xref: /illumos-gate/usr/src/tools/scripts/jstyle.pl (revision edb348833aaacfa1176e502ad38875fd0b2717ab)
1#!/usr/bin/perl -w
2#
3# CDDL HEADER START
4#
5# The contents of this file are subject to the terms of the
6# Common Development and Distribution License (the "License").
7# You may not use this file except in compliance with the License.
8#
9# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
10# or http://www.opensolaris.org/os/licensing.
11# See the License for the specific language governing permissions
12# and limitations under the License.
13#
14# When distributing Covered Code, include this CDDL HEADER in each
15# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
16# If applicable, add the following below this CDDL HEADER, with the
17# fields enclosed by brackets "[]" replaced with your own identifying
18# information: Portions Copyright [yyyy] [name of copyright owner]
19#
20# CDDL HEADER END
21#
22
23#
24# ident	"%Z%%M%	%I%	%E% SMI"
25#
26# Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
27# Use is subject to license terms.
28#
29# jstyle - check for some common stylistic errors.
30#
31
32require 5.006;
33use Getopt::Std;
34use strict;
35
36my $usage =
37"usage: jstyle [-c] [-h] [-p] [-t] [-v] [-C] file ...
38	-c	check continuation line indenting
39	-h	perform heuristic checks that are sometimes wrong
40	-p	perform some of the more picky checks
41	-t	insist on indenting by tabs
42	-v	verbose
43	-C	don't check anything in header block comments
44";
45
46my %opts;
47
48# Keep -s, as it's been around for a while.  It just doesn't do anything
49# anymore.
50if (!getopts("chpstvC", \%opts)) {
51	print $usage;
52	exit 2;
53}
54
55my $check_continuation = $opts{'c'};
56my $heuristic = $opts{'h'};
57my $picky = $opts{'p'};
58my $tabs = $opts{'t'};
59my $verbose = $opts{'v'};
60my $ignore_hdr_comment = $opts{'C'};
61
62my ($filename, $line, $prev);
63my $err_stat = 0;		# Exit status
64
65my $fmt;
66
67if ($verbose) {
68	$fmt = "%s: %d: %s\n%s\n";
69} else {
70	$fmt = "%s: %d: %s\n";
71}
72
73# Note, following must be in single quotes so that \s and \w work right.
74my $typename = '(int|char|boolean|byte|short|long|float|double)';
75my $keywords = '(for|if|while|switch|return|catch|synchronized|throw|assert)';
76# See perlre(1) for the meaning of (??{ ... })
77my $annotations = ""; $annotations = qr/@\w+\((?:(?>[^()]+)|(??{ $annotations }))*\)/;
78my $generics = ""; $generics = qr/<(([\s\w,.?[\]]| & )+|(??{ $generics }))*>/;
79my $relationalops = qr/>=|<=|<|>|!=|==/;
80my $shiftops = qr/<<<|>>>|<<|>>/;
81my $shiftassignmentops = qr/[<>]{2,3}=/;
82my $assignmentops = qr/[-+\/*|&^%]?=/;
83# These need to be in decreasing order of length
84my $allops = qr/$shiftassignmentops|$shiftops|$relationalops|$assignmentops/;
85
86if ($#ARGV >= 0) {
87	foreach my $arg (@ARGV) {
88		if (!open(STDIN, $arg)) {
89			printf "%s: can not open\n", $arg;
90		} else {
91			&jstyle($arg);
92			close STDIN;
93		}
94	}
95} else {
96	&jstyle("<stdin>");
97}
98exit $err_stat;
99
100sub err($) {
101	printf $fmt, $filename, $., $_[0], $line;
102	$err_stat = 1;
103}
104
105sub jstyle($) {
106
107my $in_comment = 0;
108my $in_header_comment = 0;
109my $in_continuation = 0;
110my $in_class = 0;
111my $in_declaration = 0;
112my $nextok = 0;
113my $nocheck = 0;
114my $expect_continuation = 0;
115my $continuation_indent;
116my $okmsg;
117my $comment_prefix;
118my $comment_done;
119my $cpp_comment;
120
121$filename = $_[0];
122
123line: while (<STDIN>) {
124	s/\r?\n$//;	# strip return and newline
125
126	# save the original line, then remove all text from within
127	# double or single quotes, we do not want to check such text.
128
129	$line = $_;
130	s/"[^"]*"/\"\"/g;
131	s/'.'/''/g;
132
133	# an /* END JSTYLED */ comment ends a no-check block.
134	if ($nocheck) {
135		if (/\/\* *END *JSTYLED *\*\//) {
136			$nocheck = 0;
137		} else {
138			next line;
139		}
140	}
141
142	# a /*JSTYLED*/ comment indicates that the next line is ok.
143	if ($nextok) {
144		if ($okmsg) {
145			err($okmsg);
146		}
147		$nextok = 0;
148		$okmsg = 0;
149		if (/\/\* *JSTYLED.*\*\//) {
150			/^.*\/\* *JSTYLED *(.*) *\*\/.*$/;
151			$okmsg = $1;
152			$nextok = 1;
153		}
154		$prev = $line;
155		next line;
156	}
157
158	# remember whether we expect to be inside a continuation line.
159	$in_continuation = $expect_continuation;
160
161	# check for proper continuation line.  blank lines
162	# in the middle of the
163	# continuation do not count.
164	# XXX - only check within functions.
165	if ($check_continuation && $expect_continuation && $in_class &&
166	    !/^\s*$/) {
167		# continuation line must start with whitespace of
168		# previous line, plus either 4 spaces or a tab, but
169		# do not check lines that start with a string constant
170		# since they are often shifted to the left to make them
171		# fit on the line.
172		if (!/^$continuation_indent    \S/ &&
173		    !/^$continuation_indent\t\S/ && !/^\s*"/) {
174			err("continuation line improperly indented");
175		}
176		$expect_continuation = 0;
177	}
178
179	# a /* BEGIN JSTYLED */ comment starts a no-check block.
180	if (/\/\* *BEGIN *JSTYLED *\*\//) {
181		$nocheck = 1;
182	}
183
184	# a /*JSTYLED*/ comment indicates that the next line is ok.
185	if (/\/\* *JSTYLED.*\*\//) {
186		/^.*\/\* *JSTYLED *(.*) *\*\/.*$/;
187		$okmsg = $1;
188		$nextok = 1;
189	}
190	if (/\/\/ *JSTYLED/) {
191		/^.*\/\/ *JSTYLED *(.*)$/;
192		$okmsg = $1;
193		$nextok = 1;
194	}
195
196	# is this the beginning or ending of a class?
197	if (/^(public\s+)*\w(class|interface)\s/) {
198		$in_class = 1;
199		$in_declaration = 1;
200		$prev = $line;
201		next line;
202	}
203	if (/^}\s*(\/\*.*\*\/\s*)*$/) {
204		$in_class = 0;
205		$prev = $line;
206		next line;
207	}
208
209	if ($comment_done) {
210		$in_comment = 0;
211		$in_header_comment = 0;
212		$comment_done = 0;
213	}
214	# does this looks like the start of a block comment?
215	if (/^\s*\/\*/ && !/^\s*\/\*.*\*\//) {
216		if (/^\s*\/\*./ && !/^\s*\/\*\*$/) {
217			err("improper first line of block comment");
218		}
219		if (!/^(\t|    )*\/\*/) {
220			err("block comment not indented properly");
221		}
222		$in_comment = 1;
223		/^(\s*)\//;
224		$comment_prefix = $1;
225		if ($comment_prefix eq "") {
226			$in_header_comment = 1;
227		}
228		$prev = $line;
229		next line;
230	}
231	# are we still in the block comment?
232	if ($in_comment) {
233		if (/^$comment_prefix \*\/$/) {
234			$comment_done = 1;
235		} elsif (/\*\//) {
236			$comment_done = 1;
237			err("improper block comment close")
238			    unless ($ignore_hdr_comment && $in_header_comment);
239		} elsif (!/^$comment_prefix \*[ \t]/ &&
240		    !/^$comment_prefix \*$/) {
241			err("improper block comment")
242			    unless ($ignore_hdr_comment && $in_header_comment);
243		}
244	}
245
246	if ($in_header_comment && $ignore_hdr_comment) {
247		$prev = $line;
248		next line;
249	}
250
251	# check for errors that might occur in comments and in code.
252
253	# check length of line.
254	# first, a quick check to see if there is any chance of being too long.
255	if ($line =~ tr/\t/\t/ * 7 + length($line) > 80) {
256		# yes, there is a chance.
257		# replace tabs with spaces and check again.
258		my $eline = $line;
259		1 while $eline =~
260		    s/\t+/' ' x (length($&) * 8 - length($`) % 8)/e;
261		if (length($eline) > 80) {
262			err("line > 80 characters");
263		}
264	}
265
266	# Allow spaces to be used to draw pictures in header comments, but
267	# disallow blocks of spaces almost everywhere else.  In particular,
268	# five spaces are also allowed at the end of a line's indentation
269	# if the rest of the line belongs to a block comment.
270	if (!$in_header_comment &&
271	    /[^ ]     / &&
272	    !(/^\t*     \*/ && !/^\t*     \*.*     /)) {
273		err("spaces instead of tabs");
274	}
275	if ($tabs && /^ / && !/^ \*[ \t\/]/ && !/^ \*$/ &&
276	    (!/^    \w/ || $in_class != 0)) {
277		err("indent by spaces instead of tabs");
278	}
279	if (!$in_comment && (/^(\t    )* {1,3}\S/ || /^(\t    )* {5,7}\S/) &&
280	    !(/^\s*[-+|&\/?:=]/ || ($prev =~ /,\s*$/))) {
281		err("indent not a multiple of 4");
282	}
283	if (/\s$/) {
284		err("space or tab at end of line");
285	}
286if (0) {
287	if (/^[\t]+ [^ \t\*]/ || /^[\t]+  \S/ || /^[\t]+   \S/) {
288		err("continuation line not indented by 4 spaces");
289	}
290}
291	if (/\/\//) {
292		$cpp_comment = 1;
293	}
294	if (!$cpp_comment && /[^ \t(\/]\/\*/ && !/\w\(\/\*.*\*\/\);/) {
295		err("comment preceded by non-blank");
296	}
297	if (/\t +\t/) {
298		err("spaces between tabs");
299	}
300	if (/ \t+ /) {
301		err("tabs between spaces");
302	}
303
304	if ($in_comment) {	# still in comment
305		$prev = $line;
306		next line;
307	}
308
309	if (!$cpp_comment && ((/\/\*\S/ && !/\/\*\*/) || /\/\*\*\S/)) {
310		err("missing blank after open comment");
311	}
312	if (!$cpp_comment && /\S\*\//) {
313		err("missing blank before close comment");
314	}
315	# check for unterminated single line comments.
316	if (/\S.*\/\*/ && !/\S.*\/\*.*\*\//) {
317		err("unterminated single line comment");
318	}
319
320	# delete any comments and check everything else.  Be sure to leave
321	# //-style comments intact, and if there are multiple comments on a
322	# line, preserve whatever's in between.
323	s/(?<!\/)\/\*.*?\*\///g;
324	# Check for //-style comments only outside of block comments
325	if (m{(//(?!$))} && substr($_, $+[0], 1) !~ /[ \t]/) {
326		err("missing blank after start comment");
327	}
328	s/\/\/.*$//;		# C++ comments
329	$cpp_comment = 0;
330
331	# delete any trailing whitespace; we have already checked for that.
332	s/\s*$//;
333
334	# We don't style (yet) what's inside annotations, so just delete them.
335	s/$annotations//;
336
337	# following checks do not apply to text in comments.
338
339	# if it looks like an operator at the end of the line, and it is
340	# not really the end of a comment (...*/), and it is not really
341	# a label (done:), and it is not a case label (case FOO:),
342	# or we are not in a function definition (ANSI C style) and the
343	# operator is a "," (to avoid hitting "int\nfoo(\n\tint i,\n\tint j)"),
344	# or we are in a function and the operator is a
345	# "*" (to avoid hitting on "char*\nfunc()").
346	if ((/[-+|&\/?:=]$/ && !/\*\/$/ && !/^\s*\w*:$/ &&
347	    !/^\s\s*case\s\s*\w*:$/) ||
348	    /,$/ ||
349	    ($in_class && /\*$/)) {
350		$expect_continuation = 1;
351		if (!$in_continuation) {
352			/^(\s*)\S/;
353			$continuation_indent = $1;
354		}
355	}
356	while (/($allops)/g) {
357		my $z = substr($_, $-[1] - 1);
358		if ($z !~ /\s\Q$1\E(?:\s|$)/) {
359			my $m = $1;
360			my $shift;
361			# @+ is available only in the currently active
362			# dynamic scope.  Assign it to a new variable
363			# to pass it into the if block.
364			if ($z =~ /($generics)/ &&
365			    ($shift = $+[1])) {
366				pos $_ += $shift;
367				next;
368			}
369
370			# These need to be in decreasing order of length
371			# (violable as long as there's no ambiguity)
372			my $nospace = "missing space around";
373			if ($m =~ $shiftassignmentops) {
374				err("$nospace assignment operator");
375			} elsif ($m =~ $shiftops) {
376				err("$nospace shift operator");
377			} elsif ($m =~ $relationalops) {
378				err("$nospace relational operator");
379			} elsif ($m =~ $assignmentops) {
380				err("$nospace assignment operator");
381			}
382		}
383	}
384	if (/[,;]\S/ && !/\bfor \(;;\)/) {
385		err("comma or semicolon followed by non-blank");
386	}
387	# allow "for" statements to have empty "while" clauses
388	if (/\s[,;]/ && !/^[\t]+;$/ && !/^\s*for \([^;]*; ;[^;]*\)/) {
389		err("comma or semicolon preceded by blank");
390	}
391if (0) {
392	if (/^\s*(&&|\|\|)/) {
393		err("improper boolean continuation");
394	}
395}
396	if ($picky && /\S   *(&&|\|\|)/ || /(&&|\|\|)   *\S/) {
397		err("more than one space around boolean operator");
398	}
399	if (/\b$keywords\(/) {
400		err("missing space between keyword and paren");
401	}
402	if (/(\b$keywords\b.*){2,}/ && !/\bcase\b.*/) { # "case" excepted
403		err("more than one keyword on line");
404	}
405	if (/\b$keywords\s\s+\(/ &&
406	    !/^#if\s+\(/) {
407		err("extra space between keyword and paren");
408	}
409	# try to detect "func (x)" but not "if (x)" or
410	# "int (*func)();"
411	if (/\w\s\(/) {
412		my $save = $_;
413		# strip off all keywords on the line
414		s/\b$keywords\s\(/XXX(/g;
415		#s/\b($typename|void)\s+\(+/XXX(/og;
416		if (/\w\s\(/) {
417			err("extra space between function name and left paren");
418		}
419		$_ = $save;
420	}
421	if (/\(\s/) {
422		err("whitespace after left paren");
423	}
424	# allow "for" statements to have empty "continue" clauses
425	if (/\s\)/ && !/^\s*for \([^;]*;[^;]*; \)/) {
426		err("whitespace before right paren");
427	}
428	if (/^\s*\(void\)[^ ]/) {
429		err("missing space after (void) cast");
430	}
431	if (/\S{/ && !/{{/) {
432		err("missing space before left brace");
433	}
434	if ($in_class && /^\s+{/ && ($prev =~ /\)\s*$/)) {
435		err("left brace starting a line");
436	}
437	if (/}(else|while)/) {
438		err("missing space after right brace");
439	}
440	if (/}\s\s+(else|while)/) {
441		err("extra space after right brace");
442	}
443	if (/\b$typename\*/o) {
444		err("missing space between type name and *");
445	}
446	if ($heuristic) {
447		# cannot check this everywhere due to "struct {\n...\n} foo;"
448		if ($in_class && !$in_declaration &&
449		    /}./ && !/}\s+=/ && !/{.*}[;,]$/ && !/}(\s|)*$/ &&
450		    !/} (else|while)/ && !/}}/) {
451			err("possible bad text following right brace");
452		}
453		# cannot check this because sub-blocks in
454		# the middle of code are ok
455		if ($in_class && /^\s+{/) {
456			err("possible left brace starting a line");
457		}
458	}
459	if (/^\s*else\W/) {
460		if ($prev =~ /^\s*}$/) {
461			my $str = "else and right brace should be on same line";
462			printf $fmt, $filename, $., $str, $prev;
463			if ($verbose) {
464				printf "%s\n", $line;
465			}
466		}
467	}
468	$prev = $line;
469}
470
471if ($picky && $prev eq "") {
472	err("last line in file is blank");
473}
474
475}
476