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