Re: segfault in read_process_output() during vc-diff

[Kim F. Storm]

Handa-san,

Can you please take a look at this bug ... it seems to be coding
system related.

It is related to this assignment (in read_process_output)

  int carryover = XINT (p->decoding_carryover);

resulting in a negative value.

For this to be true, one of the occurrences of the following pieces of code:

       text = decode_coding_string (make_unibyte_string (chars, nbytes),
                                   coding, 0);
       ...                            
      carryover = nbytes - coding->consumed;

must have resulted in coding->consumed being bigger than nbytes...


"Tim Van Holder" <address@hidden> writes:

> On 11/7/06, Kim F. Storm <address@hidden> wrote:
>> > #1  0x081864d4 in read_process_output (proc=143372852, channel=<value 
>> > optimized out>) at /home/tim/gnu/src/emacs/src/process.c:5144
>> >         odeactivate = 137457249
>> >         text = 142872947
>> >         outer_running_asynch_code = 0
>> >         waiting = -1
>> >         nbytes = <value optimized out>
>> >         outstream = 137502441
>> >         old = (struct buffer *) 0x88fb2b8
>> >         p = (struct Lisp_Process *) 0x88bb230
>> >         opoint = <value optimized out>
>> >         coding = (struct coding_system *) 0x88fb4c0
>> >         carryover = -1316
>>
>> This looks very suspicious.
>>
>> Can you print the "coding" variable and the "p" variable:
>>
>> up 1
>> p *coding
>> p *p
>
> (gdb) p *coding
> $1 = {
>  type = coding_type_iso2022,
>  eol_type = 0,
>  common_flags = 7,
>  flags = 6144,
>  mode = 0,
>  composing = 0,
>  composition_rule_follows = 0,
>  cmp_data = 0x0,
>  cmp_data_start = 0,
>  cmp_data_index = 0,
>  spec = {
>    iso2022 = {
>      current_invocation = {0, 1},
>      current_designation = {0, 129, -1, -1},
>      initial_designation = {0, 129, -1, -1},
>      last_invalid_designation_register = -1,
>      requested_designation = "\000", '\004' <repeats 128 times>,
> "\001", '\004' <repeats 125 times>,
>      charset_revision_number = '\377' <repeats 255 times>,
>      single_shifting = 0,
>      bol = 1
>    },
>    ccl = {
>      decoder = {
>        idx = 0,
>        size = 1,
>        prog = 0x0,
>        ic = 129,
>        eof_ic = -1,
>        reg = {-1, 0, 129, -1, -1, -1, 67372032, 67372036},
>        private_state = 67372036,
>        last_block = 67372036,
>        status = 67372036,
>        buf_magnification = 67372036,
>        stack_idx = 67372036,
>        eol_type = 67372036,
>        multibyte = 67372036,
>        cr_consumed = 67372036,
>        suppress_error = 67372036,
>        eight_bit_control = 67372036
>      },
>      encoder = {
>        idx = 67372036,
>        size = 67372036,
>        prog = 0x4040404,
>        ic = 67372036,
>        eof_ic = 67372036,
>        reg = {67372036, 67372036, 67372036, 67372036, 67372036,
> 67372036, 67372036, 67372036},
>        private_state = 67372036,
>        last_block = 67372036,
>        status = 67372036,
>        buf_magnification = 67372036,
>        stack_idx = 67372036,
>        eol_type = 67372036,
>        multibyte = 67372036,
>        cr_consumed = 67371268,
>        suppress_error = 67372036,
>        eight_bit_control = 67372036
>      },
>      valid_codes = '\004' <repeats 115 times>, '\377' <repeats 141 times>,
>      cr_carryover = -1,
>      eight_bit_carryover = "\377\377\377\377"
>    }
>  },
>  category_idx = 4,
>  src_multibyte = 0,
>  dst_multibyte = 1,
>  heading_ascii = 2374,
>  produced = 2374,
>  produced_char = 2374,
>  consumed = 2374,
>  consumed_char = 2374,
>  errors = 0,
>  result = 1,
>  suppress_error = 0,
>  symbol = 137802753,
>  post_read_conversion = 137457249,
>  pre_write_conversion = 137457249,
>  translation_table_for_decode = 137457249,
>  translation_table_for_encode = 139621236
> }
>
> Suspicious indeed - I can't think of any reason why
> coding_type_iso2022 would be used (locale is en_US.UTF8, and language
> environment is English).
>
> (gdb) p *p
> $2 = {
>  size = 1073742364,
>  v_next = 0x8828c08,
>  infd = 40,
>  outfd = 104,
>  tty_name = 137457249,
>  name = 141293123,
>  command = 143161269,
>  filter = 143185969,
>  sentinel = 143488149,
>  log = 137457249,
>  buffer = 142774828,
>  childp = 137457297,
>  plist = 137457249,
>  mark = 140625498,
>  kill_without_query = 137457249,
>  status = 137603945,
>  pty_flag = 137457249,
>  tick = 8,
>  update_tick = 8,
>  decode_coding_system = 137802753,
>  decoding_buf = 143668003,
>  decoding_carryover = 0,
>  encode_coding_system = 137802705,
>  encoding_buf = 143667987,
>  encoding_carryover = 0,
>  inherit_coding_system_flag = 137457249,
>  filter_multibyte = 137457297,
>  adaptive_read_buffering = 137457297,
>  read_output_delay = 0,
>  read_output_skip = 137457249,
>  pid = 10948,
>  raw_status_new = -1,
>  raw_status = 256
> }

-- 
Kim F. Storm <address@hidden> http://www.cua.dk