cont_case = 0; + next; + } + $cont_noerr = 0 if (!$cont_in); +} -- 2.40.0