gdb/copyright.py

   1 #! /usr/bin/env python3
   2
   3 # Copyright (C) 2011-2023 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 && ./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 argparse
  34 import datetime
  35 import locale
  36 import os
  37 import os.path
  38 import subprocess
  39 import sys
  40 from typing import List, Optional
  41
  42
  43 def get_update_list():
  44     """Return the list of files to update.
  45
  46     Assumes that the current working directory when called is the root
  47     of the GDB source tree (NOT the gdb/ subdirectory!).  The names of
  48     the files are relative to that root directory.
  49     """
  50     result = []
  51     for gdb_dir in (
  52         "gdb",
  53         "gdbserver",
  54         "gdbsupport",
  55         "gnulib",
  56         "sim",
  57         "include/gdb",
  58     ):
  59         for root, dirs, files in os.walk(gdb_dir, topdown=True):
  60             for dirname in dirs:
  61                 reldirname = "%s/%s" % (root, dirname)
  62                 if (
  63                     dirname in EXCLUDE_ALL_LIST
  64                     or reldirname in EXCLUDE_LIST
  65                     or reldirname in NOT_FSF_LIST
  66                     or reldirname in BY_HAND
  67                 ):
  68                     # Prune this directory from our search list.
  69                     dirs.remove(dirname)
  70             for filename in files:
  71                 relpath = "%s/%s" % (root, filename)
  72                 if (
  73                     filename in EXCLUDE_ALL_LIST
  74                     or relpath in EXCLUDE_LIST
  75                     or relpath in NOT_FSF_LIST
  76                     or relpath in BY_HAND
  77                 ):
  78                     # Ignore this file.
  79                     pass
  80                 else:
  81                     result.append(relpath)
  82     return result
  83
  84
  85 def update_files(update_list):
  86     """Update the copyright header of the files in the given list.
  87
  88     We use gnulib's update-copyright script for that.
  89     """
  90     # We want to use year intervals in the copyright notices, and
  91     # all years should be collapsed to one single year interval,
  92     # even if there are "holes" in the list of years found in the
  93     # original copyright notice (OK'ed by the FSF, case [gnu.org #719834]).
  94     os.environ["UPDATE_COPYRIGHT_USE_INTERVALS"] = "2"
  95
  96     # Perform the update, and save the output in a string.
  97     update_cmd = ["bash", "gnulib/import/extra/update-copyright"]
  98     update_cmd += update_list
  99
 100     p = subprocess.Popen(
 101         update_cmd,
 102         stdout=subprocess.PIPE,
 103         stderr=subprocess.STDOUT,
 104         encoding=locale.getpreferredencoding(),
 105     )
 106     update_out = p.communicate()[0]
 107
 108     # Process the output.  Typically, a lot of files do not have
 109     # a copyright notice :-(.  The update-copyright script prints
 110     # a well defined warning when it did not find the copyright notice.
 111     # For each of those, do a sanity check and see if they may in fact
 112     # have one.  For the files that are found not to have one, we filter
 113     # the line out from the output, since there is nothing more to do,
 114     # short of looking at each file and seeing which notice is appropriate.
 115     # Too much work! (~4,000 files listed as of 2012-01-03).
 116     update_out = update_out.splitlines(keepends=False)
 117     warning_string = ": warning: copyright statement not found"
 118     warning_len = len(warning_string)
 119
 120     for line in update_out:
 121         if line.endswith(warning_string):
 122             filename = line[:-warning_len]
 123             if may_have_copyright_notice(filename):
 124                 print(line)
 125         else:
 126             # Unrecognized file format. !?!
 127             print("*** " + line)
 128
 129
 130 def may_have_copyright_notice(filename):
 131     """Check that the given file does not seem to have a copyright notice.
 132
 133     The filename is relative to the root directory.
 134     This function assumes that the current working directory is that root
 135     directory.
 136
 137     The algorithm is fairly crude, meaning that it might return
 138     some false positives.  I do not think it will return any false
 139     negatives...  We might improve this function to handle more
 140     complex cases later...
 141     """
 142     # For now, it may have a copyright notice if we find the word
 143     # "Copyright" at the (reasonable) start of the given file, say
 144     # 50 lines...
 145     MAX_LINES = 50
 146
 147     # We don't really know what encoding each file might be following,
 148     # so just open the file as a byte stream. We only need to search
 149     # for a pattern that should be the same regardless of encoding,
 150     # so that should be good enough.
 151     with open(filename, "rb") as fd:
 152         for lineno, line in enumerate(fd, start=1):
 153             if b"Copyright" in line:
 154                 return True
 155             if lineno > MAX_LINES:
 156                 break
 157     return False
 158
 159
 160 def get_parser() -> argparse.ArgumentParser:
 161     """Get a command line parser."""
 162     parser = argparse.ArgumentParser(
 163         description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter
 164     )
 165     return parser
 166
 167
 168 def main(argv: List[str]) -> Optional[int]:
 169     """The main subprogram."""
 170     parser = get_parser()
 171     _ = parser.parse_args(argv)
 172     root_dir = os.path.dirname(os.getcwd())
 173     os.chdir(root_dir)
 174
 175     if not (
 176         os.path.isdir("gdb") and os.path.isfile("gnulib/import/extra/update-copyright")
 177     ):
 178         sys.exit("Error: This script must be called from the gdb directory.")
 179
 180     update_list = get_update_list()
 181     update_files(update_list)
 182
 183     # Remind the user that some files need to be updated by HAND...
 184
 185     if MULTIPLE_COPYRIGHT_HEADERS:
 186         print()
 187         print(
 188             "\033[31m"
 189             "REMINDER: Multiple copyright headers must be updated by hand:"
 190             "\033[0m"
 191         )
 192         for filename in MULTIPLE_COPYRIGHT_HEADERS:
 193             print("  ", filename)
 194
 195     if BY_HAND:
 196         print()
 197         print(
 198             "\033[31mREMINDER: The following files must be updated by hand." "\033[0m"
 199         )
 200         for filename in BY_HAND:
 201             print("  ", filename)
 202
 203
 204 ############################################################################
 205 #
 206 # Some constants, placed at the end because they take up a lot of room.
 207 # The actual value of these constants is not significant to the understanding
 208 # of the script.
 209 #
 210 ############################################################################
 211
 212 # Files which should not be modified, either because they are
 213 # generated, non-FSF, or otherwise special (e.g. license text,
 214 # or test cases which must be sensitive to line numbering).
 215 #
 216 # Filenames are relative to the root directory.
 217 EXCLUDE_LIST = (
 218     "gdb/nat/glibc_thread_db.h",
 219     "gdb/CONTRIBUTE",
 220     "gdbsupport/Makefile.in",
 221     "gnulib/import",
 222     "gnulib/config.in",
 223     "gnulib/Makefile.in",
 224 )
 225
 226 # Files which should not be modified, either because they are
 227 # generated, non-FSF, or otherwise special (e.g. license text,
 228 # or test cases which must be sensitive to line numbering).
 229 #
 230 # Matches any file or directory name anywhere.  Use with caution.
 231 # This is mostly for files that can be found in multiple directories.
 232 # Eg: We want all files named COPYING to be left untouched.
 233
 234 EXCLUDE_ALL_LIST = (
 235     "COPYING",
 236     "COPYING.LIB",
 237     "CVS",
 238     "configure",
 239     "copying.c",
 240     "fdl.texi",
 241     "gpl.texi",
 242     "aclocal.m4",
 243 )
 244
 245 # The list of files to update by hand.
 246 BY_HAND = (
 247     # Nothing at the moment :-).
 248 )
 249
 250 # Files containing multiple copyright headers.  This script is only
 251 # fixing the first one it finds, so we need to finish the update
 252 # by hand.
 253 MULTIPLE_COPYRIGHT_HEADERS = (
 254     "gdb/doc/gdb.texinfo",
 255     "gdb/doc/refcard.tex",
 256     "gdb/syscalls/update-netbsd.sh",
 257 )
 258
 259 # The list of file which have a copyright, but not held by the FSF.
 260 # Filenames are relative to the root directory.
 261 NOT_FSF_LIST = (
 262     "gdb/exc_request.defs",
 263     "gdb/gdbtk",
 264     "gdb/testsuite/gdb.gdbtk/",
 265     "sim/arm/armemu.h",
 266     "sim/arm/armos.c",
 267     "sim/arm/gdbhost.c",
 268     "sim/arm/dbg_hif.h",
 269     "sim/arm/dbg_conf.h",
 270     "sim/arm/communicate.h",
 271     "sim/arm/armos.h",
 272     "sim/arm/armcopro.c",
 273     "sim/arm/armemu.c",
 274     "sim/arm/kid.c",
 275     "sim/arm/thumbemu.c",
 276     "sim/arm/armdefs.h",
 277     "sim/arm/armopts.h",
 278     "sim/arm/dbg_cp.h",
 279     "sim/arm/dbg_rdi.h",
 280     "sim/arm/parent.c",
 281     "sim/arm/armsupp.c",
 282     "sim/arm/armrdi.c",
 283     "sim/arm/bag.c",
 284     "sim/arm/armvirt.c",
 285     "sim/arm/main.c",
 286     "sim/arm/bag.h",
 287     "sim/arm/communicate.c",
 288     "sim/arm/gdbhost.h",
 289     "sim/arm/armfpe.h",
 290     "sim/arm/arminit.c",
 291     "sim/common/cgen-fpu.c",
 292     "sim/common/cgen-fpu.h",
 293     "sim/common/cgen-accfp.c",
 294     "sim/mips/m16run.c",
 295     "sim/mips/sim-main.c",
 296     "sim/moxie/moxie-gdb.dts",
 297     # Not a single file in sim/ppc/ appears to be copyright FSF :-(.
 298     "sim/ppc/filter.h",
 299     "sim/ppc/gen-support.h",
 300     "sim/ppc/ld-insn.h",
 301     "sim/ppc/hw_sem.c",
 302     "sim/ppc/hw_disk.c",
 303     "sim/ppc/idecode_branch.h",
 304     "sim/ppc/sim-endian.h",
 305     "sim/ppc/table.c",
 306     "sim/ppc/hw_core.c",
 307     "sim/ppc/gen-support.c",
 308     "sim/ppc/gen-semantics.h",
 309     "sim/ppc/cpu.h",
 310     "sim/ppc/sim_callbacks.h",
 311     "sim/ppc/RUN",
 312     "sim/ppc/Makefile.in",
 313     "sim/ppc/emul_chirp.c",
 314     "sim/ppc/hw_nvram.c",
 315     "sim/ppc/dc-test.01",
 316     "sim/ppc/hw_phb.c",
 317     "sim/ppc/hw_eeprom.c",
 318     "sim/ppc/bits.h",
 319     "sim/ppc/hw_vm.c",
 320     "sim/ppc/cap.h",
 321     "sim/ppc/os_emul.h",
 322     "sim/ppc/options.h",
 323     "sim/ppc/gen-idecode.c",
 324     "sim/ppc/filter.c",
 325     "sim/ppc/corefile-n.h",
 326     "sim/ppc/std-config.h",
 327     "sim/ppc/ld-decode.h",
 328     "sim/ppc/filter_filename.h",
 329     "sim/ppc/hw_shm.c",
 330     "sim/ppc/pk_disklabel.c",
 331     "sim/ppc/dc-simple",
 332     "sim/ppc/misc.h",
 333     "sim/ppc/device_table.h",
 334     "sim/ppc/ld-insn.c",
 335     "sim/ppc/inline.c",
 336     "sim/ppc/emul_bugapi.h",
 337     "sim/ppc/hw_cpu.h",
 338     "sim/ppc/debug.h",
 339     "sim/ppc/hw_ide.c",
 340     "sim/ppc/debug.c",
 341     "sim/ppc/gen-itable.h",
 342     "sim/ppc/interrupts.c",
 343     "sim/ppc/hw_glue.c",
 344     "sim/ppc/emul_unix.c",
 345     "sim/ppc/sim_calls.c",
 346     "sim/ppc/dc-complex",
 347     "sim/ppc/ld-cache.c",
 348     "sim/ppc/registers.h",
 349     "sim/ppc/dc-test.02",
 350     "sim/ppc/options.c",
 351     "sim/ppc/igen.h",
 352     "sim/ppc/registers.c",
 353     "sim/ppc/device.h",
 354     "sim/ppc/emul_chirp.h",
 355     "sim/ppc/hw_register.c",
 356     "sim/ppc/hw_init.c",
 357     "sim/ppc/sim-endian-n.h",
 358     "sim/ppc/filter_filename.c",
 359     "sim/ppc/bits.c",
 360     "sim/ppc/idecode_fields.h",
 361     "sim/ppc/hw_memory.c",
 362     "sim/ppc/misc.c",
 363     "sim/ppc/double.c",
 364     "sim/ppc/psim.h",
 365     "sim/ppc/hw_trace.c",
 366     "sim/ppc/emul_netbsd.h",
 367     "sim/ppc/psim.c",
 368     "sim/ppc/powerpc.igen",
 369     "sim/ppc/tree.h",
 370     "sim/ppc/README",
 371     "sim/ppc/gen-icache.h",
 372     "sim/ppc/gen-model.h",
 373     "sim/ppc/ld-cache.h",
 374     "sim/ppc/mon.c",
 375     "sim/ppc/corefile.h",
 376     "sim/ppc/vm.c",
 377     "sim/ppc/INSTALL",
 378     "sim/ppc/gen-model.c",
 379     "sim/ppc/hw_cpu.c",
 380     "sim/ppc/corefile.c",
 381     "sim/ppc/hw_opic.c",
 382     "sim/ppc/gen-icache.c",
 383     "sim/ppc/events.h",
 384     "sim/ppc/os_emul.c",
 385     "sim/ppc/emul_generic.c",
 386     "sim/ppc/main.c",
 387     "sim/ppc/hw_com.c",
 388     "sim/ppc/gen-semantics.c",
 389     "sim/ppc/emul_bugapi.c",
 390     "sim/ppc/device.c",
 391     "sim/ppc/emul_generic.h",
 392     "sim/ppc/tree.c",
 393     "sim/ppc/mon.h",
 394     "sim/ppc/interrupts.h",
 395     "sim/ppc/cap.c",
 396     "sim/ppc/cpu.c",
 397     "sim/ppc/hw_phb.h",
 398     "sim/ppc/device_table.c",
 399     "sim/ppc/lf.c",
 400     "sim/ppc/lf.c",
 401     "sim/ppc/dc-stupid",
 402     "sim/ppc/hw_pal.c",
 403     "sim/ppc/ppc-spr-table",
 404     "sim/ppc/emul_unix.h",
 405     "sim/ppc/words.h",
 406     "sim/ppc/basics.h",
 407     "sim/ppc/hw_htab.c",
 408     "sim/ppc/lf.h",
 409     "sim/ppc/ld-decode.c",
 410     "sim/ppc/sim-endian.c",
 411     "sim/ppc/gen-itable.c",
 412     "sim/ppc/idecode_expression.h",
 413     "sim/ppc/table.h",
 414     "sim/ppc/dgen.c",
 415     "sim/ppc/events.c",
 416     "sim/ppc/gen-idecode.h",
 417     "sim/ppc/emul_netbsd.c",
 418     "sim/ppc/igen.c",
 419     "sim/ppc/vm_n.h",
 420     "sim/ppc/vm.h",
 421     "sim/ppc/hw_iobus.c",
 422     "sim/ppc/inline.h",
 423     "sim/testsuite/mips/mips32-dsp2.s",
 424 )
 425
 426 if __name__ == "__main__":
 427     sys.exit(main(sys.argv[1:]))