usr/src/tools/scripts/git-pbchk.py

   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) 2015, 2016 by Delphix. All rights reserved.
  22 # Copyright 2016 Nexenta Systems, Inc.
  23 # Copyright 2018 Joyent, Inc.
  24 # Copyright 2018 OmniOS Community Edition (OmniOSce) Association.
  25 #
  26
  27 from __future__ import print_function
  28
  29 import getopt
  30 import io
  31 import os
  32 import re
  33 import subprocess
  34 import sys
  35 import tempfile
  36
  37 if sys.version_info[0] < 3:
  38     from cStringIO import StringIO
  39 else:
  40     from io import StringIO
  41
  42 #
  43 # Adjust the load path based on our location and the version of python into
  44 # which it is being loaded.  This assumes the normal onbld directory
  45 # structure, where we are in bin/ and the modules are in
  46 # lib/python(version)?/onbld/Scm/.  If that changes so too must this.
  47 #
  48 sys.path.insert(1, os.path.join(os.path.dirname(__file__), "..", "lib",
  49                                 "python%d.%d" % sys.version_info[:2]))
  50
  51 #
  52 # Add the relative path to usr/src/tools to the load path, such that when run
  53 # from the source tree we use the modules also within the source tree.
  54 #
  55 sys.path.insert(2, os.path.join(os.path.dirname(__file__), ".."))
  56
  57 from onbld.Scm import Ignore
  58 from onbld.Checks import Comments, Copyright, CStyle, HdrChk, WsCheck
  59 from onbld.Checks import Keywords, ManLint, Mapfile, SpellCheck
  60
  61
  62 class GitError(Exception):
  63     pass
  64
  65 def git(command):
  66     """Run a command and return a stream containing its stdout (and write its
  67     stderr to its stdout)"""
  68
  69     if type(command) != list:
  70         command = command.split()
  71
  72     command = ["git"] + command
  73
  74     try:
  75         tmpfile = tempfile.TemporaryFile(prefix="git-nits", mode="w+b")
  76     except EnvironmentError as e:
  77         raise GitError("Could not create temporary file: %s\n" % e)
  78
  79     try:
  80         p = subprocess.Popen(command,
  81                              stdout=tmpfile,
  82                              stderr=subprocess.PIPE)
  83     except OSError as e:
  84         raise GitError("could not execute %s: %s\n" % (command, e))
  85
  86     err = p.wait()
  87     if err != 0:
  88         raise GitError(p.stderr.read())
  89
  90     tmpfile.seek(0)
  91     lines = []
  92     for l in tmpfile:
  93         lines.append(l.decode('utf-8', 'replace'))
  94     return lines
  95
  96
  97 def git_root():
  98     """Return the root of the current git workspace"""
  99
 100     p = git('rev-parse --git-dir')
 101
 102     return os.path.abspath(os.path.join(dir, os.path.pardir))
 103
 104 def git_branch():
 105     """Return the current git branch"""
 106
 107     p = git('branch')
 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 def 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 def git_comments(parent):
 140     """Return a list of any checkin comments on this git branch"""
 141
 142     p = git('log --pretty=tformat:%%B:SEP: %s..' % parent)
 143
 144     if not p:
 145         sys.stderr.write("Failed getting git comments\n")
 146         sys.exit(err)
 147
 148     return [x.strip() for x in p if x != ':SEP:\n']
 149
 150 def git_file_list(parent, paths=None):
 151     """Return the set of files which have ever changed on this branch.
 152
 153     NB: This includes files which no longer exist, or no longer actually
 154     differ."""
 155
 156     p = git("log --name-only --pretty=format: %s.. %s" %
 157              (parent, ' '.join(paths)))
 158
 159     if not p:
 160         sys.stderr.write("Failed building file-list from git\n")
 161         sys.exit(err)
 162
 163     ret = set()
 164     for fname in p:
 165         if fname and not fname.isspace() and fname not in ret:
 166             ret.add(fname.strip())
 167
 168     return ret
 169
 170 def not_check(root, cmd):
 171     """Return a function which returns True if a file given as an argument
 172     should be excluded from the check named by 'cmd'"""
 173
 174     ignorefiles = list(filter(os.path.exists,
 175                          [os.path.join(root, ".git", "%s.NOT" % cmd),
 176                           os.path.join(root, "exception_lists", cmd)]))
 177     return Ignore.ignore(root, ignorefiles)
 178
 179 def gen_files(root, parent, paths, exclude):
 180     """Return a function producing file names, relative to the current
 181     directory, of any file changed on this branch (limited to 'paths' if
 182     requested), and excluding files for which exclude returns a true value """
 183
 184     # Taken entirely from Python 2.6's os.path.relpath which we would use if we
 185     # could.
 186     def relpath(path, here):
 187         c = os.path.abspath(os.path.join(root, path)).split(os.path.sep)
 188         s = os.path.abspath(here).split(os.path.sep)
 189         l = len(os.path.commonprefix((s, c)))
 190         return os.path.join(*[os.path.pardir] * (len(s)-l) + c[l:])
 191
 192     def ret(select=None):
 193         if not select:
 194             select = lambda x: True
 195
 196         for abspath in git_file_list(parent, paths):
 197             path = relpath(abspath, '.')
 198             try:
 199                 res = git("diff %s HEAD %s" % (parent, path))
 200             except GitError as e:
 201                 # This ignores all the errors that can be thrown. Usually, this
 202                 # means that git returned non-zero because the file doesn't
 203                 # exist, but it could also fail if git can't create a new file
 204                 # or it can't be executed.  Such errors are 1) unlikely, and 2)
 205                 # will be caught by other invocations of git().
 206                 continue
 207             empty = not res
 208             if (os.path.isfile(path) and not empty and
 209                 select(path) and not exclude(abspath)):
 210                 yield path
 211     return ret
 212
 213 def comchk(root, parent, flist, output):
 214     output.write("Comments:\n")
 215
 216     return Comments.comchk(git_comments(parent), check_db=True,
 217                            output=output)
 218
 219
 220 def mapfilechk(root, parent, flist, output):
 221     ret = 0
 222
 223     # We are interested in examining any file that has the following
 224     # in its final path segment:
 225     #    - Contains the word 'mapfile'
 226     #    - Begins with 'map.'
 227     #    - Ends with '.map'
 228     # We don't want to match unless these things occur in final path segment
 229     # because directory names with these strings don't indicate a mapfile.
 230     # We also ignore files with suffixes that tell us that the files
 231     # are not mapfiles.
 232     MapfileRE = re.compile(r'.*((mapfile[^/]*)|(/map\.+[^/]*)|(\.map))$',
 233         re.IGNORECASE)
 234     NotMapSuffixRE = re.compile(r'.*\.[ch]$', re.IGNORECASE)
 235
 236     output.write("Mapfile comments:\n")
 237
 238     for f in flist(lambda x: MapfileRE.match(x) and not
 239                    NotMapSuffixRE.match(x)):
 240         with io.open(f, encoding='utf-8', errors='replace') as fh:
 241             ret |= Mapfile.mapfilechk(fh, output=output)
 242     return ret
 243
 244
 245 def copyright(root, parent, flist, output):
 246     ret = 0
 247     output.write("Copyrights:\n")
 248     for f in flist():
 249         with io.open(f, encoding='utf-8', errors='replace') as fh:
 250             ret |= Copyright.copyright(fh, output=output)
 251     return ret
 252
 253 def 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         with io.open(f, encoding='utf-8', errors='replace') as fh:
 258             ret |= HdrChk.hdrchk(fh, lenient=True, output=output)
 259     return ret
 260
 261
 262 def 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         with io.open(f, encoding='utf-8', errors='replace') as fh:
 267             ret |= CStyle.cstyle(fh, output=output, picky=True,
 268                              check_posix_types=True,
 269                              check_continuation=True)
 270     return ret
 271
 272
 273 def manlint(root, parent, flist, output):
 274     ret = 0
 275     output.write("Man page format/spelling:\n")
 276     ManfileRE = re.compile(r'.*\.[0-9][a-z]*$', re.IGNORECASE)
 277     for f in flist(lambda x: ManfileRE.match(x)):
 278         with io.open(f, encoding='utf-8', errors='replace') as fh:
 279             ret |= ManLint.manlint(fh, output=output, picky=True)
 280             ret |= SpellCheck.spellcheck(fh, output=output)
 281     return ret
 282
 283 def keywords(root, parent, flist, output):
 284     ret = 0
 285     output.write("SCCS Keywords:\n")
 286     for f in flist():
 287         with io.open(f, encoding='utf-8', errors='replace') as fh:
 288             ret |= Keywords.keywords(fh, output=output)
 289     return ret
 290
 291 def wscheck(root, parent, flist, output):
 292     ret = 0
 293     output.write("white space nits:\n")
 294     for f in flist():
 295         with io.open(f, encoding='utf-8', errors='replace') as fh:
 296             ret |= WsCheck.wscheck(fh, output=output)
 297     return ret
 298
 299 def run_checks(root, parent, cmds, paths='', opts={}):
 300     """Run the checks given in 'cmds', expected to have well-known signatures,
 301     and report results for any which fail.
 302
 303     Return failure if any of them did.
 304
 305     NB: the function name of the commands passed in is used to name the NOT
 306     file which excepts files from them."""
 307
 308     ret = 0
 309
 310     for cmd in cmds:
 311         s = StringIO()
 312
 313         exclude = not_check(root, cmd.__name__)
 314         result = cmd(root, parent, gen_files(root, parent, paths, exclude),
 315                      output=s)
 316         ret |= result
 317
 318         if result != 0:
 319             print(s.getvalue())
 320
 321     return ret
 322
 323 def nits(root, parent, paths):
 324     cmds = [copyright,
 325             cstyle,
 326             hdrchk,
 327             keywords,
 328             manlint,
 329             mapfilechk,
 330             wscheck]
 331     run_checks(root, parent, cmds, paths)
 332
 333 def pbchk(root, parent, paths):
 334     cmds = [comchk,
 335             copyright,
 336             cstyle,
 337             hdrchk,
 338             keywords,
 339             manlint,
 340             mapfilechk,
 341             wscheck]
 342     run_checks(root, parent, cmds)
 343
 344 def main(cmd, args):
 345     parent_branch = None
 346     checkname = None
 347
 348     try:
 349         opts, args = getopt.getopt(args, 'b:p:')
 350     except getopt.GetoptError, e:
 351         sys.stderr.write(str(e) + '\n')
 352         sys.stderr.write("Usage: %s [-p branch] [path...]\n" % cmd)
 353         sys.exit(1)
 354
 355     for opt, arg in opts:
 356         # We accept "-b" as an alias of "-p" for backwards compatibility.
 357         if opt == '-p' or opt == '-b':
 358             parent_branch = arg
 359
 360     if not parent_branch:
 361         parent_branch = git_parent_branch(git_branch())
 362
 363     if checkname is None:
 364         if cmd == 'git-pbchk':
 365             checkname = 'pbchk'
 366
 367     if checkname == 'pbchk':
 368         if args:
 369             sys.stderr.write("only complete workspaces may be pbchk'd\n");
 370             sys.exit(1)
 371         pbchk(git_root(), parent_branch, None)
 372     else:
 373         run_checks(git_root(), parent_branch, [eval(checkname)], args)
 374
 375 if __name__ == '__main__':
 376     try:
 377         main(os.path.basename(sys.argv[0]), sys.argv[1:])
 378     except GitError as e:
 379         sys.stderr.write("failed to run git:\n %s\n" % str(e))
 380         sys.exit(1)