xref: /illumos-gate/usr/src/tools/scripts/jstyle.pl (revision 8808ac5dae118369991f158b6ab736cb2691ecde)
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# Copyright 2015 Toomas Soome <tsoome@me.com>
25# Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
26# Use is subject to license terms.
27#
28# jstyle - check for some common stylistic errors.
29#
30
31require 5.006;
32use Getopt::Std;
33use strict;
34
35my $usage =
36"usage: jstyle [-c] [-h] [-p] [-t] [-v] [-C] file ...
37	-c	check continuation line indenting
38	-h	perform heuristic checks that are sometimes wrong
39	-p	perform some of the more picky checks
40	-t	insist on indenting by tabs
41	-v	verbose
42	-C	don't check anything in header block comments
43";
44
45my %opts;
46
47# Keep -s, as it's been around for a while.  It just doesn't do anything
48# anymore.
49if (!getopts("chpstvC", \%opts)) {
50	print $usage;
51	exit 2;
52}
53
54my $check_continuation = $opts{'c'};
55my $heuristic = $opts{'h'};
56my $picky = $opts{'p'};
57my $tabs = $opts{'t'};
58my $verbose = $opts{'v'};
59my $ignore_hdr_comment = $opts{'C'};
60
61my ($filename, $line, $prev);
62my $err_stat = 0;		# Exit status
63
64my $fmt;
65
66if ($verbose) {
67	$fmt = "%s: %d: %s\n%s\n";
68} else {
69	$fmt = "%s: %d: %s\n";
70}
71
72# Note, following must be in single quotes so that \s and \w work right.
73my $typename = '(int|char|boolean|byte|short|long|float|double)';
74my $keywords = '(for|if|while|switch|return|catch|synchronized|throw|assert)';
75# See perlre(1) for the meaning of (??{ ... })
76my $annotations = ""; $annotations = qr/@\w+\((?:(?>[^()]+)|(??{ $annotations }))*\)/;
77my $generics = ""; $generics = qr/<(([\s\w,.?[\]]| & )+|(??{ $generics }))*>/;
78my $relationalops = qr/>=|<=|<|>|!=|==/;
79my $shiftops = qr/<<<|>>>|<<|>>/;
80my $shiftassignmentops = qr/[<>]{2,3}=/;
81my $assignmentops = qr/[-+\/*|&^%]?=/;
82# These need to be in decreasing order of length
83my $allops = qr/$shiftassignmentops|$shiftops|$relationalops|$assignmentops/;
84
85if ($#ARGV >= 0) {
86	foreach my $arg (@ARGV) {
87		if (!open(STDIN, $arg)) {
88			printf "%s: can not open\n", $arg;
89		} else {
90			&jstyle($arg);
91			close STDIN;
92		}
93	}
94} else {
95	&jstyle("<stdin>");
96}
97exit $err_stat;
98
99sub err($) {
100	if ($verbose) {
101		printf $fmt, $filename, $., $_[0], $line;
102	} else {
103		printf $fmt, $filename, $., $_[0];
104	}
105	$err_stat = 1;
106}
107
108sub jstyle($) {
109
110my $in_comment = 0;
111my $in_header_comment = 0;
112my $in_continuation = 0;
113my $in_class = 0;
114my $in_declaration = 0;
115my $nextok = 0;
116my $nocheck = 0;
117my $expect_continuation = 0;
118my $continuation_indent;
119my $okmsg;
120my $comment_prefix;
121my $comment_done;
122my $cpp_comment;
123
124$filename = $_[0];
125
126line: while (<STDIN>) {
127	s/\r?\n$//;	# strip return and newline
128
129	# save the original line, then remove all text from within
130	# double or single quotes, we do not want to check such text.
131
132	$line = $_;
133	s/"[^"]*"/\"\"/g;
134	s/'.'/''/g;
135
136	# an /* END JSTYLED */ comment ends a no-check block.
137	if ($nocheck) {
138		if (/\/\* *END *JSTYLED *\*\//) {
139			$nocheck = 0;
140		} else {
141			next line;
142		}
143	}
144
145	# a /*JSTYLED*/ comment indicates that the next line is ok.
146	if ($nextok) {
147		if ($okmsg) {
148			err($okmsg);
149		}
150		$nextok = 0;
151		$okmsg = 0;
152		if (/\/\* *JSTYLED.*\*\//) {
153			/^.*\/\* *JSTYLED *(.*) *\*\/.*$/;
154			$okmsg = $1;
155			$nextok = 1;
156		}
157		$prev = $line;
158		next line;
159	}
160
161	# remember whether we expect to be inside a continuation line.
162	$in_continuation = $expect_continuation;
163
164	# check for proper continuation line.  blank lines
165	# in the middle of the
166	# continuation do not count.
167	# XXX - only check within functions.
168	if ($check_continuation && $expect_continuation && $in_class &&
169	    !/^\s*$/) {
170		# continuation line must start with whitespace of
171		# previous line, plus either 4 spaces or a tab, but
172		# do not check lines that start with a string constant
173		# since they are often shifted to the left to make them
174		# fit on the line.
175		if (!/^$continuation_indent    \S/ &&
176		    !/^$continuation_indent\t\S/ && !/^\s*"/) {
177			err("continuation line improperly indented");
178		}
179		$expect_continuation = 0;
180	}
181
182	# a /* BEGIN JSTYLED */ comment starts a no-check block.
183	if (/\/\* *BEGIN *JSTYLED *\*\//) {
184		$nocheck = 1;
185	}
186
187	# a /*JSTYLED*/ comment indicates that the next line is ok.
188	if (/\/\* *JSTYLED.*\*\//) {
189		/^.*\/\* *JSTYLED *(.*) *\*\/.*$/;
190		$okmsg = $1;
191		$nextok = 1;
192	}
193	if (/\/\/ *JSTYLED/) {
194		/^.*\/\/ *JSTYLED *(.*)$/;
195		$okmsg = $1;
196		$nextok = 1;
197	}
198
199	# is this the beginning or ending of a class?
200	if (/^(public\s+)*\w(class|interface)\s/) {
201		$in_class = 1;
202		$in_declaration = 1;
203		$prev = $line;
204		next line;
205	}
206	if (/^}\s*(\/\*.*\*\/\s*)*$/) {
207		$in_class = 0;
208		$prev = $line;
209		next line;
210	}
211
212	if ($comment_done) {
213		$in_comment = 0;
214		$in_header_comment = 0;
215		$comment_done = 0;
216	}
217	# does this looks like the start of a block comment?
218	if (/^\s*\/\*/ && !/^\s*\/\*.*\*\//) {
219		if (/^\s*\/\*./ && !/^\s*\/\*\*$/) {
220			err("improper first line of block comment");
221		}
222		if (!/^(\t|    )*\/\*/) {
223			err("block comment not indented properly");
224		}
225		$in_comment = 1;
226		/^(\s*)\//;
227		$comment_prefix = $1;
228		if ($comment_prefix eq "") {
229			$in_header_comment = 1;
230		}
231		$prev = $line;
232		next line;
233	}
234	# are we still in the block comment?
235	if ($in_comment) {
236		if (/^$comment_prefix \*\/$/) {
237			$comment_done = 1;
238		} elsif (/\*\//) {
239			$comment_done = 1;
240			err("improper block comment close")
241			    unless ($ignore_hdr_comment && $in_header_comment);
242		} elsif (!/^$comment_prefix \*[ \t]/ &&
243		    !/^$comment_prefix \*$/) {
244			err("improper block comment")
245			    unless ($ignore_hdr_comment && $in_header_comment);
246		}
247	}
248
249	if ($in_header_comment && $ignore_hdr_comment) {
250		$prev = $line;
251		next line;
252	}
253
254	# check for errors that might occur in comments and in code.
255
256	# check length of line.
257	# first, a quick check to see if there is any chance of being too long.
258	if ($line =~ tr/\t/\t/ * 7 + length($line) > 80) {
259		# yes, there is a chance.
260		# replace tabs with spaces and check again.
261		my $eline = $line;
262		1 while $eline =~
263		    s/\t+/' ' x (length($&) * 8 - length($`) % 8)/e;
264		if (length($eline) > 80) {
265			err("line > 80 characters");
266		}
267	}
268
269	# Allow spaces to be used to draw pictures in header comments, but
270	# disallow blocks of spaces almost everywhere else.  In particular,
271	# five spaces are also allowed at the end of a line's indentation
272	# if the rest of the line belongs to a block comment.
273	if (!$in_header_comment &&
274	    /[^ ]     / &&
275	    !(/^\t*     \*/ && !/^\t*     \*.*     /)) {
276		err("spaces instead of tabs");
277	}
278	if ($tabs && /^ / && !/^ \*[ \t\/]/ && !/^ \*$/ &&
279	    (!/^    \w/ || $in_class != 0)) {
280		err("indent by spaces instead of tabs");
281	}
282	if (!$in_comment && (/^(\t    )* {1,3}\S/ || /^(\t    )* {5,7}\S/) &&
283	    !(/^\s*[-+|&\/?:=]/ || ($prev =~ /,\s*$/))) {
284		err("indent not a multiple of 4");
285	}
286	if (/\s$/) {
287		err("space or tab at end of line");
288	}
289if (0) {
290	if (/^[\t]+ [^ \t\*]/ || /^[\t]+  \S/ || /^[\t]+   \S/) {
291		err("continuation line not indented by 4 spaces");
292	}
293}
294	if (/\/\//) {
295		$cpp_comment = 1;
296	}
297	if (!$cpp_comment && /[^ \t(\/]\/\*/ && !/\w\(\/\*.*\*\/\);/) {
298		err("comment preceded by non-blank");
299	}
300	if (/\t +\t/) {
301		err("spaces between tabs");
302	}
303	if (/ \t+ /) {
304		err("tabs between spaces");
305	}
306
307	if ($in_comment) {	# still in comment
308		$prev = $line;
309		next line;
310	}
311
312	if (!$cpp_comment && ((/\/\*\S/ && !/\/\*\*/) || /\/\*\*\S/)) {
313		err("missing blank after open comment");
314	}
315	if (!$cpp_comment && /\S\*\//) {
316		err("missing blank before close comment");
317	}
318	# check for unterminated single line comments.
319	if (/\S.*\/\*/ && !/\S.*\/\*.*\*\//) {
320		err("unterminated single line comment");
321	}
322
323	# delete any comments and check everything else.  Be sure to leave
324	# //-style comments intact, and if there are multiple comments on a
325	# line, preserve whatever's in between.
326	s/(?<!\/)\/\*.*?\*\///g;
327	# Check for //-style comments only outside of block comments
328	if (m{(//(?!$))} && substr($_, $+[0], 1) !~ /[ \t]/) {
329		err("missing blank after start comment");
330	}
331	s/\/\/.*$//;		# C++ comments
332	$cpp_comment = 0;
333
334	# delete any trailing whitespace; we have already checked for that.
335	s/\s*$//;
336
337	# We don't style (yet) what's inside annotations, so just delete them.
338	s/$annotations//;
339
340	# following checks do not apply to text in comments.
341
342	# if it looks like an operator at the end of the line, and it is
343	# not really the end of a comment (...*/), and it is not really
344	# a label (done:), and it is not a case label (case FOO:),
345	# or we are not in a function definition (ANSI C style) and the
346	# operator is a "," (to avoid hitting "int\nfoo(\n\tint i,\n\tint j)"),
347	# or we are in a function and the operator is a
348	# "*" (to avoid hitting on "char*\nfunc()").
349	if ((/[-+|&\/?:=]$/ && !/\*\/$/ && !/^\s*\w*:$/ &&
350	    !/^\s\s*case\s\s*\w*:$/) ||
351	    /,$/ ||
352	    ($in_class && /\*$/)) {
353		$expect_continuation = 1;
354		if (!$in_continuation) {
355			/^(\s*)\S/;
356			$continuation_indent = $1;
357		}
358	}
359	while (/($allops)/g) {
360		my $z = substr($_, $-[1] - 1);
361		if ($z !~ /\s\Q$1\E(?:\s|$)/) {
362			my $m = $1;
363			my $shift;
364			# @+ is available only in the currently active
365			# dynamic scope.  Assign it to a new variable
366			# to pass it into the if block.
367			if ($z =~ /($generics)/ &&
368			    ($shift = $+[1])) {
369				pos $_ += $shift;
370				next;
371			}
372
373			# These need to be in decreasing order of length
374			# (violable as long as there's no ambiguity)
375			my $nospace = "missing space around";
376			if ($m =~ $shiftassignmentops) {
377				err("$nospace assignment operator");
378			} elsif ($m =~ $shiftops) {
379				err("$nospace shift operator");
380			} elsif ($m =~ $relationalops) {
381				err("$nospace relational operator");
382			} elsif ($m =~ $assignmentops) {
383				err("$nospace assignment operator");
384			}
385		}
386	}
387	if (/[,;]\S/ && !/\bfor \(;;\)/) {
388		err("comma or semicolon followed by non-blank");
389	}
390	# allow "for" statements to have empty "while" clauses
391	if (/\s[,;]/ && !/^[\t]+;$/ && !/^\s*for \([^;]*; ;[^;]*\)/) {
392		err("comma or semicolon preceded by blank");
393	}
394if (0) {
395	if (/^\s*(&&|\|\|)/) {
396		err("improper boolean continuation");
397	}
398}
399	if ($picky && /\S   *(&&|\|\|)/ || /(&&|\|\|)   *\S/) {
400		err("more than one space around boolean operator");
401	}
402	if (/\b$keywords\(/) {
403		err("missing space between keyword and paren");
404	}
405	if (/(\b$keywords\b.*){2,}/ && !/\bcase\b.*/) { # "case" excepted
406		err("more than one keyword on line");
407	}
408	if (/\b$keywords\s\s+\(/ &&
409	    !/^#if\s+\(/) {
410		err("extra space between keyword and paren");
411	}
412	# try to detect "func (x)" but not "if (x)" or
413	# "int (*func)();"
414	if (/\w\s\(/) {
415		my $save = $_;
416		# strip off all keywords on the line
417		s/\b$keywords\s\(/XXX(/g;
418		#s/\b($typename|void)\s+\(+/XXX(/og;
419		if (/\w\s\(/) {
420			err("extra space between function name and left paren");
421		}
422		$_ = $save;
423	}
424	if (/\(\s/) {
425		err("whitespace after left paren");
426	}
427	# allow "for" statements to have empty "continue" clauses
428	if (/\s\)/ && !/^\s*for \([^;]*;[^;]*; \)/) {
429		err("whitespace before right paren");
430	}
431	if (/^\s*\(void\)[^ ]/) {
432		err("missing space after (void) cast");
433	}
434	if (/\S\{/ && !/\{\{/) {
435		err("missing space before left brace");
436	}
437	if ($in_class && /^\s+{/ && ($prev =~ /\)\s*$/)) {
438		err("left brace starting a line");
439	}
440	if (/}(else|while)/) {
441		err("missing space after right brace");
442	}
443	if (/}\s\s+(else|while)/) {
444		err("extra space after right brace");
445	}
446	if (/\b$typename\*/o) {
447		err("missing space between type name and *");
448	}
449	if ($heuristic) {
450		# cannot check this everywhere due to "struct {\n...\n} foo;"
451		if ($in_class && !$in_declaration &&
452		    /}./ && !/}\s+=/ && !/{.*}[;,]$/ && !/}(\s|)*$/ &&
453		    !/} (else|while)/ && !/}}/) {
454			err("possible bad text following right brace");
455		}
456		# cannot check this because sub-blocks in
457		# the middle of code are ok
458		if ($in_class && /^\s+{/) {
459			err("possible left brace starting a line");
460		}
461	}
462	if (/^\s*else\W/) {
463		if ($prev =~ /^\s*}$/) {
464			my $str = "else and right brace should be on same line";
465			printf $fmt, $filename, $., $str, $prev;
466			if ($verbose) {
467				printf "%s\n", $line;
468			}
469		}
470	}
471	$prev = $line;
472}
473
474if ($picky && $prev eq "") {
475	err("last line in file is blank");
476}
477
478}
479