gdb/copyright.py

   1 #! /usr/bin/env python3
   2
   3 # Copyright (C) 2011-2021 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 (
  50         'gdb', 'gdbserver', 'gdbsupport', 'gnulib', 'sim', 'include/gdb',
  51     ):
  52         for root, dirs, files in os.walk(gdb_dir, topdown=True):
  53             for dirname in dirs:
  54                 reldirname = "%s/%s" % (root, dirname)
  55                 if (dirname in EXCLUDE_ALL_LIST
  56                     or reldirname in EXCLUDE_LIST
  57                     or reldirname in NOT_FSF_LIST
  58                     or reldirname in BY_HAND):
  59                     # Prune this directory from our search list.
  60                     dirs.remove(dirname)
  61             for filename in files:
  62                 relpath = "%s/%s" % (root, filename)
  63                 if (filename in EXCLUDE_ALL_LIST
  64                     or relpath in EXCLUDE_LIST
  65                     or relpath in NOT_FSF_LIST
  66                     or relpath in BY_HAND):
  67                     # Ignore this file.
  68                     pass
  69                 else:
  70                     result.append(relpath)
  71     return result
  72
  73
  74 def update_files(update_list):
  75     """Update the copyright header of the files in the given list.
  76
  77     We use gnulib's update-copyright script for that.
  78     """
  79     # We want to use year intervals in the copyright notices, and
  80     # all years should be collapsed to one single year interval,
  81     # even if there are "holes" in the list of years found in the
  82     # original copyright notice (OK'ed by the FSF, case [gnu.org #719834]).
  83     os.environ['UPDATE_COPYRIGHT_USE_INTERVALS'] = '2'
  84
  85     # Perform the update, and save the output in a string.
  86     update_cmd = ['bash', 'gnulib/import/extra/update-copyright']
  87     update_cmd += update_list
  88
  89     p = subprocess.Popen(update_cmd, stdout=subprocess.PIPE,
  90                          stderr=subprocess.STDOUT,
  91                          encoding=locale.getpreferredencoding())
  92     update_out = p.communicate()[0]
  93
  94     # Process the output.  Typically, a lot of files do not have
  95     # a copyright notice :-(.  The update-copyright script prints
  96     # a well defined warning when it did not find the copyright notice.
  97     # For each of those, do a sanity check and see if they may in fact
  98     # have one.  For the files that are found not to have one, we filter
  99     # the line out from the output, since there is nothing more to do,
 100     # short of looking at each file and seeing which notice is appropriate.
 101     # Too much work! (~4,000 files listed as of 2012-01-03).
 102     update_out = update_out.splitlines(keepends=False)
 103     warning_string = ': warning: copyright statement not found'
 104     warning_len = len(warning_string)
 105
 106     for line in update_out:
 107         if line.endswith(warning_string):
 108             filename = line[:-warning_len]
 109             if may_have_copyright_notice(filename):
 110                 print(line)
 111         else:
 112             # Unrecognized file format. !?!
 113             print("*** " + line)
 114
 115
 116 def may_have_copyright_notice(filename):
 117     """Check that the given file does not seem to have a copyright notice.
 118
 119     The filename is relative to the root directory.
 120     This function assumes that the current working directory is that root
 121     directory.
 122
 123     The algorigthm is fairly crude, meaning that it might return
 124     some false positives.  I do not think it will return any false
 125     negatives...  We might improve this function to handle more
 126     complex cases later...
 127     """
 128     # For now, it may have a copyright notice if we find the word
 129     # "Copyright" at the (reasonable) start of the given file, say
 130     # 50 lines...
 131     MAX_LINES = 50
 132
 133     # We don't really know what encoding each file might be following,
 134     # so just open the file as a byte stream. We only need to search
 135     # for a pattern that should be the same regardless of encoding,
 136     # so that should be good enough.
 137     fd = open(filename, 'rb')
 138
 139     lineno = 1
 140     for line in fd:
 141         if b'Copyright' in line:
 142             return True
 143         lineno += 1
 144         if lineno > 50:
 145             return False
 146     return False
 147
 148
 149 def main ():
 150     """The main subprogram."""
 151     root_dir = os.path.dirname(os.getcwd())
 152     os.chdir(root_dir)
 153
 154     if not (os.path.isdir('gdb') and
 155             os.path.isfile("gnulib/import/extra/update-copyright")):
 156         print("Error: This script must be called from the gdb directory.")
 157         sys.exit(1)
 158
 159     update_list = get_update_list()
 160     update_files (update_list)
 161
 162     # Remind the user that some files need to be updated by HAND...
 163
 164     if MULTIPLE_COPYRIGHT_HEADERS:
 165         print()
 166         print("\033[31m"
 167               "REMINDER: Multiple copyright headers must be updated by hand:"
 168               "\033[0m")
 169         for filename in MULTIPLE_COPYRIGHT_HEADERS:
 170             print("  ", filename)
 171
 172     if BY_HAND:
 173         print()
 174         print("\033[31mREMINDER: The following files must be updated by hand." \
 175               "\033[0m")
 176         for filename in BY_HAND:
 177             print("  ", filename)
 178
 179 ############################################################################
 180 #
 181 # Some constants, placed at the end because they take up a lot of room.
 182 # The actual value of these constants is not significant to the understanding
 183 # of the script.
 184 #
 185 ############################################################################
 186
 187 # Files which should not be modified, either because they are
 188 # generated, non-FSF, or otherwise special (e.g. license text,
 189 # or test cases which must be sensitive to line numbering).
 190 #
 191 # Filenames are relative to the root directory.
 192 EXCLUDE_LIST = (
 193     'gdb/nat/glibc_thread_db.h',
 194     'gdb/CONTRIBUTE',
 195     'gnulib/import',
 196     'gnulib/config.in',
 197     'gnulib/Makefile.in',
 198 )
 199
 200 # Files which should not be modified, either because they are
 201 # generated, non-FSF, or otherwise special (e.g. license text,
 202 # or test cases which must be sensitive to line numbering).
 203 #
 204 # Matches any file or directory name anywhere.  Use with caution.
 205 # This is mostly for files that can be found in multiple directories.
 206 # Eg: We want all files named COPYING to be left untouched.
 207
 208 EXCLUDE_ALL_LIST = (
 209     "COPYING", "COPYING.LIB", "CVS", "configure", "copying.c",
 210     "fdl.texi", "gpl.texi", "aclocal.m4",
 211 )
 212
 213 # The list of files to update by hand.
 214 BY_HAND = (
 215     # Nothing at the moment :-).
 216 )
 217
 218 # Files containing multiple copyright headers.  This script is only
 219 # fixing the first one it finds, so we need to finish the update
 220 # by hand.
 221 MULTIPLE_COPYRIGHT_HEADERS = (
 222     "gdb/doc/gdb.texinfo",
 223     "gdb/doc/refcard.tex",
 224     "gdb/gdbarch.sh",
 225 )
 226
 227 # The list of file which have a copyright, but not head by the FSF.
 228 # Filenames are relative to the root directory.
 229 NOT_FSF_LIST = (
 230     "gdb/exc_request.defs",
 231     "gdb/gdbtk",
 232     "gdb/testsuite/gdb.gdbtk/",
 233     "sim/arm/armemu.h", "sim/arm/armos.c", "sim/arm/gdbhost.c",
 234     "sim/arm/dbg_hif.h", "sim/arm/dbg_conf.h", "sim/arm/communicate.h",
 235     "sim/arm/armos.h", "sim/arm/armcopro.c", "sim/arm/armemu.c",
 236     "sim/arm/kid.c", "sim/arm/thumbemu.c", "sim/arm/armdefs.h",
 237     "sim/arm/armopts.h", "sim/arm/dbg_cp.h", "sim/arm/dbg_rdi.h",
 238     "sim/arm/parent.c", "sim/arm/armsupp.c", "sim/arm/armrdi.c",
 239     "sim/arm/bag.c", "sim/arm/armvirt.c", "sim/arm/main.c", "sim/arm/bag.h",
 240     "sim/arm/communicate.c", "sim/arm/gdbhost.h", "sim/arm/armfpe.h",
 241     "sim/arm/arminit.c",
 242     "sim/common/cgen-fpu.c", "sim/common/cgen-fpu.h",
 243     "sim/common/cgen-accfp.c",
 244     "sim/mips/m16run.c", "sim/mips/sim-main.c",
 245     "sim/moxie/moxie-gdb.dts",
 246     # Not a single file in sim/ppc/ appears to be copyright FSF :-(.
 247     "sim/ppc/filter.h", "sim/ppc/gen-support.h", "sim/ppc/ld-insn.h",
 248     "sim/ppc/hw_sem.c", "sim/ppc/hw_disk.c", "sim/ppc/idecode_branch.h",
 249     "sim/ppc/sim-endian.h", "sim/ppc/table.c", "sim/ppc/hw_core.c",
 250     "sim/ppc/gen-support.c", "sim/ppc/gen-semantics.h", "sim/ppc/cpu.h",
 251     "sim/ppc/sim_callbacks.h", "sim/ppc/RUN", "sim/ppc/Makefile.in",
 252     "sim/ppc/emul_chirp.c", "sim/ppc/hw_nvram.c", "sim/ppc/dc-test.01",
 253     "sim/ppc/hw_phb.c", "sim/ppc/hw_eeprom.c", "sim/ppc/bits.h",
 254     "sim/ppc/hw_vm.c", "sim/ppc/cap.h", "sim/ppc/os_emul.h",
 255     "sim/ppc/options.h", "sim/ppc/gen-idecode.c", "sim/ppc/filter.c",
 256     "sim/ppc/corefile-n.h", "sim/ppc/std-config.h", "sim/ppc/ld-decode.h",
 257     "sim/ppc/filter_filename.h", "sim/ppc/hw_shm.c",
 258     "sim/ppc/pk_disklabel.c", "sim/ppc/dc-simple", "sim/ppc/misc.h",
 259     "sim/ppc/device_table.h", "sim/ppc/ld-insn.c", "sim/ppc/inline.c",
 260     "sim/ppc/emul_bugapi.h", "sim/ppc/hw_cpu.h", "sim/ppc/debug.h",
 261     "sim/ppc/hw_ide.c", "sim/ppc/debug.c", "sim/ppc/gen-itable.h",
 262     "sim/ppc/interrupts.c", "sim/ppc/hw_glue.c", "sim/ppc/emul_unix.c",
 263     "sim/ppc/sim_calls.c", "sim/ppc/dc-complex", "sim/ppc/ld-cache.c",
 264     "sim/ppc/registers.h", "sim/ppc/dc-test.02", "sim/ppc/options.c",
 265     "sim/ppc/igen.h", "sim/ppc/registers.c", "sim/ppc/device.h",
 266     "sim/ppc/emul_chirp.h", "sim/ppc/hw_register.c", "sim/ppc/hw_init.c",
 267     "sim/ppc/sim-endian-n.h", "sim/ppc/filter_filename.c",
 268     "sim/ppc/bits.c", "sim/ppc/idecode_fields.h", "sim/ppc/hw_memory.c",
 269     "sim/ppc/misc.c", "sim/ppc/double.c", "sim/ppc/psim.h",
 270     "sim/ppc/hw_trace.c", "sim/ppc/emul_netbsd.h", "sim/ppc/psim.c",
 271     "sim/ppc/ppc-instructions", "sim/ppc/tree.h", "sim/ppc/README",
 272     "sim/ppc/gen-icache.h", "sim/ppc/gen-model.h", "sim/ppc/ld-cache.h",
 273     "sim/ppc/mon.c", "sim/ppc/corefile.h", "sim/ppc/vm.c",
 274     "sim/ppc/INSTALL", "sim/ppc/gen-model.c", "sim/ppc/hw_cpu.c",
 275     "sim/ppc/corefile.c", "sim/ppc/hw_opic.c", "sim/ppc/gen-icache.c",
 276     "sim/ppc/events.h", "sim/ppc/os_emul.c", "sim/ppc/emul_generic.c",
 277     "sim/ppc/main.c", "sim/ppc/hw_com.c", "sim/ppc/gen-semantics.c",
 278     "sim/ppc/emul_bugapi.c", "sim/ppc/device.c", "sim/ppc/emul_generic.h",
 279     "sim/ppc/tree.c", "sim/ppc/mon.h", "sim/ppc/interrupts.h",
 280     "sim/ppc/cap.c", "sim/ppc/cpu.c", "sim/ppc/hw_phb.h",
 281     "sim/ppc/device_table.c", "sim/ppc/lf.c", "sim/ppc/lf.c",
 282     "sim/ppc/dc-stupid", "sim/ppc/hw_pal.c", "sim/ppc/ppc-spr-table",
 283     "sim/ppc/emul_unix.h", "sim/ppc/words.h", "sim/ppc/basics.h",
 284     "sim/ppc/hw_htab.c", "sim/ppc/lf.h", "sim/ppc/ld-decode.c",
 285     "sim/ppc/sim-endian.c", "sim/ppc/gen-itable.c",
 286     "sim/ppc/idecode_expression.h", "sim/ppc/table.h", "sim/ppc/dgen.c",
 287     "sim/ppc/events.c", "sim/ppc/gen-idecode.h", "sim/ppc/emul_netbsd.c",
 288     "sim/ppc/igen.c", "sim/ppc/vm_n.h", "sim/ppc/vm.h",
 289     "sim/ppc/hw_iobus.c", "sim/ppc/inline.h",
 290     "sim/testsuite/sim/mips/mips32-dsp2.s",
 291 )
 292
 293 if __name__ == "__main__":
 294     main()
 295