[horde] switching to UTF8 database

[Jan Schneider]

Zitat von Andrew Morgan <morgan at orst.edu>:

> On Tue, 7 Feb 2012, Jan Schneider wrote:
>
>>> dump all the databases to an .sql file.
>>> move that file offsite.
>>> do the UTF-8 change you're comfortable with
>>> use mysqladmin to reimport it
>>> stop/start mysql and see if it complains.
>>> tell H3 to use the "new" db..
>>>
>>> Someone else can advise if you need to do more than that.
>>
>> It's not quite that easy, because that won't work with any  
>> serialized data that contains non-ascii strings. Those cannot  
>> simply be convert because they'd have different string lengths  
>> after that. This might affect identity data for example.
>
> Is this really a problem for latin1 to utf8?  If there are no  
> conversions that result in a double-byte character, then I think it  
> would be okay.

That has nothing to do with the charset though. If you data only  
contains ascii data, it doesn't matter to which charset you convert,  
indeed. As soon it latin1-specific characters, it will break.

> I wonder if I could test this with a string containing every  
> possible latin1 character, convert to utf8, and verify the string  
> length is the same.
>
> Thanks,
> 	Andy


Jan.

-- 
Do you need professional PHP or Horde consulting?
http://horde.org/consulting/