xref: /illumos-gate/usr/src/tools/scripts/git-pbchk.py (revision 694c35faa87b858ecdadfe4fc592615f4eefbb07)
1#!/usr/bin/python2.6
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#
21
22import getopt
23import os
24import re
25import subprocess
26import sys
27import tempfile
28
29from cStringIO import StringIO
30
31# This is necessary because, in a fit of pique, we used hg-format ignore lists
32# for NOT files.
33from mercurial import ignore
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.Checks import Comments, Copyright, CStyle, HdrChk
51from onbld.Checks import JStyle, Keywords, Mapfile
52
53
54class GitError(Exception):
55    pass
56
57def git(command):
58    """Run a command and return a stream containing its stdout (and write its
59    stderr to its stdout)"""
60
61    if type(command) != list:
62        command = command.split()
63
64    command = ["git"] + command
65
66    try:
67        tmpfile = tempfile.TemporaryFile(prefix="git-nits")
68    except EnvironmentError, e:
69        raise GitError("Could not create temporary file: %s\n" % e)
70
71    try:
72        p = subprocess.Popen(command,
73                             stdout=tmpfile,
74                             stderr=subprocess.STDOUT)
75    except OSError, e:
76        raise GitError("could not execute %s: %s\n" (command, e))
77
78    err = p.wait()
79    if err != 0:
80        raise GitError(p.stdout.read())
81
82    tmpfile.seek(0)
83    return tmpfile
84
85
86def git_root():
87    """Return the root of the current git workspace"""
88
89    p = git('rev-parse --git-dir')
90
91    if not p:
92        sys.stderr.write("Failed finding git workspace\n")
93        sys.exit(err)
94
95    return os.path.abspath(os.path.join(p.readlines()[0],
96                                        os.path.pardir))
97
98
99def git_branch():
100    """Return the current git branch"""
101
102    p = git('branch')
103
104    if not p:
105        sys.stderr.write("Failed finding git branch\n")
106        sys.exit(err)
107
108    for elt in p:
109        if elt[0] == '*':
110            if elt.endswith('(no branch)'):
111                return None
112            return elt.split()[1]
113
114
115def git_parent_branch(branch):
116    """Return the parent of the current git branch.
117
118    If this branch tracks a remote branch, return the remote branch which is
119    tracked.  If not, default to origin/master."""
120
121    if not branch:
122        return None
123
124    p = git("for-each-ref --format=%(refname:short) %(upstream:short) " +
125            "refs/heads/")
126
127    if not p:
128        sys.stderr.write("Failed finding git parent branch\n")
129        sys.exit(err)
130
131    for line in p:
132        # Git 1.7 will leave a ' ' trailing any non-tracking branch
133        if ' ' in line and not line.endswith(' \n'):
134            local, remote = line.split()
135            if local == branch:
136                return remote
137    return 'origin/master'
138
139
140def git_comments(parent):
141    """Return a list of any checkin comments on this git branch"""
142
143    p = git('log --pretty=tformat:%%B:SEP: %s..' % parent)
144
145    if not p:
146        sys.stderr.write("Failed getting git comments\n")
147        sys.exit(err)
148
149    return [x.strip() for x in p.readlines() if x != ':SEP:\n']
150
151
152def git_file_list(parent, paths=None):
153    """Return the set of files which have ever changed on this branch.
154
155    NB: This includes files which no longer exist, or no longer actually
156    differ."""
157
158    p = git("log --name-only --pretty=format: %s.. %s" %
159             (parent, ' '.join(paths)))
160
161    if not p:
162        sys.stderr.write("Failed building file-list from git\n")
163        sys.exit(err)
164
165    ret = set()
166    for fname in p:
167        if fname and not fname.isspace() and fname not in ret:
168            ret.add(fname.strip())
169
170    return ret
171
172
173def not_check(root, cmd):
174    """Return a function which returns True if a file given as an argument
175    should be excluded from the check named by 'cmd'"""
176
177    ignorefiles = filter(os.path.exists,
178                         [os.path.join(root, ".git", "%s.NOT" % cmd),
179                          os.path.join(root, "exception_lists", cmd)])
180    if len(ignorefiles) > 0:
181        return ignore.ignore(root, ignorefiles, sys.stderr.write)
182    else:
183        return lambda x: False
184
185
186def gen_files(root, parent, paths, exclude):
187    """Return a function producing file names, relative to the current
188    directory, of any file changed on this branch (limited to 'paths' if
189    requested), and excluding files for which exclude returns a true value """
190
191    # Taken entirely from Python 2.6's os.path.relpath which we would use if we
192    # could.
193    def relpath(path, here):
194        c = os.path.abspath(os.path.join(root, path)).split(os.path.sep)
195        s = os.path.abspath(here).split(os.path.sep)
196        l = len(os.path.commonprefix((s, c)))
197        return os.path.join(*[os.path.pardir] * (len(s)-l) + c[l:])
198
199    def ret(select=None):
200        if not select:
201            select = lambda x: True
202
203        for f in git_file_list(parent, paths):
204            f = relpath(f, '.')
205            if (os.path.exists(f) and select(f) and not exclude(f)):
206                yield f
207    return ret
208
209
210def comchk(root, parent, flist, output):
211    output.write("Comments:\n")
212
213    return Comments.comchk(git_comments(parent), check_db=True,
214                           output=output)
215
216
217def mapfilechk(root, parent, flist, output):
218    ret = 0
219
220    # We are interested in examining any file that has the following
221    # in its final path segment:
222    #    - Contains the word 'mapfile'
223    #    - Begins with 'map.'
224    #    - Ends with '.map'
225    # We don't want to match unless these things occur in final path segment
226    # because directory names with these strings don't indicate a mapfile.
227    # We also ignore files with suffixes that tell us that the files
228    # are not mapfiles.
229    MapfileRE = re.compile(r'.*((mapfile[^/]*)|(/map\.+[^/]*)|(\.map))$',
230        re.IGNORECASE)
231    NotMapSuffixRE = re.compile(r'.*\.[ch]$', re.IGNORECASE)
232
233    output.write("Mapfile comments:\n")
234
235    for f in flist(lambda x: MapfileRE.match(x) and not
236                   NotMapSuffixRE.match(x)):
237        fh = open(f, 'r')
238        ret |= Mapfile.mapfilechk(fh, output=output)
239        fh.close()
240    return ret
241
242
243def copyright(root, parent, flist, output):
244    ret = 0
245    output.write("Copyrights:\n")
246    for f in flist():
247        fh = open(f, 'r')
248        ret |= Copyright.copyright(fh, output=output)
249        fh.close()
250    return ret
251
252
253def hdrchk(root, parent, flist, output):
254    ret = 0
255    output.write("Header format:\n")
256    for f in flist(lambda x: x.endswith('.h')):
257        fh = open(f, 'r')
258        ret |= HdrChk.hdrchk(fh, lenient=True, output=output)
259        fh.close()
260    return ret
261
262
263def cstyle(root, parent, flist, output):
264    ret = 0
265    output.write("C style:\n")
266    for f in flist(lambda x: x.endswith('.c') or x.endswith('.h')):
267        fh = open(f, 'r')
268        ret |= CStyle.cstyle(fh, output=output, picky=True,
269                             check_posix_types=True,
270                             check_continuation=True)
271        fh.close()
272    return ret
273
274
275def jstyle(root, parent, flist, output):
276    ret = 0
277    output.write("Java style:\n")
278    for f in flist(lambda x: x.endswith('.java')):
279        fh = open(f, 'r')
280        ret |= JStyle.jstyle(fh, output=output, picky=True)
281        fh.close()
282    return ret
283
284
285def keywords(root, parent, flist, output):
286    ret = 0
287    output.write("SCCS Keywords:\n")
288    for f in flist():
289        fh = open(f, 'r')
290        ret |= Keywords.keywords(fh, output=output)
291        fh.close()
292    return ret
293
294
295def run_checks(root, parent, cmds, paths='', opts={}):
296    """Run the checks given in 'cmds', expected to have well-known signatures,
297    and report results for any which fail.
298
299    Return failure if any of them did.
300
301    NB: the function name of the commands passed in is used to name the NOT
302    file which excepts files from them."""
303
304    ret = 0
305
306    for cmd in cmds:
307        s = StringIO()
308
309        exclude = not_check(root, cmd.func_name)
310        result = cmd(root, parent, gen_files(root, parent, paths, exclude),
311                     output=s)
312        ret |= result
313
314        if result != 0:
315            print s.getvalue()
316
317    return ret
318
319
320def nits(root, parent, paths):
321    cmds = [copyright,
322            cstyle,
323            hdrchk,
324            jstyle,
325            keywords,
326            mapfilechk]
327    run_checks(root, parent, cmds, paths)
328
329
330def pbchk(root, parent, paths):
331    cmds = [comchk,
332            copyright,
333            cstyle,
334            hdrchk,
335            jstyle,
336            keywords,
337            mapfilechk]
338    run_checks(root, parent, cmds)
339
340
341def main(cmd, args):
342    parent_branch = None
343
344    try:
345        opts, args = getopt.getopt(args, 'b:')
346    except getopt.GetoptError, e:
347        sys.stderr.write(str(e) + '\n')
348        sys.stderr.write("Usage: %s [-b branch] [path...]\n" % cmd)
349        sys.exit(1)
350
351    for opt, arg in opts:
352        if opt == '-b':
353            parent_branch = arg
354
355    if not parent_branch:
356        parent_branch = git_parent_branch(git_branch())
357
358    func = nits
359    if cmd == 'git-pbchk':
360        func = pbchk
361        if args:
362            sys.stderr.write("only complete workspaces may be pbchk'd\n");
363            sys.exit(1)
364
365    func(git_root(), parent_branch, args)
366
367if __name__ == '__main__':
368    try:
369        main(os.path.basename(sys.argv[0]), sys.argv[1:])
370    except GitError, e:
371        sys.stderr.write("failed to run git:\n %s\n" % str(e))
372        sys.exit(1)
373