[Patch] Avoid duplicate names in /proc/registry (which may crash find)

[Christian Franke]

Hi Corinna,

Corinna Vinschen wrote:
> ...
>> With the attached patch, a duplicate name "foo" is handled as follows:
>>
>> - readdir() returns the key as "foo" and the value as "foo%val".
>> - If the name is "foo%val", stat() and open() consider only the value 
>> "foo".
>>
>> This keeps the names 'as is' if possible and allows access to the (very 
>> few) entries with duplicate names. The "%val" is at least somewhat 
>> self-explanatory.
>>     
>
> Cool.  Can you please send a ChangeLog entry as well?
>
>   

Of course:

2008-12-07  Christian Franke  <franke@computer.org>

	* fhandler_registry.cc (encode_regname): Add Parameter add_val.
	Append "%val" if add_val is set.
	(decode_regname): Remove trailing "%val". Change returncode
	accordingly.
	(__DIR_hash): New class.
	(d_hash): New macro.
	(key_exists): New function.
	(fhandler_registry::exists): Remove encode of registry name
	before path compare, decode file part of path instead.
	Skip checks for keys if trailing "%val" detected.
	(fhandler_registry::fstat): Change check of return
	value of decode_regname ().
	(fhandler_registry::readdir): Allocate __DIR_hash.
	Record key names in hash table. Append "%val" if key with
	same name exists. Fix error handling of encode_regname ().
	Set dirent.d_type.
	(fhandler_registry::closedir): Delete __DIR_hash.
	(fhandler_registry::open): Don't open key if trailing "%val"
	detected by decode_regname ().
	(open_key): Ditto.



Christian