Re: [sqlite] Release / development timeline

[Konstantin Malakhanov]

On 11/03/2004 02:16 PM, Peter Bartholdsson wrote:
> On Mon, 1 Nov 2004 19:23:58 -0800, Jonathon Blake
> <[EMAIL PROTECTED]> wrote:
> 
>> How about a simple page that gives projected release date for upgrades?
>>
>> Something like
>>
>> Version 3.0.9 beta  10 November
>> Version 3.0.9 final   20 November
>> Version 3.0.10 beta 1 December
>> Version 3.0.10 final 15 December
> 
> 
> All this would do is create pressure and people asking why stuff ain't
> done yet like the projected release date states. Especially since coders
> are optimists by nature when it comes to time estimates. :P
> 

It's always two weeks to the next release ;-)


-- 
Konstantin Malakhanov <[EMAIL PROTECTED]>

Re: [sqlite] Flush to disc

[Konstantin Malakhanov]

On 11/03/2004 08:48 PM, Reid Thompson wrote:
> [EMAIL PROTECTED] wrote:
> 
>>In a message dated 11/3/2004 12:20:36 PM Eastern Standard Time,
>>[EMAIL PROTECTED] writes: 
>>
>>How can  I make sure after an insert that everything has been flushed
>>to disc? 
>>
>>I  mean, I want even my program to become slower as long as I know
>> everything have been written to disc.
>>
>>At the moment I am using:
>>PRAGMA  default_synchronous = FULL;
>>PRAGMA synchronous = FULL;
>>
>>
>>And the  queries are all INSERT's without a  transaction.
>>
>>
>>Thx,
>>
>>KaReL (aka  Steven)
>>
>>Main Webpage : _http://www.karels0ft.be/
>>_ (http://www.karels0ft.be/) ICQ  #:  35217584
>>
>>
>>
>>If synchronus=FULL is set. every insert without a transaction is
>>flushed to disk before sqlite3_finalized is returned.  You are safe.
>>
>>Wei
> 
> 
> Unless you are using IDE disks that respond that they've written to the
> disk when in fact they have not.

..or you have some transaction-based/journaling filesystem which delays
data writing to disk ;-)

> reid
> 
> 


-- 
Konstantin Malakhanov <[EMAIL PROTECTED]>