[TYPO3-core] RFC: Fix bug #5152: "Duplicate entry" in cache_pagesection when using frames

[Ernesto Baschny [cron IT]]

Michael Stucki wrote: on 04.02.2008 01:41:
> Martin Kutschker wrote:
> 
>>>> Sidenote: the system uses a super-short md5 hash (the first 7 bytes as
>>>> plain integer) for the MPvars. Would it make a difference to move on to
>>>> a full md5 hash here?
>>> Are you thinking if this would solve the duplicate error?
>> Unlikely, as MP vars are not nearly frequent enough. But it would
>> prevent a random conflict.
> 
> It could, yes. However, getting rid of t3lib_div::shortMD5() and
> t3lib_div::md5int() is nothing I consider as urgent, as we never got a
> proof of a failure caused by this shortage.

There are also indexed_search problems with the use of md5int (different 
words giving out the same md5int). While it is nice to have a short and 
fast int(32) for that purpose, the conflict potential is there.

Cheers,
Ernesto