This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Factor out rbreak_command with concat


I feel silly for not seeing the for loop..... :-)

What you say makes sense.

I wish there was a concat in libiberty that would reuse the buffer for us.

Patch withdrawn.

On Sun, Jul 26, 2015 at 7:12 PM, Doug Evans <xdje42@gmail.com> wrote:
> Ciro Santilli <ciro.santilli@gmail.com> writes:
>> The function behaviour should be unchanged.
>>
>> len was always 0 , so I removed it.
>>
>> xrealloc was always called on NULL string, so equivalent to xmalloc,
>> which concat uses.
>>
>> ---
>>  gdb/ChangeLog |  4 ++++
>>  gdb/symtab.c  | 35 +++++++++--------------------------
>>  2 files changed, 13 insertions(+), 26 deletions(-)
>>
>> diff --git a/gdb/ChangeLog b/gdb/ChangeLog
>> index 407dfb4..3d8c416 100644
>> --- a/gdb/ChangeLog
>> +++ b/gdb/ChangeLog
>> @@ -1,3 +1,7 @@
>> +4015-07-20  Ciro Santilli  <ciro.santilli@gmail.com>
>> +
>> +    * symtab.c (rbreak_command): Factor with concat.
>> +
>>  2015-07-15  Markus Metzger  <markus.t.metzger@intel.com>
>>          Pedro Alves <palves@redhat.com>
>>
>> diff --git a/gdb/symtab.c b/gdb/symtab.c
>> index decc5a9..015cb56 100644
>> --- a/gdb/symtab.c
>> +++ b/gdb/symtab.c
>> @@ -4889,7 +4889,6 @@ rbreak_command (char *regexp, int from_tty)
>>    struct symbol_search *p;
>>    struct cleanup *old_chain;
>>    char *string = NULL;
>> -  int len = 0;
>>    const char **files = NULL;
>>    const char *file_name;
>>    int nfiles = 0;
>> @@ -4928,20 +4927,11 @@ rbreak_command (char *regexp, int from_tty)
>>      {
>>        struct symtab *symtab = symbol_symtab (p->symbol);
>>        const char *fullname = symtab_to_fullname (symtab);
>> -
>> -      int newlen = (strlen (fullname)
>> -            + strlen (SYMBOL_LINKAGE_NAME (p->symbol))
>> -            + 4);
>> -
>> -      if (newlen > len)
>> -        {
>> -          string = xrealloc (string, newlen);
>> -          len = newlen;
>> -        }
>> -      strcpy (string, fullname);
>> -      strcat (string, ":'");
>> -      strcat (string, SYMBOL_LINKAGE_NAME (p->symbol));
>> -      strcat (string, "'");
>> +      string = concat(fullname,
>> +              ":'",
>> +              SYMBOL_LINKAGE_NAME (p->symbol),
>> +              "'",
>> +              (char *)NULL);
>>        break_command (string, from_tty);
>>        print_symbol_info (FUNCTIONS_DOMAIN,
>>                   p->symbol,
>> @@ -4950,17 +4940,10 @@ rbreak_command (char *regexp, int from_tty)
>>      }
>>        else
>>      {
>> -      int newlen = (strlen (MSYMBOL_LINKAGE_NAME (p->msymbol.minsym)) + 3);
>> -
>> -      if (newlen > len)
>> -        {
>> -          string = xrealloc (string, newlen);
>> -          len = newlen;
>> -        }
>> -      strcpy (string, "'");
>> -      strcat (string, MSYMBOL_LINKAGE_NAME (p->msymbol.minsym));
>> -      strcat (string, "'");
>> -
>> +      string = concat("'",
>> +              MSYMBOL_LINKAGE_NAME (p->msymbol.minsym),
>> +              "'",
>> +              (char *)NULL);
>>        break_command (string, from_tty);
>>        printf_filtered ("<function, no debug info> %s;\n",
>>                 MSYMBOL_PRINT_NAME (p->msymbol.minsym));
>
> Hi.
>
> Thanks for the patch.
>
> Things are a bit odd, no disagreement there.
> They are that way because note that all of this is
> done in a loop, and we reuse the buffer for each
> iteration.
>
> The above patch will leak memory.
>
> We *could* just concat the string for each iteration,
> but we have to ensure the string is freed, even if, say,
> break_command throws an error, so that means setting
> up a cleanup inside the loop. At this point it's
> probably a toss-up as to which is simpler, so I'd say
> let's leave things as is.
>
> Sound ok?


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]