xref: /illumos-gate/usr/src/tools/scripts/git-pbchk.py (revision 5626beece2e5dedec7197ecf325cfaa1854a6c2e)
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# Copyright 2014 Garrett D'Amore <garrett@damore.org>
21# Copyright (c) 2014, Joyent, Inc.
22# Copyright (c) 2015 by Delphix. All rights reserved.
23#
24
25import getopt
26import os
27import re
28import subprocess
29import sys
30import tempfile
31
32from cStringIO import StringIO
33
34#
35# Adjust the load path based on our location and the version of python into
36# which it is being loaded.  This assumes the normal onbld directory
37# structure, where we are in bin/ and the modules are in
38# lib/python(version)?/onbld/Scm/.  If that changes so too must this.
39#
40sys.path.insert(1, os.path.join(os.path.dirname(__file__), "..", "lib",
41                                "python%d.%d" % sys.version_info[:2]))
42
43#
44# Add the relative path to usr/src/tools to the load path, such that when run
45# from the source tree we use the modules also within the source tree.
46#
47sys.path.insert(2, os.path.join(os.path.dirname(__file__), ".."))
48
49from onbld.Scm import Ignore
50from onbld.Checks import Comments, Copyright, CStyle, HdrChk
51from onbld.Checks import JStyle, Keywords, ManLint, 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        res = git("diff %s HEAD %s" % (parent, fname))
168        empty = not res.readline()
169        if fname and not fname.isspace() and fname not in ret and not empty:
170            ret.add(fname.strip())
171
172    return ret
173
174
175def not_check(root, cmd):
176    """Return a function which returns True if a file given as an argument
177    should be excluded from the check named by 'cmd'"""
178
179    ignorefiles = filter(os.path.exists,
180                         [os.path.join(root, ".git", "%s.NOT" % cmd),
181                          os.path.join(root, "exception_lists", cmd)])
182    return Ignore.ignore(root, ignorefiles)
183
184
185def gen_files(root, parent, paths, exclude):
186    """Return a function producing file names, relative to the current
187    directory, of any file changed on this branch (limited to 'paths' if
188    requested), and excluding files for which exclude returns a true value """
189
190    # Taken entirely from Python 2.6's os.path.relpath which we would use if we
191    # could.
192    def relpath(path, here):
193        c = os.path.abspath(os.path.join(root, path)).split(os.path.sep)
194        s = os.path.abspath(here).split(os.path.sep)
195        l = len(os.path.commonprefix((s, c)))
196        return os.path.join(*[os.path.pardir] * (len(s)-l) + c[l:])
197
198    def ret(select=None):
199        if not select:
200            select = lambda x: True
201
202        for f in git_file_list(parent, paths):
203            f = relpath(f, '.')
204            if (os.path.exists(f) and select(f) and not exclude(f)):
205                yield f
206    return ret
207
208
209def comchk(root, parent, flist, output):
210    output.write("Comments:\n")
211
212    return Comments.comchk(git_comments(parent), check_db=True,
213                           output=output)
214
215
216def mapfilechk(root, parent, flist, output):
217    ret = 0
218
219    # We are interested in examining any file that has the following
220    # in its final path segment:
221    #    - Contains the word 'mapfile'
222    #    - Begins with 'map.'
223    #    - Ends with '.map'
224    # We don't want to match unless these things occur in final path segment
225    # because directory names with these strings don't indicate a mapfile.
226    # We also ignore files with suffixes that tell us that the files
227    # are not mapfiles.
228    MapfileRE = re.compile(r'.*((mapfile[^/]*)|(/map\.+[^/]*)|(\.map))$',
229        re.IGNORECASE)
230    NotMapSuffixRE = re.compile(r'.*\.[ch]$', re.IGNORECASE)
231
232    output.write("Mapfile comments:\n")
233
234    for f in flist(lambda x: MapfileRE.match(x) and not
235                   NotMapSuffixRE.match(x)):
236        fh = open(f, 'r')
237        ret |= Mapfile.mapfilechk(fh, output=output)
238        fh.close()
239    return ret
240
241
242def copyright(root, parent, flist, output):
243    ret = 0
244    output.write("Copyrights:\n")
245    for f in flist():
246        fh = open(f, 'r')
247        ret |= Copyright.copyright(fh, output=output)
248        fh.close()
249    return ret
250
251
252def hdrchk(root, parent, flist, output):
253    ret = 0
254    output.write("Header format:\n")
255    for f in flist(lambda x: x.endswith('.h')):
256        fh = open(f, 'r')
257        ret |= HdrChk.hdrchk(fh, lenient=True, output=output)
258        fh.close()
259    return ret
260
261
262def cstyle(root, parent, flist, output):
263    ret = 0
264    output.write("C style:\n")
265    for f in flist(lambda x: x.endswith('.c') or x.endswith('.h')):
266        fh = open(f, 'r')
267        ret |= CStyle.cstyle(fh, output=output, picky=True,
268                             check_posix_types=True,
269                             check_continuation=True)
270        fh.close()
271    return ret
272
273
274def jstyle(root, parent, flist, output):
275    ret = 0
276    output.write("Java style:\n")
277    for f in flist(lambda x: x.endswith('.java')):
278        fh = open(f, 'r')
279        ret |= JStyle.jstyle(fh, output=output, picky=True)
280        fh.close()
281    return ret
282
283
284def manlint(root, parent, flist, output):
285    ret = 0
286    output.write("Man page format:\n")
287    ManfileRE = re.compile(r'.*\.[0-9][a-z]*$', re.IGNORECASE)
288    for f in flist(lambda x: ManfileRE.match(x)):
289        fh = open(f, 'r')
290        ret |= ManLint.manlint(fh, output=output, picky=True)
291	fh.close()
292    return ret
293
294def keywords(root, parent, flist, output):
295    ret = 0
296    output.write("SCCS Keywords:\n")
297    for f in flist():
298        fh = open(f, 'r')
299        ret |= Keywords.keywords(fh, output=output)
300        fh.close()
301    return ret
302
303
304def run_checks(root, parent, cmds, paths='', opts={}):
305    """Run the checks given in 'cmds', expected to have well-known signatures,
306    and report results for any which fail.
307
308    Return failure if any of them did.
309
310    NB: the function name of the commands passed in is used to name the NOT
311    file which excepts files from them."""
312
313    ret = 0
314
315    for cmd in cmds:
316        s = StringIO()
317
318        exclude = not_check(root, cmd.func_name)
319        result = cmd(root, parent, gen_files(root, parent, paths, exclude),
320                     output=s)
321        ret |= result
322
323        if result != 0:
324            print s.getvalue()
325
326    return ret
327
328
329def nits(root, parent, paths):
330    cmds = [copyright,
331            cstyle,
332            hdrchk,
333            jstyle,
334            keywords,
335	    manlint,
336            mapfilechk]
337    run_checks(root, parent, cmds, paths)
338
339
340def pbchk(root, parent, paths):
341    cmds = [comchk,
342            copyright,
343            cstyle,
344            hdrchk,
345            jstyle,
346            keywords,
347	    manlint,
348            mapfilechk]
349    run_checks(root, parent, cmds)
350
351
352def main(cmd, args):
353    parent_branch = None
354
355    try:
356        opts, args = getopt.getopt(args, 'b:')
357    except getopt.GetoptError, e:
358        sys.stderr.write(str(e) + '\n')
359        sys.stderr.write("Usage: %s [-b branch] [path...]\n" % cmd)
360        sys.exit(1)
361
362    for opt, arg in opts:
363        if opt == '-b':
364            parent_branch = arg
365
366    if not parent_branch:
367        parent_branch = git_parent_branch(git_branch())
368
369    func = nits
370    if cmd == 'git-pbchk':
371        func = pbchk
372        if args:
373            sys.stderr.write("only complete workspaces may be pbchk'd\n");
374            sys.exit(1)
375
376    func(git_root(), parent_branch, args)
377
378if __name__ == '__main__':
379    try:
380        main(os.path.basename(sys.argv[0]), sys.argv[1:])
381    except GitError, e:
382        sys.stderr.write("failed to run git:\n %s\n" % str(e))
383        sys.exit(1)
384