Re: [Libcdio-devel] New git branch for bug 45017 ? Or re-use branch for 45015 ?

[Rocky Bernstein]

I've made a few minor changes to that branch so run git pull before
committing and pushing.

I still don't understand what "below does not exactly round up." means.

On Sun, Jul 30, 2017 at 7:30 PM, Rocky Bernstein <address@hidden> wrote:

> This is great. Thanks. Just commit in the branch you checked out. Then git
> push.
>
> When the dust is settled, we'll give others a chance to review and comment
> on before we merge that back into the main branch, which we'll make another
> release from.
>
>
> On Sun, Jul 30, 2017 at 2:22 PM, Thomas Schmitt <address@hidden> wrote:
>
>> Hi,
>>
>> may i commit the following change for
>>
>>   https://savannah.gnu.org/bugs/?45017
>>
>> (a neighbor of 45015) in the TS-RockRidge-Fix branch ?
>> Or shall i ask for a TS-iso-info-Fix branch ?
>>
>> -----------------------------------------------------------------------
>>
>> diff --git a/src/iso-info.c b/src/iso-info.c
>> index 212ab33..984d21a 100644
>> --- a/src/iso-info.c
>> +++ b/src/iso-info.c
>> @@ -61,6 +61,13 @@
>>  #define NORMAL ""
>>  #endif
>>
>> +/* TODO: Find a better place from where cd-info can read it too. */
>> +/*
>> +   ECMA-119 allows only a depth of 8 directories. Nobody obeys.
>> +   Rock Ridge allows path length 1023. This would be max depth 512.
>> +*/
>> +#define CDIO_MAX_DIR_RECURSION 512
>> +
>>  /* Used by `main' to communicate with `parse_opt'. And global options
>>   */
>>  static struct arguments
>> @@ -213,7 +220,8 @@ _log_handler (cdio_log_level_t level, const char
>> message[])
>>  }
>>
>>  static void
>> -print_iso9660_recurse (iso9660_t *p_iso, const char psz_path[])
>> +print_iso9660_recurse (iso9660_t *p_iso, const char psz_path[],
>> +                      unsigned int rec_counter)
>>  {
>>    CdioList_t *entlist;
>>    CdioList_t *dirlist =  _cdio_list_new ();
>> @@ -234,6 +242,15 @@ print_iso9660_recurse (iso9660_t *p_iso, const char
>> psz_path[])
>>      return;
>>    }
>>
>> +  rec_counter++;
>> +  if (rec_counter > CDIO_MAX_DIR_RECURSION) {
>> +    free(translated_name);
>> +    free(dirlist);
>> +    _cdio_list_free (entlist, true);
>> +    report( stderr,
>> +            "Directory recursion too deep. ISO most probably damaged.\n"
>> );
>> +    return;
>> +  }
>> +
>>    /* Iterate over files in this directory */
>>
>>    _CDIO_LIST_FOREACH (entnode, entlist)
>> @@ -297,7 +315,7 @@ print_iso9660_recurse (iso9660_t *p_iso, const char
>> psz_path[])
>>      {
>>        char *_fullname = _cdio_list_node_data (entnode);
>>
>> -      print_iso9660_recurse (p_iso, _fullname);
>> +      print_iso9660_recurse (p_iso, _fullname, rec_counter);
>>      }
>>
>>    _cdio_list_free (dirlist, true);
>> @@ -306,7 +324,7 @@ print_iso9660_recurse (iso9660_t *p_iso, const char
>> psz_path[])
>>  static void
>>  print_iso9660_fs (iso9660_t *iso)
>>  {
>> -  print_iso9660_recurse (iso, "/");
>> +  print_iso9660_recurse (iso, "/", 0);
>>  }
>>
>>  static void
>>
>> -----------------------------------------------------------------------
>>
>> With this change, iso-info --no-joliet -f still floods the terminal
>> with insane long paths "/a/0/0/0/..." but bails out with
>>   Directory recursion too deep. ISO most probably damaged.
>> before valgrind can crash.
>>   ==11439== All heap blocks were freed -- no leaks are possible
>>   ...
>>   ==11439== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
>>
>>
>> Have a nice day :)
>>
>> Thomas
>>
>>
>>
>