[PATCH v2: setxkbmap 2/3] Bug 34151: Potential bug and buffer overflow due to misused rules filename

[Alan Coopersmith]

On 02/12/11 12:48 AM, Dirk Wallenstein wrote:
> On Thu, Feb 10, 2011 at 11:01:58PM -0800, Alan Coopersmith wrote:
>> https://bugs.freedesktop.org/show_bug.cgi?id=34151
>>
>> Use rfName consistently, instead of sometimes reverting to svValue[RULES_NDX]
>>
>> Signed-off-by: Alan Coopersmith <alan.coopersmith at oracle.com>
>> ---
>>
>> This patch is actually appearing for the first time in version 2 of this
>> series.
>>
>>  setxkbmap.c |    4 ++--
>>  1 files changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/setxkbmap.c b/setxkbmap.c
>> index 511659b..9f92e2e 100644
>> --- a/setxkbmap.c
>> +++ b/setxkbmap.c
>> @@ -850,13 +850,13 @@ applyRules(void)
>>                            inclPath[i], rfName);
>>                      continue;
>>                  }
>> -                sprintf(buf, "%s/rules/%s", inclPath[i], svValue[RULES_NDX]);
>> +                sprintf(buf, "%s/rules/%s", inclPath[i], rfName);
>>                  rules = XkbRF_Load(buf, svValue[LOCALE_NDX], True, True);
>>              }
>>          }
>>          if (!rules)
>>          {
>> -            ERR1("Couldn't find rules file (%s) \n", svValue[RULES_NDX]);
>> +            ERR1("Couldn't find rules file (%s) \n", rfname);
> Typo:                                                   ^^^^^^ rfName

Yeah, that was a manual merge error when rebasing to insert in the middle of the
existing patch series.   I'd fixed in my tree but mailed the wrong version.

>>              return False;
>>          }
>>          /* Let the rules file to the magic, then update the svValues with
>> -- 
>> 1.7.3.2
>>
> 
> There is another use of svValue[RULES_NDX] further down.

Ah, yes, I guess we should report the same rules file we used, I'll fix that:

             MSG1("Applied rules from %s:\n", svValue[RULES_NDX]);



-- 
	-Alan Coopersmith-        alan.coopersmith at oracle.com
	 Oracle Solaris Platform Engineering: X Window System