Re: [cowiki-dev] Releasing coWiki with the "working" parser

Daniel T. Gorski wrote:

>On 13 Jul 08:01, Paul Hanchett wrote:
>
>  
>
>>Apparently Daniel is checking to see if the database supports InnoDB, 
>>and if it doesn't the install fails and you can't get any further.  :-(  
>>    
>>
>
>Using a non-transactional persistence engine like MyISAM with coWiki (or
>better any software) is negligent, thats all.
>
>The installer process fails (or better: stops with a qualified message) to
>avoid any "wrong security feeling" for the user.
>
>We decided to support MySQL >= 4.0 which supports InnoDB without problems
>and I don't really see where the problems are now. Maybe it would be better
>to concentrate on other important issues. This one is solved IMO.
>  
>
4.0 does not universally support InnoDB Daniel.  Nor does 4.1.  The 
InnoDB engine has to be explicitly turned on and it may not be in all 
hosting situations.  I am really quite disappointed in this. 

I am also not satisfied with the analysis of where the coWiki "Magic" 
issue comes from.  The "magic" itself is not a /database /issue, it 
looks more like program logic.  Transactions will not address this 
problem at any level.  coWiki transactions take around 250mSec to 
complete.  How likely is it for a hardware failure to occur during any 
particular transactions?  If a failure occurs and one page is lost when 
it is being saved, I don't care too much as the Internet wire is less 
reliable than that.

Having transactions and using them when appropriate is good.  But for 
now, it looks more like a solution to a problem we don't have, and it 
creates other issues that /do/ affect the product.

:-(