xref: /illumos-gate/usr/src/tools/scripts/git-pbchk.py (revision 8c69cc8fbe729fa7b091e901c4b50508ccc6bb33)
1#!@PYTHON@
2#
3#  This program is free software; you can redistribute it and/or modify
4#  it under the terms of the GNU General Public License version 2
5#  as published by the Free Software Foundation.
6#
7#  This program is distributed in the hope that it will be useful,
8#  but WITHOUT ANY WARRANTY; without even the implied warranty of
9#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
10#  GNU General Public License for more details.
11#
12#  You should have received a copy of the GNU General Public License
13#  along with this program; if not, write to the Free Software
14#  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
15#
16
17#
18# Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
19# Copyright 2008, 2012 Richard Lowe
20# Copyright 2014 Garrett D'Amore <garrett@damore.org>
21# Copyright (c) 2014, Joyent, Inc.
22# Copyright (c) 2015, 2016 by Delphix. All rights reserved.
23# Copyright 2016 Nexenta Systems, Inc.
24#
25
26import getopt
27import os
28import re
29import subprocess
30import sys
31import tempfile
32
33from cStringIO import StringIO
34
35#
36# Adjust the load path based on our location and the version of python into
37# which it is being loaded.  This assumes the normal onbld directory
38# structure, where we are in bin/ and the modules are in
39# lib/python(version)?/onbld/Scm/.  If that changes so too must this.
40#
41sys.path.insert(1, os.path.join(os.path.dirname(__file__), "..", "lib",
42                                "python%d.%d" % sys.version_info[:2]))
43
44#
45# Add the relative path to usr/src/tools to the load path, such that when run
46# from the source tree we use the modules also within the source tree.
47#
48sys.path.insert(2, os.path.join(os.path.dirname(__file__), ".."))
49
50from onbld.Scm import Ignore
51from onbld.Checks import Comments, Copyright, CStyle, HdrChk
52from onbld.Checks import JStyle, Keywords, ManLint, Mapfile, SpellCheck
53
54
55class GitError(Exception):
56    pass
57
58def git(command):
59    """Run a command and return a stream containing its stdout (and write its
60    stderr to its stdout)"""
61
62    if type(command) != list:
63        command = command.split()
64
65    command = ["git"] + command
66
67    try:
68        tmpfile = tempfile.TemporaryFile(prefix="git-nits")
69    except EnvironmentError, e:
70        raise GitError("Could not create temporary file: %s\n" % e)
71
72    try:
73        p = subprocess.Popen(command,
74                             stdout=tmpfile,
75                             stderr=subprocess.PIPE)
76    except OSError, e:
77        raise GitError("could not execute %s: %s\n" % (command, e))
78
79    err = p.wait()
80    if err != 0:
81        raise GitError(p.stderr.read())
82
83    tmpfile.seek(0)
84    return tmpfile
85
86
87def git_root():
88    """Return the root of the current git workspace"""
89
90    p = git('rev-parse --git-dir')
91
92    if not p:
93        sys.stderr.write("Failed finding git workspace\n")
94        sys.exit(err)
95
96    return os.path.abspath(os.path.join(p.readlines()[0],
97                                        os.path.pardir))
98
99
100def git_branch():
101    """Return the current git branch"""
102
103    p = git('branch')
104
105    if not p:
106        sys.stderr.write("Failed finding git branch\n")
107        sys.exit(err)
108
109    for elt in p:
110        if elt[0] == '*':
111            if elt.endswith('(no branch)'):
112                return None
113            return elt.split()[1]
114
115
116def git_parent_branch(branch):
117    """Return the parent of the current git branch.
118
119    If this branch tracks a remote branch, return the remote branch which is
120    tracked.  If not, default to origin/master."""
121
122    if not branch:
123        return None
124
125    p = git(["for-each-ref", "--format=%(refname:short) %(upstream:short)",
126            "refs/heads/"])
127
128    if not p:
129        sys.stderr.write("Failed finding git parent branch\n")
130        sys.exit(err)
131
132    for line in p:
133        # Git 1.7 will leave a ' ' trailing any non-tracking branch
134        if ' ' in line and not line.endswith(' \n'):
135            local, remote = line.split()
136            if local == branch:
137                return remote
138    return 'origin/master'
139
140
141def git_comments(parent):
142    """Return a list of any checkin comments on this git branch"""
143
144    p = git('log --pretty=tformat:%%B:SEP: %s..' % parent)
145
146    if not p:
147        sys.stderr.write("Failed getting git comments\n")
148        sys.exit(err)
149
150    return [x.strip() for x in p.readlines() if x != ':SEP:\n']
151
152
153def git_file_list(parent, paths=None):
154    """Return the set of files which have ever changed on this branch.
155
156    NB: This includes files which no longer exist, or no longer actually
157    differ."""
158
159    p = git("log --name-only --pretty=format: %s.. %s" %
160             (parent, ' '.join(paths)))
161
162    if not p:
163        sys.stderr.write("Failed building file-list from git\n")
164        sys.exit(err)
165
166    ret = set()
167    for fname in p:
168        if fname and not fname.isspace() and fname not in ret:
169            ret.add(fname.strip())
170
171    return ret
172
173
174def not_check(root, cmd):
175    """Return a function which returns True if a file given as an argument
176    should be excluded from the check named by 'cmd'"""
177
178    ignorefiles = filter(os.path.exists,
179                         [os.path.join(root, ".git", "%s.NOT" % cmd),
180                          os.path.join(root, "exception_lists", cmd)])
181    return Ignore.ignore(root, ignorefiles)
182
183
184def gen_files(root, parent, paths, exclude):
185    """Return a function producing file names, relative to the current
186    directory, of any file changed on this branch (limited to 'paths' if
187    requested), and excluding files for which exclude returns a true value """
188
189    # Taken entirely from Python 2.6's os.path.relpath which we would use if we
190    # could.
191    def relpath(path, here):
192        c = os.path.abspath(os.path.join(root, path)).split(os.path.sep)
193        s = os.path.abspath(here).split(os.path.sep)
194        l = len(os.path.commonprefix((s, c)))
195        return os.path.join(*[os.path.pardir] * (len(s)-l) + c[l:])
196
197    def ret(select=None):
198        if not select:
199            select = lambda x: True
200
201        for f in git_file_list(parent, paths):
202            f = relpath(f, '.')
203            try:
204                res = git("diff %s HEAD %s" % (parent, f))
205            except GitError, e:
206                # This ignores all the errors that can be thrown. Usually, this means
207                # that git returned non-zero because the file doesn't exist, but it
208                # could also fail if git can't create a new file or it can't be
209                # executed.  Such errors are 1) unlikely, and 2) will be caught by other
210                # invocations of git().
211                continue
212            empty = not res.readline()
213            if (os.path.isfile(f) and not empty and select(f) and not exclude(f)):
214                yield f
215    return ret
216
217
218def comchk(root, parent, flist, output):
219    output.write("Comments:\n")
220
221    return Comments.comchk(git_comments(parent), check_db=True,
222                           output=output)
223
224
225def mapfilechk(root, parent, flist, output):
226    ret = 0
227
228    # We are interested in examining any file that has the following
229    # in its final path segment:
230    #    - Contains the word 'mapfile'
231    #    - Begins with 'map.'
232    #    - Ends with '.map'
233    # We don't want to match unless these things occur in final path segment
234    # because directory names with these strings don't indicate a mapfile.
235    # We also ignore files with suffixes that tell us that the files
236    # are not mapfiles.
237    MapfileRE = re.compile(r'.*((mapfile[^/]*)|(/map\.+[^/]*)|(\.map))$',
238        re.IGNORECASE)
239    NotMapSuffixRE = re.compile(r'.*\.[ch]$', re.IGNORECASE)
240
241    output.write("Mapfile comments:\n")
242
243    for f in flist(lambda x: MapfileRE.match(x) and not
244                   NotMapSuffixRE.match(x)):
245        fh = open(f, 'r')
246        ret |= Mapfile.mapfilechk(fh, output=output)
247        fh.close()
248    return ret
249
250
251def copyright(root, parent, flist, output):
252    ret = 0
253    output.write("Copyrights:\n")
254    for f in flist():
255        fh = open(f, 'r')
256        ret |= Copyright.copyright(fh, output=output)
257        fh.close()
258    return ret
259
260
261def hdrchk(root, parent, flist, output):
262    ret = 0
263    output.write("Header format:\n")
264    for f in flist(lambda x: x.endswith('.h')):
265        fh = open(f, 'r')
266        ret |= HdrChk.hdrchk(fh, lenient=True, output=output)
267        fh.close()
268    return ret
269
270
271def cstyle(root, parent, flist, output):
272    ret = 0
273    output.write("C style:\n")
274    for f in flist(lambda x: x.endswith('.c') or x.endswith('.h')):
275        fh = open(f, 'r')
276        ret |= CStyle.cstyle(fh, output=output, picky=True,
277                             check_posix_types=True,
278                             check_continuation=True)
279        fh.close()
280    return ret
281
282
283def jstyle(root, parent, flist, output):
284    ret = 0
285    output.write("Java style:\n")
286    for f in flist(lambda x: x.endswith('.java')):
287        fh = open(f, 'r')
288        ret |= JStyle.jstyle(fh, output=output, picky=True)
289        fh.close()
290    return ret
291
292
293def manlint(root, parent, flist, output):
294    ret = 0
295    output.write("Man page format/spelling:\n")
296    ManfileRE = re.compile(r'.*\.[0-9][a-z]*$', re.IGNORECASE)
297    for f in flist(lambda x: ManfileRE.match(x)):
298        fh = open(f, 'r')
299        ret |= ManLint.manlint(fh, output=output, picky=True)
300        ret |= SpellCheck.spellcheck(fh, output=output)
301        fh.close()
302    return ret
303
304def keywords(root, parent, flist, output):
305    ret = 0
306    output.write("SCCS Keywords:\n")
307    for f in flist():
308        fh = open(f, 'r')
309        ret |= Keywords.keywords(fh, output=output)
310        fh.close()
311    return ret
312
313
314def run_checks(root, parent, cmds, paths='', opts={}):
315    """Run the checks given in 'cmds', expected to have well-known signatures,
316    and report results for any which fail.
317
318    Return failure if any of them did.
319
320    NB: the function name of the commands passed in is used to name the NOT
321    file which excepts files from them."""
322
323    ret = 0
324
325    for cmd in cmds:
326        s = StringIO()
327
328        exclude = not_check(root, cmd.func_name)
329        result = cmd(root, parent, gen_files(root, parent, paths, exclude),
330                     output=s)
331        ret |= result
332
333        if result != 0:
334            print s.getvalue()
335
336    return ret
337
338
339def nits(root, parent, paths):
340    cmds = [copyright,
341            cstyle,
342            hdrchk,
343            jstyle,
344            keywords,
345            manlint,
346            mapfilechk]
347    run_checks(root, parent, cmds, paths)
348
349
350def pbchk(root, parent, paths):
351    cmds = [comchk,
352            copyright,
353            cstyle,
354            hdrchk,
355            jstyle,
356            keywords,
357            manlint,
358            mapfilechk]
359    run_checks(root, parent, cmds)
360
361
362def main(cmd, args):
363    parent_branch = None
364
365    try:
366        opts, args = getopt.getopt(args, 'b:')
367    except getopt.GetoptError, e:
368        sys.stderr.write(str(e) + '\n')
369        sys.stderr.write("Usage: %s [-b branch] [path...]\n" % cmd)
370        sys.exit(1)
371
372    for opt, arg in opts:
373        if opt == '-b':
374            parent_branch = arg
375
376    if not parent_branch:
377        parent_branch = git_parent_branch(git_branch())
378
379    func = nits
380    if cmd == 'git-pbchk':
381        func = pbchk
382        if args:
383            sys.stderr.write("only complete workspaces may be pbchk'd\n");
384            sys.exit(1)
385
386    func(git_root(), parent_branch, args)
387
388if __name__ == '__main__':
389    try:
390        main(os.path.basename(sys.argv[0]), sys.argv[1:])
391    except GitError, e:
392        sys.stderr.write("failed to run git:\n %s\n" % str(e))
393        sys.exit(1)
394