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 #
  25
  26 import getopt
  27 import os
  28 import re
  29 import subprocess
  30 import sys
  31 import tempfile
  32
  33 from 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 #
  41 sys.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 #
  48 sys.path.insert(2, os.path.join(os.path.dirname(__file__), ".."))
  49
  50 from onbld.Scm import Ignore
  51 from onbld.Checks import Comments, Copyright, CStyle, HdrChk, WsCheck
  52 from onbld.Checks import Keywords, ManLint, Mapfile, SpellCheck
  53
  54
  55 class GitError(Exception):
  56     pass
  57
  58 def 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
  87 def 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
 100 def 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
 116 def 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
 141 def 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
 153 def 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
 174 def 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
 184 def 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 abspath in git_file_list(parent, paths):
 202             path = relpath(abspath, '.')
 203             try:
 204                 res = git("diff %s HEAD %s" % (parent, path))
 205             except GitError, e:
 206                 # This ignores all the errors that can be thrown. Usually, this
 207                 # means that git returned non-zero because the file doesn't
 208                 # exist, but it could also fail if git can't create a new file
 209                 # or it can't be executed.  Such errors are 1) unlikely, and 2)
 210                 # will be caught by other invocations of git().
 211                 continue
 212             empty = not res.readline()
 213             if (os.path.isfile(path) and not empty and
 214                 select(path) and not exclude(abspath)):
 215                 yield path
 216     return ret
 217
 218
 219 def comchk(root, parent, flist, output):
 220     output.write("Comments:\n")
 221
 222     return Comments.comchk(git_comments(parent), check_db=True,
 223                            output=output)
 224
 225
 226 def mapfilechk(root, parent, flist, output):
 227     ret = 0
 228
 229     # We are interested in examining any file that has the following
 230     # in its final path segment:
 231     #    - Contains the word 'mapfile'
 232     #    - Begins with 'map.'
 233     #    - Ends with '.map'
 234     # We don't want to match unless these things occur in final path segment
 235     # because directory names with these strings don't indicate a mapfile.
 236     # We also ignore files with suffixes that tell us that the files
 237     # are not mapfiles.
 238     MapfileRE = re.compile(r'.*((mapfile[^/]*)|(/map\.+[^/]*)|(\.map))$',
 239         re.IGNORECASE)
 240     NotMapSuffixRE = re.compile(r'.*\.[ch]$', re.IGNORECASE)
 241
 242     output.write("Mapfile comments:\n")
 243
 244     for f in flist(lambda x: MapfileRE.match(x) and not
 245                    NotMapSuffixRE.match(x)):
 246         fh = open(f, 'r')
 247         ret |= Mapfile.mapfilechk(fh, output=output)
 248         fh.close()
 249     return ret
 250
 251
 252 def copyright(root, parent, flist, output):
 253     ret = 0
 254     output.write("Copyrights:\n")
 255     for f in flist():
 256         fh = open(f, 'r')
 257         ret |= Copyright.copyright(fh, output=output)
 258         fh.close()
 259     return ret
 260
 261
 262 def hdrchk(root, parent, flist, output):
 263     ret = 0
 264     output.write("Header format:\n")
 265     for f in flist(lambda x: x.endswith('.h')):
 266         fh = open(f, 'r')
 267         ret |= HdrChk.hdrchk(fh, lenient=True, output=output)
 268         fh.close()
 269     return ret
 270
 271
 272 def cstyle(root, parent, flist, output):
 273     ret = 0
 274     output.write("C style:\n")
 275     for f in flist(lambda x: x.endswith('.c') or x.endswith('.h')):
 276         fh = open(f, 'r')
 277         ret |= CStyle.cstyle(fh, output=output, picky=True,
 278                              check_posix_types=True,
 279                              check_continuation=True)
 280         fh.close()
 281     return ret
 282
 283
 284 def manlint(root, parent, flist, output):
 285     ret = 0
 286     output.write("Man page format/spelling:\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         ret |= SpellCheck.spellcheck(fh, output=output)
 292         fh.close()
 293     return ret
 294
 295 def keywords(root, parent, flist, output):
 296     ret = 0
 297     output.write("SCCS Keywords:\n")
 298     for f in flist():
 299         fh = open(f, 'r')
 300         ret |= Keywords.keywords(fh, output=output)
 301         fh.close()
 302     return ret
 303
 304 def wscheck(root, parent, flist, output):
 305     ret = 0
 306     output.write("white space nits:\n")
 307     for f in flist():
 308         fh = open(f, 'r')
 309         ret |= WsCheck.wscheck(fh, output=output)
 310         fh.close()
 311     return ret
 312
 313 def run_checks(root, parent, cmds, paths='', opts={}):
 314     """Run the checks given in 'cmds', expected to have well-known signatures,
 315     and report results for any which fail.
 316
 317     Return failure if any of them did.
 318
 319     NB: the function name of the commands passed in is used to name the NOT
 320     file which excepts files from them."""
 321
 322     ret = 0
 323
 324     for cmd in cmds:
 325         s = StringIO()
 326
 327         exclude = not_check(root, cmd.func_name)
 328         result = cmd(root, parent, gen_files(root, parent, paths, exclude),
 329                      output=s)
 330         ret |= result
 331
 332         if result != 0:
 333             print s.getvalue()
 334
 335     return ret
 336
 337
 338 def nits(root, parent, paths):
 339     cmds = [copyright,
 340             cstyle,
 341             hdrchk,
 342             keywords,
 343             manlint,
 344             mapfilechk,
 345             wscheck]
 346     run_checks(root, parent, cmds, paths)
 347
 348
 349 def pbchk(root, parent, paths):
 350     cmds = [comchk,
 351             copyright,
 352             cstyle,
 353             hdrchk,
 354             keywords,
 355             manlint,
 356             mapfilechk,
 357             wscheck]
 358     run_checks(root, parent, cmds)
 359
 360
 361 def main(cmd, args):
 362     parent_branch = None
 363     checkname = None
 364
 365     try:
 366         opts, args = getopt.getopt(args, 'b:p:')
 367     except getopt.GetoptError, e:
 368         sys.stderr.write(str(e) + '\n')
 369         sys.stderr.write("Usage: %s [-p branch] [path...]\n" % cmd)
 370         sys.exit(1)
 371
 372     for opt, arg in opts:
 373         # We accept "-b" as an alias of "-p" for backwards compatibility.
 374         if opt == '-p' or opt == '-b':
 375             parent_branch = arg
 376
 377     if not parent_branch:
 378         parent_branch = git_parent_branch(git_branch())
 379
 380     if checkname is None:
 381         if cmd == 'git-pbchk':
 382             checkname = 'pbchk'
 383
 384     if checkname == 'pbchk':
 385         if args:
 386             sys.stderr.write("only complete workspaces may be pbchk'd\n");
 387             sys.exit(1)
 388         pbchk(git_root(), parent_branch, None)
 389     else:
 390         run_checks(git_root(), parent_branch, [eval(checkname)], args)
 391
 392 if __name__ == '__main__':
 393     try:
 394         main(os.path.basename(sys.argv[0]), sys.argv[1:])
 395     except GitError, e:
 396         sys.stderr.write("failed to run git:\n %s\n" % str(e))
 397         sys.exit(1)