gdb/copyright.py

   1 #! /usr/bin/env python3
   2
   3 # Copyright (C) 2011-2020 Free Software Foundation, Inc.
   4 #
   5 # This file is part of GDB.
   6 #
   7 # This program is free software; you can redistribute it and/or modify
   8 # it under the terms of the GNU General Public License as published by
   9 # the Free Software Foundation; either version 3 of the License, or
  10 # (at your option) any later version.
  11 #
  12 # This program is distributed in the hope that it will be useful,
  13 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15 # GNU General Public License for more details.
  16 #
  17 # You should have received a copy of the GNU General Public License
  18 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
  19
  20 """copyright.py
  21
  22 This script updates the list of years in the copyright notices in
  23 most files maintained by the GDB project.
  24
  25 Usage: cd src/gdb && python copyright.py
  26
  27 Always review the output of this script before committing it!
  28 A useful command to review the output is:
  29     % filterdiff -x \*.c -x \*.cc -x \*.h -x \*.exp updates.diff
  30 This removes the bulk of the changes which are most likely to be correct.
  31 """
  32
  33 import datetime
  34 import locale
  35 import os
  36 import os.path
  37 import subprocess
  38 import sys
  39
  40
  41 def get_update_list():
  42     """Return the list of files to update.
  43
  44     Assumes that the current working directory when called is the root
  45     of the GDB source tree (NOT the gdb/ subdirectory!).  The names of
  46     the files are relative to that root directory.
  47     """
  48     result = []
  49     for gdb_dir in ('gdb', 'gnulib', 'sim', 'include/gdb'):
  50         for root, dirs, files in os.walk(gdb_dir, topdown=True):
  51             for dirname in dirs:
  52                 reldirname = "%s/%s" % (root, dirname)
  53                 if (dirname in EXCLUDE_ALL_LIST
  54                     or reldirname in EXCLUDE_LIST
  55                     or reldirname in NOT_FSF_LIST
  56                     or reldirname in BY_HAND):
  57                     # Prune this directory from our search list.
  58                     dirs.remove(dirname)
  59             for filename in files:
  60                 relpath = "%s/%s" % (root, filename)
  61                 if (filename in EXCLUDE_ALL_LIST
  62                     or relpath in EXCLUDE_LIST
  63                     or relpath in NOT_FSF_LIST
  64                     or relpath in BY_HAND):
  65                     # Ignore this file.
  66                     pass
  67                 else:
  68                     result.append(relpath)
  69     return result
  70
  71
  72 def update_files(update_list):
  73     """Update the copyright header of the files in the given list.
  74
  75     We use gnulib's update-copyright script for that.
  76     """
  77     # We want to use year intervals in the copyright notices, and
  78     # all years should be collapsed to one single year interval,
  79     # even if there are "holes" in the list of years found in the
  80     # original copyright notice (OK'ed by the FSF, case [gnu.org #719834]).
  81     os.environ['UPDATE_COPYRIGHT_USE_INTERVALS'] = '2'
  82
  83     # Perform the update, and save the output in a string.
  84     update_cmd = ['bash', 'gnulib/import/extra/update-copyright']
  85     update_cmd += update_list
  86
  87     p = subprocess.Popen(update_cmd, stdout=subprocess.PIPE,
  88                          stderr=subprocess.STDOUT,
  89                          encoding=locale.getpreferredencoding())
  90     update_out = p.communicate()[0]
  91
  92     # Process the output.  Typically, a lot of files do not have
  93     # a copyright notice :-(.  The update-copyright script prints
  94     # a well defined warning when it did not find the copyright notice.
  95     # For each of those, do a sanity check and see if they may in fact
  96     # have one.  For the files that are found not to have one, we filter
  97     # the line out from the output, since there is nothing more to do,
  98     # short of looking at each file and seeing which notice is appropriate.
  99     # Too much work! (~4,000 files listed as of 2012-01-03).
 100     update_out = update_out.splitlines(keepends=False)
 101     warning_string = ': warning: copyright statement not found'
 102     warning_len = len(warning_string)
 103
 104     for line in update_out:
 105         if line.endswith(warning_string):
 106             filename = line[:-warning_len]
 107             if may_have_copyright_notice(filename):
 108                 print(line)
 109         else:
 110             # Unrecognized file format. !?!
 111             print("*** " + line)
 112
 113
 114 def may_have_copyright_notice(filename):
 115     """Check that the given file does not seem to have a copyright notice.
 116
 117     The filename is relative to the root directory.
 118     This function assumes that the current working directory is that root
 119     directory.
 120
 121     The algorigthm is fairly crude, meaning that it might return
 122     some false positives.  I do not think it will return any false
 123     negatives...  We might improve this function to handle more
 124     complex cases later...
 125     """
 126     # For now, it may have a copyright notice if we find the word
 127     # "Copyright" at the (reasonable) start of the given file, say
 128     # 50 lines...
 129     MAX_LINES = 50
 130
 131     # We don't really know what encoding each file might be following,
 132     # so just open the file as a byte stream. We only need to search
 133     # for a pattern that should be the same regardless of encoding,
 134     # so that should be good enough.
 135     fd = open(filename, 'rb')
 136
 137     lineno = 1
 138     for line in fd:
 139         if b'Copyright' in line:
 140             return True
 141         lineno += 1
 142         if lineno > 50:
 143             return False
 144     return False
 145
 146
 147 def main ():
 148     """The main subprogram."""
 149     root_dir = os.path.dirname(os.getcwd())
 150     os.chdir(root_dir)
 151
 152     if not (os.path.isdir('gdb') and
 153             os.path.isfile("gnulib/import/extra/update-copyright")):
 154         print("Error: This script must be called from the gdb directory.")
 155         sys.exit(1)
 156
 157     update_list = get_update_list()
 158     update_files (update_list)
 159
 160     # Remind the user that some files need to be updated by HAND...
 161
 162     if MULTIPLE_COPYRIGHT_HEADERS:
 163         print()
 164         print("\033[31m"
 165               "REMINDER: Multiple copyright headers must be updated by hand:"
 166               "\033[0m")
 167         for filename in MULTIPLE_COPYRIGHT_HEADERS:
 168             print("  ", filename)
 169
 170     if BY_HAND:
 171         print()
 172         print("\033[31mREMINDER: The following files must be updated by hand." \
 173               "\033[0m")
 174         for filename in BY_HAND:
 175             print("  ", filename)
 176
 177 ############################################################################
 178 #
 179 # Some constants, placed at the end because they take up a lot of room.
 180 # The actual value of these constants is not significant to the understanding
 181 # of the script.
 182 #
 183 ############################################################################
 184
 185 # Files which should not be modified, either because they are
 186 # generated, non-FSF, or otherwise special (e.g. license text,
 187 # or test cases which must be sensitive to line numbering).
 188 #
 189 # Filenames are relative to the root directory.
 190 EXCLUDE_LIST = (
 191     'gdb/nat/glibc_thread_db.h',
 192     'gdb/CONTRIBUTE',
 193     'gnulib/import'
 194 )
 195
 196 # Files which should not be modified, either because they are
 197 # generated, non-FSF, or otherwise special (e.g. license text,
 198 # or test cases which must be sensitive to line numbering).
 199 #
 200 # Matches any file or directory name anywhere.  Use with caution.
 201 # This is mostly for files that can be found in multiple directories.
 202 # Eg: We want all files named COPYING to be left untouched.
 203
 204 EXCLUDE_ALL_LIST = (
 205     "COPYING", "COPYING.LIB", "CVS", "configure", "copying.c",
 206     "fdl.texi", "gpl.texi", "aclocal.m4",
 207 )
 208
 209 # The list of files to update by hand.
 210 BY_HAND = (
 211     # Nothing at the moment :-).
 212 )
 213
 214 # Files containing multiple copyright headers.  This script is only
 215 # fixing the first one it finds, so we need to finish the update
 216 # by hand.
 217 MULTIPLE_COPYRIGHT_HEADERS = (
 218     "gdb/doc/gdb.texinfo",
 219     "gdb/doc/refcard.tex",
 220     "gdb/gdbarch.sh",
 221 )
 222
 223 # The list of file which have a copyright, but not head by the FSF.
 224 # Filenames are relative to the root directory.
 225 NOT_FSF_LIST = (
 226     "gdb/exc_request.defs",
 227     "gdb/gdbtk",
 228     "gdb/testsuite/gdb.gdbtk/",
 229     "sim/arm/armemu.h", "sim/arm/armos.c", "sim/arm/gdbhost.c",
 230     "sim/arm/dbg_hif.h", "sim/arm/dbg_conf.h", "sim/arm/communicate.h",
 231     "sim/arm/armos.h", "sim/arm/armcopro.c", "sim/arm/armemu.c",
 232     "sim/arm/kid.c", "sim/arm/thumbemu.c", "sim/arm/armdefs.h",
 233     "sim/arm/armopts.h", "sim/arm/dbg_cp.h", "sim/arm/dbg_rdi.h",
 234     "sim/arm/parent.c", "sim/arm/armsupp.c", "sim/arm/armrdi.c",
 235     "sim/arm/bag.c", "sim/arm/armvirt.c", "sim/arm/main.c", "sim/arm/bag.h",
 236     "sim/arm/communicate.c", "sim/arm/gdbhost.h", "sim/arm/armfpe.h",
 237     "sim/arm/arminit.c",
 238     "sim/common/cgen-fpu.c", "sim/common/cgen-fpu.h",
 239     "sim/common/cgen-accfp.c",
 240     "sim/mips/m16run.c", "sim/mips/sim-main.c",
 241     "sim/moxie/moxie-gdb.dts",
 242     # Not a single file in sim/ppc/ appears to be copyright FSF :-(.
 243     "sim/ppc/filter.h", "sim/ppc/gen-support.h", "sim/ppc/ld-insn.h",
 244     "sim/ppc/hw_sem.c", "sim/ppc/hw_disk.c", "sim/ppc/idecode_branch.h",
 245     "sim/ppc/sim-endian.h", "sim/ppc/table.c", "sim/ppc/hw_core.c",
 246     "sim/ppc/gen-support.c", "sim/ppc/gen-semantics.h", "sim/ppc/cpu.h",
 247     "sim/ppc/sim_callbacks.h", "sim/ppc/RUN", "sim/ppc/Makefile.in",
 248     "sim/ppc/emul_chirp.c", "sim/ppc/hw_nvram.c", "sim/ppc/dc-test.01",
 249     "sim/ppc/hw_phb.c", "sim/ppc/hw_eeprom.c", "sim/ppc/bits.h",
 250     "sim/ppc/hw_vm.c", "sim/ppc/cap.h", "sim/ppc/os_emul.h",
 251     "sim/ppc/options.h", "sim/ppc/gen-idecode.c", "sim/ppc/filter.c",
 252     "sim/ppc/corefile-n.h", "sim/ppc/std-config.h", "sim/ppc/ld-decode.h",
 253     "sim/ppc/filter_filename.h", "sim/ppc/hw_shm.c",
 254     "sim/ppc/pk_disklabel.c", "sim/ppc/dc-simple", "sim/ppc/misc.h",
 255     "sim/ppc/device_table.h", "sim/ppc/ld-insn.c", "sim/ppc/inline.c",
 256     "sim/ppc/emul_bugapi.h", "sim/ppc/hw_cpu.h", "sim/ppc/debug.h",
 257     "sim/ppc/hw_ide.c", "sim/ppc/debug.c", "sim/ppc/gen-itable.h",
 258     "sim/ppc/interrupts.c", "sim/ppc/hw_glue.c", "sim/ppc/emul_unix.c",
 259     "sim/ppc/sim_calls.c", "sim/ppc/dc-complex", "sim/ppc/ld-cache.c",
 260     "sim/ppc/registers.h", "sim/ppc/dc-test.02", "sim/ppc/options.c",
 261     "sim/ppc/igen.h", "sim/ppc/registers.c", "sim/ppc/device.h",
 262     "sim/ppc/emul_chirp.h", "sim/ppc/hw_register.c", "sim/ppc/hw_init.c",
 263     "sim/ppc/sim-endian-n.h", "sim/ppc/filter_filename.c",
 264     "sim/ppc/bits.c", "sim/ppc/idecode_fields.h", "sim/ppc/hw_memory.c",
 265     "sim/ppc/misc.c", "sim/ppc/double.c", "sim/ppc/psim.h",
 266     "sim/ppc/hw_trace.c", "sim/ppc/emul_netbsd.h", "sim/ppc/psim.c",
 267     "sim/ppc/ppc-instructions", "sim/ppc/tree.h", "sim/ppc/README",
 268     "sim/ppc/gen-icache.h", "sim/ppc/gen-model.h", "sim/ppc/ld-cache.h",
 269     "sim/ppc/mon.c", "sim/ppc/corefile.h", "sim/ppc/vm.c",
 270     "sim/ppc/INSTALL", "sim/ppc/gen-model.c", "sim/ppc/hw_cpu.c",
 271     "sim/ppc/corefile.c", "sim/ppc/hw_opic.c", "sim/ppc/gen-icache.c",
 272     "sim/ppc/events.h", "sim/ppc/os_emul.c", "sim/ppc/emul_generic.c",
 273     "sim/ppc/main.c", "sim/ppc/hw_com.c", "sim/ppc/gen-semantics.c",
 274     "sim/ppc/emul_bugapi.c", "sim/ppc/device.c", "sim/ppc/emul_generic.h",
 275     "sim/ppc/tree.c", "sim/ppc/mon.h", "sim/ppc/interrupts.h",
 276     "sim/ppc/cap.c", "sim/ppc/cpu.c", "sim/ppc/hw_phb.h",
 277     "sim/ppc/device_table.c", "sim/ppc/lf.c", "sim/ppc/lf.c",
 278     "sim/ppc/dc-stupid", "sim/ppc/hw_pal.c", "sim/ppc/ppc-spr-table",
 279     "sim/ppc/emul_unix.h", "sim/ppc/words.h", "sim/ppc/basics.h",
 280     "sim/ppc/hw_htab.c", "sim/ppc/lf.h", "sim/ppc/ld-decode.c",
 281     "sim/ppc/sim-endian.c", "sim/ppc/gen-itable.c",
 282     "sim/ppc/idecode_expression.h", "sim/ppc/table.h", "sim/ppc/dgen.c",
 283     "sim/ppc/events.c", "sim/ppc/gen-idecode.h", "sim/ppc/emul_netbsd.c",
 284     "sim/ppc/igen.c", "sim/ppc/vm_n.h", "sim/ppc/vm.h",
 285     "sim/ppc/hw_iobus.c", "sim/ppc/inline.h",
 286     "sim/testsuite/sim/bfin/s21.s", "sim/testsuite/sim/mips/mips32-dsp2.s",
 287 )
 288
 289 if __name__ == "__main__":
 290     main()
 291