Harsha, are you working on this?

Yes Casrten , in 1.87+ i am not seeing such an issue. Any Suggestions

harsha, you dumped the DB and imported it without problems?

Harsha, what is the status on this?

Carsten , Sorry for the delay on this, actually i forgot this. Now i checked again and found out the issue is there. It is indeed better convert them to varchar (http://forums.mysql.com/read.php?24,105964,105964 ) for performance perspective as well.I will look into this.

They need to be text since varchar would be too small (they can contain several email addresses separated with a semikolon.

Ok , those tables having text email definitions are lime_old_tokens_49128_20100322225507 and lime_old_tokens_49128_20100322225619 (All are the old token tables).

but at some points (lime_surveys , lime_saved_control) it is defined email as varchar(320).

Carten, so what would better i do in this case.

You should fix the export so that it properly exports right SQL for the keys.

Carsten , for the raw tables there is no such issue at all as all the email fields are defined as varchar but for the token tables generated on the fly contain text definitions. So i made it correct. Please have a look.

Harsha, in your patch you are changing the type of the email field.
So on import the DB will look different than before. Please talk to me.

Exactly , that was my previous problem. Yes without changing to varchar(320) we can do Text(xxx).Is it better ? In that case can u suggest me a suitable length for the field.

i did try with Text(320) but no success. Same issue as before, it does not allow to import correctly. By changing them to varchar it works fine. So Carsten what should i do.

Carsten, Thanks for the time took on this issue. I learnt lot