Damaged dbs when killing app in debugger

jda jda at his.com
Tue May 31 14:47:29 CDT 2005


>On 5/31/05 3:01 PM, "jda" <jda at his.com> wrote:
>
>>  I flush after every 500 right now (and after the last one). But
>>  usually people import 1 or 2 at a time. What can I do, if anything,
>>  to prevent corruption (other than, I suppose, creating a db in RAM,
>>  import to that, and then transferring to the real database, which
>>  might work but I image would be pretty slow)?
>
>But you again can get corruption during transfer.
>In ideal we need transactions and double record of data.
>
>Why happens crash during import?
>
>You can reproduce crash ?
>

No, the crash was in my user's experience. I can get the corrupt 
database, perhaps, but that's too late.

Kem said he had a reproducible way of corrupting a database with no 
read/write going on -- did you get his database to test (and if not, 
Kem, could you send it)?

Jon


More information about the Valentina-beta mailing list