RE: GNU make 3.81rc1 released -- please test.

[Paul D. Smith]

%% "Norman Wilson" <address@hidden> writes:

  nw> I don't want to rain on your parade, but there is at least one more core
  nw> dump lurking in this version:

Ouch!

  nw> (dbx) run
  nw> Running: make -k -j 4 variants=dyn_dbg --warn-undefined-variables
  nw> everything 
  nw> (process id 25869)
  nw> signal SEGV (no mapping at the fault address) in variable_hash_1 at line
  nw> 109 in file "variable.c"
  nw> (dbx) where
  nw> =>[1] variable_hash_1(keyv = 0x37b8e8), line 109 in "variable.c"
  nw>   [2] hash_find_slot(ht = 0xc47d0, key = 0x37b8e8), line 91 in "hash.c"
  nw>   [3] merge_variable_sets(to_set = 0xc47d0, from_set = 0x699530), line
  nw> 650 in "variable.c"
  nw>   [4] merge_variable_set_lists(setlist0 = 0x8b9ec4, setlist1 = 0x87ed8),
  nw> line 682 in "variable.c"
  nw>   [5] rehash_file(from_file = 0xad8e38, to_hname = 0x1a6c368
  nw> "/view/normanw_9.1/vobs/common/messages/src/Messages.msg"), line 298 in
  nw> "file.c"
  nw>   [6] f_mtime(file = 0xad8e38, search = 1), line 1247 in "remake.c"
  nw>   [7] check_dep(file = 0xad8e38, depth = 3U, this_mtime = 1ULL,
  nw> must_make_ptr = 0xffbfe658), line 971 in "remake.c"
  nw>   [8] check_dep(file = 0x4ddec0, depth = 2U, this_mtime = 1ULL,
  nw> must_make_ptr = 0xffbfe70c), line 1010 in "remake.c"
  nw>   [9] update_file_1(file = 0x98268, depth = 1U), line 504 in "remake.c"
  nw>   [10] update_file(file = 0x98268, depth = 0), line 306 in "remake.c"
  nw>   [11] update_goal_chain(goals = 0xa5afb0), line 154 in "remake.c"
  nw>   [12] main(argc = 7, argv = 0xffbff584, envp = 0xffbff5a4), line 2200
  nw> in "main.c"

  nw> This was there in the pervious version too. Do you want me to post a
  nw> bug?

By "previous version" do you mean 3.81beta4, or 3.80?


Yes, we definitely need to figure this out.  I'm happy to do it on the
mailing list, but it's probably a good idea to file a bug as well.
Offhand this backtrace doesn't tell me much unfortunately.  Can you
examine the value of *key in variable_hash_1 (the first frame)?  Does it
basically seem corrupted, like already-freed or overwritten memory?


This one looks like a bear :-/.

-- 
-------------------------------------------------------------------------------
 Paul D. Smith <address@hidden>          Find some GNU make tips at:
 http://www.gnu.org                      http://make.paulandlesley.org
 "Please remain calm...I may be mad, but I am a professional." --Mad Scientist